{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "global_step": 25177,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.984126984126984e-06,
      "loss": 1.3412,
      "theoretical_loss": 3.515379199078347,
      "tokens_seen": 1650065408
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.968253968253968e-06,
      "loss": 1.2724,
      "theoretical_loss": 3.515367210984307,
      "tokens_seen": 1650130944
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.9523809523809525e-06,
      "loss": 1.2857,
      "theoretical_loss": 3.515355223499678,
      "tokens_seen": 1650196480
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.936507936507936e-06,
      "loss": 1.3914,
      "theoretical_loss": 3.515343236624406,
      "tokens_seen": 1650262016
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.92063492063492e-06,
      "loss": 1.2866,
      "theoretical_loss": 3.5153312503584346,
      "tokens_seen": 1650327552
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.1904761904761905e-05,
      "loss": 1.3488,
      "theoretical_loss": 3.5153192647017093,
      "tokens_seen": 1650393088
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.3888888888888888e-05,
      "loss": 1.2748,
      "theoretical_loss": 3.5153072796541744,
      "tokens_seen": 1650458624
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.5873015873015872e-05,
      "loss": 1.4147,
      "theoretical_loss": 3.5152952952157754,
      "tokens_seen": 1650524160
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.7857142857142855e-05,
      "loss": 1.3383,
      "theoretical_loss": 3.515283311386457,
      "tokens_seen": 1650589696
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.984126984126984e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.515271328166164,
      "tokens_seen": 1650655232
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.1825396825396824e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.515259345554841,
      "tokens_seen": 1650720768
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.380952380952381e-05,
      "loss": 1.2859,
      "theoretical_loss": 3.515247363552433,
      "tokens_seen": 1650786304
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.5793650793650793e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.515235382158885,
      "tokens_seen": 1650851840
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.7777777777777776e-05,
      "loss": 1.2359,
      "theoretical_loss": 3.5152234013741417,
      "tokens_seen": 1650917376
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.9761904761904762e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.5152114211981487,
      "tokens_seen": 1650982912
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.1746031746031745e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.51519944163085,
      "tokens_seen": 1651048448
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.373015873015873e-05,
      "loss": 1.3605,
      "theoretical_loss": 3.5151874626721904,
      "tokens_seen": 1651113984
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.571428571428571e-05,
      "loss": 1.3939,
      "theoretical_loss": 3.515175484322116,
      "tokens_seen": 1651179520
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.76984126984127e-05,
      "loss": 1.3198,
      "theoretical_loss": 3.5151635065805706,
      "tokens_seen": 1651245056
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.968253968253968e-05,
      "loss": 1.3317,
      "theoretical_loss": 3.5151515294474995,
      "tokens_seen": 1651310592
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.1666666666666665e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.515139552922848,
      "tokens_seen": 1651376128
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.365079365079365e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.5151275770065604,
      "tokens_seen": 1651441664
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.563492063492063e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.5151156016985823,
      "tokens_seen": 1651507200
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.761904761904762e-05,
      "loss": 1.318,
      "theoretical_loss": 3.5151036269988585,
      "tokens_seen": 1651572736
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.96031746031746e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.5150916529073335,
      "tokens_seen": 1651638272
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.1587301587301586e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.5150796794239527,
      "tokens_seen": 1651703808
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.357142857142857e-05,
      "loss": 1.332,
      "theoretical_loss": 3.5150677065486615,
      "tokens_seen": 1651769344
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.555555555555555e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.515055734281404,
      "tokens_seen": 1651834880
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.753968253968254e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.515043762622125,
      "tokens_seen": 1651900416
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.9523809523809524e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.5150317915707707,
      "tokens_seen": 1651965952
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.15079365079365e-05,
      "loss": 1.212,
      "theoretical_loss": 3.5150198211272854,
      "tokens_seen": 1652031488
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.349206349206349e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.5150078512916143,
      "tokens_seen": 1652097024
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.547619047619048e-05,
      "loss": 1.41,
      "theoretical_loss": 3.514995882063702,
      "tokens_seen": 1652162560
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.746031746031745e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.5149839134434946,
      "tokens_seen": 1652228096
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.944444444444444e-05,
      "loss": 1.3606,
      "theoretical_loss": 3.5149719454309363,
      "tokens_seen": 1652293632
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.142857142857142e-05,
      "loss": 1.4097,
      "theoretical_loss": 3.5149599780259715,
      "tokens_seen": 1652359168
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.341269841269842e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.514948011228547,
      "tokens_seen": 1652424704
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.53968253968254e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.5149360450386062,
      "tokens_seen": 1652490240
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.738095238095239e-05,
      "loss": 1.2061,
      "theoretical_loss": 3.514924079456095,
      "tokens_seen": 1652555776
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.936507936507937e-05,
      "loss": 1.305,
      "theoretical_loss": 3.5149121144809587,
      "tokens_seen": 1652621312
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.134920634920635e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.5149001501131423,
      "tokens_seen": 1652686848
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.333333333333333e-05,
      "loss": 1.227,
      "theoretical_loss": 3.5148881863525903,
      "tokens_seen": 1652752384
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.531746031746032e-05,
      "loss": 1.3487,
      "theoretical_loss": 3.5148762231992476,
      "tokens_seen": 1652817920
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.73015873015873e-05,
      "loss": 1.3457,
      "theoretical_loss": 3.5148642606530602,
      "tokens_seen": 1652883456
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.928571428571429e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.514852298713973,
      "tokens_seen": 1652948992
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.126984126984126e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.514840337381931,
      "tokens_seen": 1653014528
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.325396825396825e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.5148283766568795,
      "tokens_seen": 1653080064
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.523809523809524e-05,
      "loss": 1.3705,
      "theoretical_loss": 3.514816416538763,
      "tokens_seen": 1653145600
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.722222222222223e-05,
      "loss": 1.3992,
      "theoretical_loss": 3.5148044570275276,
      "tokens_seen": 1653211136
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 603740,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2631263732910156,
      "objective/train/theoretical_loss": 3.5147924981231173,
      "objective/train/tokens_used": 23735776,
      "theoretical_loss": 3.5147924981231173,
      "tokens_seen": 1653276672
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.92063492063492e-05,
      "loss": 1.261,
      "theoretical_loss": 3.5147924981231173,
      "tokens_seen": 1653276672
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001011904761904762,
      "loss": 1.3669,
      "theoretical_loss": 3.5147805398254786,
      "tokens_seen": 1653342208
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010317460317460317,
      "loss": 1.3943,
      "theoretical_loss": 3.514768582134555,
      "tokens_seen": 1653407744
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010515873015873016,
      "loss": 1.2742,
      "theoretical_loss": 3.5147566250502935,
      "tokens_seen": 1653473280
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010714285714285714,
      "loss": 1.2705,
      "theoretical_loss": 3.514744668572638,
      "tokens_seen": 1653538816
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010912698412698413,
      "loss": 1.3269,
      "theoretical_loss": 3.514732712701534,
      "tokens_seen": 1653604352
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001111111111111111,
      "loss": 1.3336,
      "theoretical_loss": 3.514720757436927,
      "tokens_seen": 1653669888
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011309523809523809,
      "loss": 1.2208,
      "theoretical_loss": 3.5147088027787623,
      "tokens_seen": 1653735424
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011507936507936508,
      "loss": 1.3446,
      "theoretical_loss": 3.5146968487269845,
      "tokens_seen": 1653800960
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011706349206349207,
      "loss": 1.2881,
      "theoretical_loss": 3.5146848952815386,
      "tokens_seen": 1653866496
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011904761904761905,
      "loss": 1.3218,
      "theoretical_loss": 3.5146729424423713,
      "tokens_seen": 1653932032
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012103174603174604,
      "loss": 1.2394,
      "theoretical_loss": 3.514660990209426,
      "tokens_seen": 1653997568
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000123015873015873,
      "loss": 1.3168,
      "theoretical_loss": 3.5146490385826494,
      "tokens_seen": 1654063104
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000125,
      "loss": 1.3394,
      "theoretical_loss": 3.5146370875619857,
      "tokens_seen": 1654128640
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012698412698412698,
      "loss": 1.1952,
      "theoretical_loss": 3.5146251371473807,
      "tokens_seen": 1654194176
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012896825396825398,
      "loss": 1.4024,
      "theoretical_loss": 3.51461318733878,
      "tokens_seen": 1654259712
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013095238095238096,
      "loss": 1.3534,
      "theoretical_loss": 3.5146012381361276,
      "tokens_seen": 1654325248
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013293650793650793,
      "loss": 1.2637,
      "theoretical_loss": 3.51458928953937,
      "tokens_seen": 1654390784
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001349206349206349,
      "loss": 1.3883,
      "theoretical_loss": 3.514577341548452,
      "tokens_seen": 1654456320
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001369047619047619,
      "loss": 1.3304,
      "theoretical_loss": 3.5145653941633186,
      "tokens_seen": 1654521856
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001388888888888889,
      "loss": 1.3348,
      "theoretical_loss": 3.514553447383916,
      "tokens_seen": 1654587392
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014087301587301586,
      "loss": 1.3753,
      "theoretical_loss": 3.5145415012101884,
      "tokens_seen": 1654652928
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014285714285714284,
      "loss": 1.2047,
      "theoretical_loss": 3.514529555642082,
      "tokens_seen": 1654718464
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014484126984126984,
      "loss": 1.4171,
      "theoretical_loss": 3.5145176106795417,
      "tokens_seen": 1654784000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014682539682539685,
      "loss": 1.2871,
      "theoretical_loss": 3.514505666322513,
      "tokens_seen": 1654849536
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014880952380952382,
      "loss": 1.2237,
      "theoretical_loss": 3.5144937225709407,
      "tokens_seen": 1654915072
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001507936507936508,
      "loss": 1.3116,
      "theoretical_loss": 3.514481779424771,
      "tokens_seen": 1654980608
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001527777777777778,
      "loss": 1.3522,
      "theoretical_loss": 3.5144698368839484,
      "tokens_seen": 1655046144
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015476190476190478,
      "loss": 1.2708,
      "theoretical_loss": 3.5144578949484186,
      "tokens_seen": 1655111680
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015674603174603175,
      "loss": 1.2083,
      "theoretical_loss": 3.514445953618127,
      "tokens_seen": 1655177216
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015873015873015873,
      "loss": 1.3442,
      "theoretical_loss": 3.5144340128930196,
      "tokens_seen": 1655242752
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016071428571428573,
      "loss": 1.3725,
      "theoretical_loss": 3.51442207277304,
      "tokens_seen": 1655308288
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001626984126984127,
      "loss": 1.2751,
      "theoretical_loss": 3.5144101332581354,
      "tokens_seen": 1655373824
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016468253968253969,
      "loss": 1.3079,
      "theoretical_loss": 3.5143981943482503,
      "tokens_seen": 1655439360
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016666666666666666,
      "loss": 1.4171,
      "theoretical_loss": 3.5143862560433305,
      "tokens_seen": 1655504896
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016865079365079366,
      "loss": 1.3876,
      "theoretical_loss": 3.5143743183433207,
      "tokens_seen": 1655570432
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017063492063492064,
      "loss": 1.2589,
      "theoretical_loss": 3.514362381248167,
      "tokens_seen": 1655635968
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017261904761904762,
      "loss": 1.2565,
      "theoretical_loss": 3.514350444757815,
      "tokens_seen": 1655701504
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001746031746031746,
      "loss": 1.3118,
      "theoretical_loss": 3.5143385088722088,
      "tokens_seen": 1655767040
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001765873015873016,
      "loss": 1.3472,
      "theoretical_loss": 3.5143265735912954,
      "tokens_seen": 1655832576
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017857142857142857,
      "loss": 1.2835,
      "theoretical_loss": 3.5143146389150193,
      "tokens_seen": 1655898112
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018055555555555555,
      "loss": 1.2546,
      "theoretical_loss": 3.5143027048433266,
      "tokens_seen": 1655963648
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018253968253968252,
      "loss": 1.3388,
      "theoretical_loss": 3.514290771376162,
      "tokens_seen": 1656029184
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018452380952380953,
      "loss": 1.3113,
      "theoretical_loss": 3.514278838513471,
      "tokens_seen": 1656094720
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001865079365079365,
      "loss": 1.3084,
      "theoretical_loss": 3.5142669062552,
      "tokens_seen": 1656160256
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001884920634920635,
      "loss": 1.2925,
      "theoretical_loss": 3.514254974601293,
      "tokens_seen": 1656225792
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019047619047619048,
      "loss": 1.2899,
      "theoretical_loss": 3.5142430435516974,
      "tokens_seen": 1656291328
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019246031746031748,
      "loss": 1.2437,
      "theoretical_loss": 3.514231113106357,
      "tokens_seen": 1656356864
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019444444444444446,
      "loss": 1.2868,
      "theoretical_loss": 3.514219183265218,
      "tokens_seen": 1656422400
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019642857142857144,
      "loss": 1.3238,
      "theoretical_loss": 3.514207254028226,
      "tokens_seen": 1656487936
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 604805,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4689288139343262,
      "objective/train/theoretical_loss": 3.5141953253953258,
      "objective/train/tokens_used": 27012576,
      "theoretical_loss": 3.5141953253953258,
      "tokens_seen": 1656553472
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001984126984126984,
      "loss": 1.3694,
      "theoretical_loss": 3.5141953253953258,
      "tokens_seen": 1656553472
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020039682539682542,
      "loss": 1.3317,
      "theoretical_loss": 3.5141833973664633,
      "tokens_seen": 1656619008
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002023809523809524,
      "loss": 1.2069,
      "theoretical_loss": 3.5141714699415845,
      "tokens_seen": 1656684544
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020436507936507937,
      "loss": 1.3143,
      "theoretical_loss": 3.5141595431206345,
      "tokens_seen": 1656750080
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020634920634920634,
      "loss": 1.287,
      "theoretical_loss": 3.514147616903559,
      "tokens_seen": 1656815616
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020833333333333335,
      "loss": 1.3504,
      "theoretical_loss": 3.5141356912903032,
      "tokens_seen": 1656881152
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021031746031746032,
      "loss": 1.2455,
      "theoretical_loss": 3.5141237662808127,
      "tokens_seen": 1656946688
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002123015873015873,
      "loss": 1.5026,
      "theoretical_loss": 3.5141118418750334,
      "tokens_seen": 1657012224
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021428571428571427,
      "loss": 1.2715,
      "theoretical_loss": 3.514099918072911,
      "tokens_seen": 1657077760
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021626984126984128,
      "loss": 1.2512,
      "theoretical_loss": 3.514087994874391,
      "tokens_seen": 1657143296
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021825396825396825,
      "loss": 1.2572,
      "theoretical_loss": 3.5140760722794178,
      "tokens_seen": 1657208832
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022023809523809523,
      "loss": 1.2935,
      "theoretical_loss": 3.5140641502879384,
      "tokens_seen": 1657274368
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002222222222222222,
      "loss": 1.2324,
      "theoretical_loss": 3.5140522288998977,
      "tokens_seen": 1657339904
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002242063492063492,
      "loss": 1.2474,
      "theoretical_loss": 3.514040308115242,
      "tokens_seen": 1657405440
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022619047619047618,
      "loss": 1.3313,
      "theoretical_loss": 3.514028387933916,
      "tokens_seen": 1657470976
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022817460317460316,
      "loss": 1.3137,
      "theoretical_loss": 3.514016468355866,
      "tokens_seen": 1657536512
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023015873015873016,
      "loss": 1.1736,
      "theoretical_loss": 3.5140045493810366,
      "tokens_seen": 1657602048
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023214285714285717,
      "loss": 1.3316,
      "theoretical_loss": 3.5139926310093745,
      "tokens_seen": 1657667584
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023412698412698414,
      "loss": 1.3608,
      "theoretical_loss": 3.5139807132408256,
      "tokens_seen": 1657733120
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023611111111111112,
      "loss": 1.2163,
      "theoretical_loss": 3.513968796075334,
      "tokens_seen": 1657798656
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002380952380952381,
      "loss": 1.3069,
      "theoretical_loss": 3.513956879512847,
      "tokens_seen": 1657864192
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002400793650793651,
      "loss": 1.2888,
      "theoretical_loss": 3.513944963553309,
      "tokens_seen": 1657929728
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024206349206349207,
      "loss": 1.3412,
      "theoretical_loss": 3.513933048196667,
      "tokens_seen": 1657995264
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024404761904761905,
      "loss": 1.4273,
      "theoretical_loss": 3.5139211334428646,
      "tokens_seen": 1658060800
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000246031746031746,
      "loss": 1.2545,
      "theoretical_loss": 3.5139092192918495,
      "tokens_seen": 1658126336
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000248015873015873,
      "loss": 1.2022,
      "theoretical_loss": 3.5138973057435665,
      "tokens_seen": 1658191872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025,
      "loss": 1.2708,
      "theoretical_loss": 3.5138853927979614,
      "tokens_seen": 1658257408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000251984126984127,
      "loss": 1.2957,
      "theoretical_loss": 3.51387348045498,
      "tokens_seen": 1658322944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025396825396825396,
      "loss": 1.3885,
      "theoretical_loss": 3.513861568714568,
      "tokens_seen": 1658388480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025595238095238096,
      "loss": 1.4945,
      "theoretical_loss": 3.513849657576671,
      "tokens_seen": 1658454016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025793650793650796,
      "loss": 1.254,
      "theoretical_loss": 3.513837747041234,
      "tokens_seen": 1658519552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025992063492063497,
      "loss": 1.3749,
      "theoretical_loss": 3.513825837108204,
      "tokens_seen": 1658585088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002619047619047619,
      "loss": 1.3178,
      "theoretical_loss": 3.513813927777526,
      "tokens_seen": 1658650624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002638888888888889,
      "loss": 1.3493,
      "theoretical_loss": 3.5138020190491464,
      "tokens_seen": 1658716160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026587301587301587,
      "loss": 1.304,
      "theoretical_loss": 3.5137901109230096,
      "tokens_seen": 1658781696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026785714285714287,
      "loss": 1.3476,
      "theoretical_loss": 3.513778203399063,
      "tokens_seen": 1658847232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002698412698412698,
      "loss": 1.2261,
      "theoretical_loss": 3.513766296477251,
      "tokens_seen": 1658912768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002718253968253968,
      "loss": 1.3979,
      "theoretical_loss": 3.51375439015752,
      "tokens_seen": 1658978304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002738095238095238,
      "loss": 1.3722,
      "theoretical_loss": 3.5137424844398164,
      "tokens_seen": 1659043840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027579365079365083,
      "loss": 1.3695,
      "theoretical_loss": 3.5137305793240845,
      "tokens_seen": 1659109376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002777777777777778,
      "loss": 1.3552,
      "theoretical_loss": 3.5137186748102716,
      "tokens_seen": 1659174912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002797619047619048,
      "loss": 1.32,
      "theoretical_loss": 3.513706770898322,
      "tokens_seen": 1659240448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028174603174603173,
      "loss": 1.398,
      "theoretical_loss": 3.5136948675881827,
      "tokens_seen": 1659305984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028373015873015873,
      "loss": 1.2293,
      "theoretical_loss": 3.5136829648797985,
      "tokens_seen": 1659371520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002857142857142857,
      "loss": 1.3544,
      "theoretical_loss": 3.5136710627731165,
      "tokens_seen": 1659437056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002876984126984127,
      "loss": 1.2462,
      "theoretical_loss": 3.5136591612680816,
      "tokens_seen": 1659502592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002896825396825397,
      "loss": 1.2311,
      "theoretical_loss": 3.51364726036464,
      "tokens_seen": 1659568128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002916666666666667,
      "loss": 1.4142,
      "theoretical_loss": 3.513635360062737,
      "tokens_seen": 1659633664
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002936507936507937,
      "loss": 1.4822,
      "theoretical_loss": 3.513623460362319,
      "tokens_seen": 1659699200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029563492063492064,
      "loss": 1.3401,
      "theoretical_loss": 3.5136115612633314,
      "tokens_seen": 1659764736
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 606029,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3237011432647705,
      "objective/train/theoretical_loss": 3.5135996627657207,
      "objective/train/tokens_used": 30289376,
      "theoretical_loss": 3.5135996627657207,
      "tokens_seen": 1659830272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029761904761904765,
      "loss": 1.3534,
      "theoretical_loss": 3.5135996627657207,
      "tokens_seen": 1659830272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002996031746031746,
      "loss": 1.3512,
      "theoretical_loss": 3.513587764869432,
      "tokens_seen": 1659895808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003015873015873016,
      "loss": 1.3128,
      "theoretical_loss": 3.5135758675744118,
      "tokens_seen": 1659961344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030357142857142855,
      "loss": 1.1756,
      "theoretical_loss": 3.5135639708806057,
      "tokens_seen": 1660026880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003055555555555556,
      "loss": 1.3124,
      "theoretical_loss": 3.5135520747879596,
      "tokens_seen": 1660092416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030753968253968255,
      "loss": 1.3834,
      "theoretical_loss": 3.51354017929642,
      "tokens_seen": 1660157952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030952380952380956,
      "loss": 1.4062,
      "theoretical_loss": 3.5135282844059317,
      "tokens_seen": 1660223488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003115079365079365,
      "loss": 1.2078,
      "theoretical_loss": 3.5135163901164406,
      "tokens_seen": 1660289024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003134920634920635,
      "loss": 1.4016,
      "theoretical_loss": 3.5135044964278936,
      "tokens_seen": 1660354560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031547619047619046,
      "loss": 1.3976,
      "theoretical_loss": 3.5134926033402367,
      "tokens_seen": 1660420096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031746031746031746,
      "loss": 1.213,
      "theoretical_loss": 3.5134807108534147,
      "tokens_seen": 1660485632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003194444444444444,
      "loss": 1.2538,
      "theoretical_loss": 3.513468818967374,
      "tokens_seen": 1660551168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032142857142857147,
      "loss": 1.2863,
      "theoretical_loss": 3.5134569276820606,
      "tokens_seen": 1660616704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003234126984126984,
      "loss": 1.2696,
      "theoretical_loss": 3.513445036997421,
      "tokens_seen": 1660682240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003253968253968254,
      "loss": 1.2223,
      "theoretical_loss": 3.5134331469134006,
      "tokens_seen": 1660747776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032738095238095237,
      "loss": 1.301,
      "theoretical_loss": 3.513421257429945,
      "tokens_seen": 1660813312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032936507936507937,
      "loss": 1.2914,
      "theoretical_loss": 3.5134093685470007,
      "tokens_seen": 1660878848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003313492063492063,
      "loss": 1.3476,
      "theoretical_loss": 3.513397480264514,
      "tokens_seen": 1660944384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003333333333333333,
      "loss": 1.3696,
      "theoretical_loss": 3.5133855925824298,
      "tokens_seen": 1661009920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003353174603174603,
      "loss": 1.4239,
      "theoretical_loss": 3.513373705500695,
      "tokens_seen": 1661075456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033730158730158733,
      "loss": 1.3598,
      "theoretical_loss": 3.5133618190192553,
      "tokens_seen": 1661140992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033928571428571433,
      "loss": 1.3815,
      "theoretical_loss": 3.513349933138057,
      "tokens_seen": 1661206528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003412698412698413,
      "loss": 1.3253,
      "theoretical_loss": 3.5133380478570455,
      "tokens_seen": 1661272064
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003432539682539683,
      "loss": 1.3166,
      "theoretical_loss": 3.513326163176167,
      "tokens_seen": 1661337600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034523809523809523,
      "loss": 1.2531,
      "theoretical_loss": 3.5133142790953684,
      "tokens_seen": 1661403136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034722222222222224,
      "loss": 1.2607,
      "theoretical_loss": 3.513302395614595,
      "tokens_seen": 1661468672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003492063492063492,
      "loss": 1.4398,
      "theoretical_loss": 3.513290512733792,
      "tokens_seen": 1661534208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003511904761904762,
      "loss": 1.3071,
      "theoretical_loss": 3.513278630452907,
      "tokens_seen": 1661599744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003531746031746032,
      "loss": 1.3396,
      "theoretical_loss": 3.513266748771885,
      "tokens_seen": 1661665280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003551587301587302,
      "loss": 1.2607,
      "theoretical_loss": 3.5132548676906725,
      "tokens_seen": 1661730816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035714285714285714,
      "loss": 1.3685,
      "theoretical_loss": 3.5132429872092157,
      "tokens_seen": 1661796352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035912698412698415,
      "loss": 1.2958,
      "theoretical_loss": 3.51323110732746,
      "tokens_seen": 1661861888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003611111111111111,
      "loss": 1.2959,
      "theoretical_loss": 3.5132192280453527,
      "tokens_seen": 1661927424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003630952380952381,
      "loss": 1.3178,
      "theoretical_loss": 3.5132073493628386,
      "tokens_seen": 1661992960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036507936507936505,
      "loss": 1.398,
      "theoretical_loss": 3.5131954712798645,
      "tokens_seen": 1662058496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036706349206349205,
      "loss": 1.3103,
      "theoretical_loss": 3.513183593796376,
      "tokens_seen": 1662124032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036904761904761905,
      "loss": 1.3852,
      "theoretical_loss": 3.5131717169123196,
      "tokens_seen": 1662189568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037103174603174606,
      "loss": 1.3375,
      "theoretical_loss": 3.5131598406276416,
      "tokens_seen": 1662255104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000373015873015873,
      "loss": 1.2439,
      "theoretical_loss": 3.513147964942288,
      "tokens_seen": 1662320640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000375,
      "loss": 1.3022,
      "theoretical_loss": 3.5131360898562045,
      "tokens_seen": 1662386176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000376984126984127,
      "loss": 1.2795,
      "theoretical_loss": 3.513124215369338,
      "tokens_seen": 1662451712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037896825396825396,
      "loss": 1.3061,
      "theoretical_loss": 3.5131123414816336,
      "tokens_seen": 1662517248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038095238095238096,
      "loss": 1.4298,
      "theoretical_loss": 3.513100468193038,
      "tokens_seen": 1662582784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038293650793650797,
      "loss": 1.2707,
      "theoretical_loss": 3.513088595503498,
      "tokens_seen": 1662648320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038492063492063497,
      "loss": 1.3499,
      "theoretical_loss": 3.5130767234129587,
      "tokens_seen": 1662713856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003869047619047619,
      "loss": 1.3215,
      "theoretical_loss": 3.5130648519213663,
      "tokens_seen": 1662779392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003888888888888889,
      "loss": 1.3652,
      "theoretical_loss": 3.513052981028668,
      "tokens_seen": 1662844928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039087301587301587,
      "loss": 1.3625,
      "theoretical_loss": 3.513041110734809,
      "tokens_seen": 1662910464
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003928571428571429,
      "loss": 1.4092,
      "theoretical_loss": 3.513029241039737,
      "tokens_seen": 1662976000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003948412698412698,
      "loss": 1.2667,
      "theoretical_loss": 3.5130173719433957,
      "tokens_seen": 1663041536
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 607267,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.356825590133667,
      "objective/train/theoretical_loss": 3.5130055034457337,
      "objective/train/tokens_used": 33566176,
      "theoretical_loss": 3.5130055034457337,
      "tokens_seen": 1663107072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003968253968253968,
      "loss": 1.3268,
      "theoretical_loss": 3.5130055034457337,
      "tokens_seen": 1663107072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039880952380952383,
      "loss": 1.3616,
      "theoretical_loss": 3.5129936355466955,
      "tokens_seen": 1663172608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040079365079365083,
      "loss": 1.3558,
      "theoretical_loss": 3.5129817682462283,
      "tokens_seen": 1663238144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004027777777777778,
      "loss": 1.3607,
      "theoretical_loss": 3.512969901544278,
      "tokens_seen": 1663303680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004047619047619048,
      "loss": 1.3915,
      "theoretical_loss": 3.512958035440791,
      "tokens_seen": 1663369216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040674603174603173,
      "loss": 1.3687,
      "theoretical_loss": 3.512946169935713,
      "tokens_seen": 1663434752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040873015873015874,
      "loss": 1.2725,
      "theoretical_loss": 3.512934305028991,
      "tokens_seen": 1663500288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004107142857142857,
      "loss": 1.3532,
      "theoretical_loss": 3.512922440720571,
      "tokens_seen": 1663565824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004126984126984127,
      "loss": 1.2948,
      "theoretical_loss": 3.512910577010399,
      "tokens_seen": 1663631360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004146825396825397,
      "loss": 1.3592,
      "theoretical_loss": 3.5128987138984216,
      "tokens_seen": 1663696896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004166666666666667,
      "loss": 1.3072,
      "theoretical_loss": 3.512886851384585,
      "tokens_seen": 1663762432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004186507936507937,
      "loss": 1.3437,
      "theoretical_loss": 3.512874989468835,
      "tokens_seen": 1663827968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042063492063492065,
      "loss": 1.3196,
      "theoretical_loss": 3.5128631281511185,
      "tokens_seen": 1663893504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042261904761904765,
      "loss": 1.3875,
      "theoretical_loss": 3.512851267431382,
      "tokens_seen": 1663959040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004246031746031746,
      "loss": 1.3927,
      "theoretical_loss": 3.5128394073095706,
      "tokens_seen": 1664024576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004265873015873016,
      "loss": 1.2842,
      "theoretical_loss": 3.512827547785632,
      "tokens_seen": 1664090112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042857142857142855,
      "loss": 1.3583,
      "theoretical_loss": 3.512815688859512,
      "tokens_seen": 1664155648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004305555555555556,
      "loss": 1.2375,
      "theoretical_loss": 3.5128038305311566,
      "tokens_seen": 1664221184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043253968253968256,
      "loss": 1.4035,
      "theoretical_loss": 3.512791972800512,
      "tokens_seen": 1664286720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043452380952380956,
      "loss": 1.2929,
      "theoretical_loss": 3.5127801156675256,
      "tokens_seen": 1664352256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004365079365079365,
      "loss": 1.3587,
      "theoretical_loss": 3.5127682591321427,
      "tokens_seen": 1664417792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004384920634920635,
      "loss": 1.3087,
      "theoretical_loss": 3.5127564031943104,
      "tokens_seen": 1664483328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044047619047619046,
      "loss": 1.269,
      "theoretical_loss": 3.5127445478539743,
      "tokens_seen": 1664548864
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044246031746031746,
      "loss": 1.2273,
      "theoretical_loss": 3.512732693111081,
      "tokens_seen": 1664614400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004444444444444444,
      "loss": 1.3967,
      "theoretical_loss": 3.5127208389655777,
      "tokens_seen": 1664679936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044642857142857147,
      "loss": 1.3524,
      "theoretical_loss": 3.51270898541741,
      "tokens_seen": 1664745472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004484126984126984,
      "loss": 1.3109,
      "theoretical_loss": 3.5126971324665237,
      "tokens_seen": 1664811008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004503968253968254,
      "loss": 1.3021,
      "theoretical_loss": 3.512685280112866,
      "tokens_seen": 1664876544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045238095238095237,
      "loss": 1.3615,
      "theoretical_loss": 3.5126734283563836,
      "tokens_seen": 1664942080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045436507936507937,
      "loss": 1.264,
      "theoretical_loss": 3.512661577197022,
      "tokens_seen": 1665007616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004563492063492063,
      "loss": 1.401,
      "theoretical_loss": 3.5126497266347285,
      "tokens_seen": 1665073152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004583333333333333,
      "loss": 1.3086,
      "theoretical_loss": 3.512637876669449,
      "tokens_seen": 1665138688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046031746031746033,
      "loss": 1.2858,
      "theoretical_loss": 3.5126260273011303,
      "tokens_seen": 1665204224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046230158730158733,
      "loss": 1.3467,
      "theoretical_loss": 3.5126141785297182,
      "tokens_seen": 1665269760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046428571428571433,
      "loss": 1.4068,
      "theoretical_loss": 3.5126023303551595,
      "tokens_seen": 1665335296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004662698412698413,
      "loss": 1.3382,
      "theoretical_loss": 3.5125904827774006,
      "tokens_seen": 1665400832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004682539682539683,
      "loss": 1.3011,
      "theoretical_loss": 3.5125786357963884,
      "tokens_seen": 1665466368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047023809523809523,
      "loss": 1.3789,
      "theoretical_loss": 3.5125667894120687,
      "tokens_seen": 1665531904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047222222222222224,
      "loss": 1.472,
      "theoretical_loss": 3.5125549436243886,
      "tokens_seen": 1665597440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004742063492063492,
      "loss": 1.3007,
      "theoretical_loss": 3.512543098433294,
      "tokens_seen": 1665662976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004761904761904762,
      "loss": 1.3371,
      "theoretical_loss": 3.5125312538387314,
      "tokens_seen": 1665728512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004781746031746032,
      "loss": 1.3715,
      "theoretical_loss": 3.512519409840648,
      "tokens_seen": 1665794048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004801587301587302,
      "loss": 1.4025,
      "theoretical_loss": 3.5125075664389893,
      "tokens_seen": 1665859584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048214285714285715,
      "loss": 1.3333,
      "theoretical_loss": 3.5124957236337027,
      "tokens_seen": 1665925120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048412698412698415,
      "loss": 1.4573,
      "theoretical_loss": 3.512483881424734,
      "tokens_seen": 1665990656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004861111111111111,
      "loss": 1.3031,
      "theoretical_loss": 3.5124720398120304,
      "tokens_seen": 1666056192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004880952380952381,
      "loss": 1.317,
      "theoretical_loss": 3.5124601987955377,
      "tokens_seen": 1666121728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004900793650793651,
      "loss": 1.3426,
      "theoretical_loss": 3.5124483583752033,
      "tokens_seen": 1666187264
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000492063492063492,
      "loss": 1.4375,
      "theoretical_loss": 3.5124365185509725,
      "tokens_seen": 1666252800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004940476190476191,
      "loss": 1.2994,
      "theoretical_loss": 3.512424679322793,
      "tokens_seen": 1666318336
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 608508,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4184010028839111,
      "objective/train/theoretical_loss": 3.5124128406906108,
      "objective/train/tokens_used": 36842976,
      "theoretical_loss": 3.5124128406906108,
      "tokens_seen": 1666383872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000496031746031746,
      "loss": 1.3533,
      "theoretical_loss": 3.5124128406906108,
      "tokens_seen": 1666383872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498015873015873,
      "loss": 1.4404,
      "theoretical_loss": 3.512401002654373,
      "tokens_seen": 1666449408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005,
      "loss": 1.4442,
      "theoretical_loss": 3.5123891652140253,
      "tokens_seen": 1666514944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999799398194584,
      "loss": 1.3187,
      "theoretical_loss": 3.5123773283695146,
      "tokens_seen": 1666580480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999598796389167,
      "loss": 1.4164,
      "theoretical_loss": 3.512365492120788,
      "tokens_seen": 1666646016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999398194583751,
      "loss": 1.4281,
      "theoretical_loss": 3.5123536564677913,
      "tokens_seen": 1666711552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999197592778335,
      "loss": 1.3973,
      "theoretical_loss": 3.5123418214104722,
      "tokens_seen": 1666777088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998996990972919,
      "loss": 1.2053,
      "theoretical_loss": 3.5123299869487763,
      "tokens_seen": 1666842624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998796389167503,
      "loss": 1.2739,
      "theoretical_loss": 3.51231815308265,
      "tokens_seen": 1666908160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998595787362087,
      "loss": 1.3377,
      "theoretical_loss": 3.5123063198120406,
      "tokens_seen": 1666973696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499839518555667,
      "loss": 1.3845,
      "theoretical_loss": 3.5122944871368946,
      "tokens_seen": 1667039232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998194583751254,
      "loss": 1.2625,
      "theoretical_loss": 3.5122826550571586,
      "tokens_seen": 1667104768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997993981945837,
      "loss": 1.2543,
      "theoretical_loss": 3.5122708235727798,
      "tokens_seen": 1667170304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997793380140421,
      "loss": 1.3257,
      "theoretical_loss": 3.5122589926837033,
      "tokens_seen": 1667235840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997592778335005,
      "loss": 1.3879,
      "theoretical_loss": 3.512247162389877,
      "tokens_seen": 1667301376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997392176529588,
      "loss": 1.3611,
      "theoretical_loss": 3.5122353326912474,
      "tokens_seen": 1667366912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997191574724173,
      "loss": 1.3385,
      "theoretical_loss": 3.512223503587761,
      "tokens_seen": 1667432448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996990972918757,
      "loss": 1.2879,
      "theoretical_loss": 3.512211675079364,
      "tokens_seen": 1667497984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499679037111334,
      "loss": 1.2707,
      "theoretical_loss": 3.512199847166004,
      "tokens_seen": 1667563520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996589769307924,
      "loss": 1.3643,
      "theoretical_loss": 3.512188019847627,
      "tokens_seen": 1667629056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996389167502507,
      "loss": 1.3053,
      "theoretical_loss": 3.5121761931241804,
      "tokens_seen": 1667694592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996188565697091,
      "loss": 1.2215,
      "theoretical_loss": 3.5121643669956097,
      "tokens_seen": 1667760128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995987963891675,
      "loss": 1.4085,
      "theoretical_loss": 3.512152541461863,
      "tokens_seen": 1667825664
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995787362086258,
      "loss": 1.5177,
      "theoretical_loss": 3.512140716522886,
      "tokens_seen": 1667891200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995586760280842,
      "loss": 1.2917,
      "theoretical_loss": 3.512128892178626,
      "tokens_seen": 1667956736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995386158475427,
      "loss": 1.261,
      "theoretical_loss": 3.5121170684290286,
      "tokens_seen": 1668022272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499518555667001,
      "loss": 1.3779,
      "theoretical_loss": 3.5121052452740424,
      "tokens_seen": 1668087808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994984954864594,
      "loss": 1.2618,
      "theoretical_loss": 3.5120934227136127,
      "tokens_seen": 1668153344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994784353059178,
      "loss": 1.3569,
      "theoretical_loss": 3.512081600747687,
      "tokens_seen": 1668218880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994583751253761,
      "loss": 1.2593,
      "theoretical_loss": 3.5120697793762115,
      "tokens_seen": 1668284416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994383149448345,
      "loss": 1.3794,
      "theoretical_loss": 3.512057958599133,
      "tokens_seen": 1668349952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994182547642928,
      "loss": 1.3364,
      "theoretical_loss": 3.512046138416399,
      "tokens_seen": 1668415488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993981945837512,
      "loss": 1.5521,
      "theoretical_loss": 3.512034318827955,
      "tokens_seen": 1668481024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993781344032096,
      "loss": 1.2601,
      "theoretical_loss": 3.512022499833749,
      "tokens_seen": 1668546560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499358074222668,
      "loss": 1.3203,
      "theoretical_loss": 3.5120106814337273,
      "tokens_seen": 1668612096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993380140421264,
      "loss": 1.3877,
      "theoretical_loss": 3.511998863627837,
      "tokens_seen": 1668677632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993179538615848,
      "loss": 1.4574,
      "theoretical_loss": 3.511987046416024,
      "tokens_seen": 1668743168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992978936810431,
      "loss": 1.3442,
      "theoretical_loss": 3.511975229798236,
      "tokens_seen": 1668808704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992778335005015,
      "loss": 1.3632,
      "theoretical_loss": 3.5119634137744193,
      "tokens_seen": 1668874240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992577733199598,
      "loss": 1.3682,
      "theoretical_loss": 3.5119515983445213,
      "tokens_seen": 1668939776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992377131394183,
      "loss": 1.2747,
      "theoretical_loss": 3.5119397835084882,
      "tokens_seen": 1669005312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992176529588767,
      "loss": 1.4697,
      "theoretical_loss": 3.5119279692662673,
      "tokens_seen": 1669070848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499197592778335,
      "loss": 1.2718,
      "theoretical_loss": 3.5119161556178047,
      "tokens_seen": 1669136384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991775325977934,
      "loss": 1.4316,
      "theoretical_loss": 3.511904342563048,
      "tokens_seen": 1669201920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991574724172518,
      "loss": 1.3667,
      "theoretical_loss": 3.511892530101944,
      "tokens_seen": 1669267456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991374122367101,
      "loss": 1.3733,
      "theoretical_loss": 3.5118807182344396,
      "tokens_seen": 1669332992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991173520561685,
      "loss": 1.4507,
      "theoretical_loss": 3.511868906960481,
      "tokens_seen": 1669398528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499097291875627,
      "loss": 1.4666,
      "theoretical_loss": 3.511857096280016,
      "tokens_seen": 1669464064
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990772316950853,
      "loss": 1.3904,
      "theoretical_loss": 3.5118452861929907,
      "tokens_seen": 1669529600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990571715145437,
      "loss": 1.3522,
      "theoretical_loss": 3.5118334766993526,
      "tokens_seen": 1669595136
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 609762,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4398548603057861,
      "objective/train/theoretical_loss": 3.5118216677990484,
      "objective/train/tokens_used": 40119776,
      "theoretical_loss": 3.5118216677990484,
      "tokens_seen": 1669660672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499037111334002,
      "loss": 1.4915,
      "theoretical_loss": 3.5118216677990484,
      "tokens_seen": 1669660672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990170511534604,
      "loss": 1.3718,
      "theoretical_loss": 3.5118098594920246,
      "tokens_seen": 1669726208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989969909729188,
      "loss": 1.4447,
      "theoretical_loss": 3.5117980517782286,
      "tokens_seen": 1669791744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989769307923771,
      "loss": 1.2739,
      "theoretical_loss": 3.511786244657607,
      "tokens_seen": 1669857280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989568706118355,
      "loss": 1.347,
      "theoretical_loss": 3.5117744381301064,
      "tokens_seen": 1669922816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989368104312939,
      "loss": 1.3521,
      "theoretical_loss": 3.5117626321956745,
      "tokens_seen": 1669988352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989167502507523,
      "loss": 1.3728,
      "theoretical_loss": 3.5117508268542585,
      "tokens_seen": 1670053888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988966900702107,
      "loss": 1.4433,
      "theoretical_loss": 3.5117390221058042,
      "tokens_seen": 1670119424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988766298896691,
      "loss": 1.4599,
      "theoretical_loss": 3.511727217950259,
      "tokens_seen": 1670184960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988565697091274,
      "loss": 1.4139,
      "theoretical_loss": 3.5117154143875706,
      "tokens_seen": 1670250496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988365095285858,
      "loss": 1.3556,
      "theoretical_loss": 3.511703611417685,
      "tokens_seen": 1670316032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988164493480441,
      "loss": 1.4031,
      "theoretical_loss": 3.5116918090405496,
      "tokens_seen": 1670381568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987963891675025,
      "loss": 1.3096,
      "theoretical_loss": 3.5116800072561114,
      "tokens_seen": 1670447104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987763289869609,
      "loss": 1.2903,
      "theoretical_loss": 3.511668206064317,
      "tokens_seen": 1670512640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987562688064192,
      "loss": 1.3447,
      "theoretical_loss": 3.5116564054651134,
      "tokens_seen": 1670578176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987362086258777,
      "loss": 1.2898,
      "theoretical_loss": 3.5116446054584483,
      "tokens_seen": 1670643712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987161484453361,
      "loss": 1.5065,
      "theoretical_loss": 3.5116328060442683,
      "tokens_seen": 1670709248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986960882647944,
      "loss": 1.3288,
      "theoretical_loss": 3.5116210072225202,
      "tokens_seen": 1670774784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986760280842528,
      "loss": 1.3256,
      "theoretical_loss": 3.511609208993151,
      "tokens_seen": 1670840320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986559679037111,
      "loss": 1.3815,
      "theoretical_loss": 3.5115974113561084,
      "tokens_seen": 1670905856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986359077231695,
      "loss": 1.3216,
      "theoretical_loss": 3.511585614311339,
      "tokens_seen": 1670971392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986158475426279,
      "loss": 1.3473,
      "theoretical_loss": 3.5115738178587894,
      "tokens_seen": 1671036928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985957873620862,
      "loss": 1.3758,
      "theoretical_loss": 3.511562021998407,
      "tokens_seen": 1671102464
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985757271815446,
      "loss": 1.3575,
      "theoretical_loss": 3.5115502267301393,
      "tokens_seen": 1671168000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985556670010031,
      "loss": 1.3945,
      "theoretical_loss": 3.5115384320539325,
      "tokens_seen": 1671233536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985356068204614,
      "loss": 1.5017,
      "theoretical_loss": 3.5115266379697343,
      "tokens_seen": 1671299072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985155466399198,
      "loss": 1.2951,
      "theoretical_loss": 3.5115148444774915,
      "tokens_seen": 1671364608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984954864593782,
      "loss": 1.4262,
      "theoretical_loss": 3.5115030515771517,
      "tokens_seen": 1671430144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984754262788365,
      "loss": 1.3975,
      "theoretical_loss": 3.511491259268661,
      "tokens_seen": 1671495680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984553660982949,
      "loss": 1.3532,
      "theoretical_loss": 3.5114794675519674,
      "tokens_seen": 1671561216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984353059177532,
      "loss": 1.4,
      "theoretical_loss": 3.5114676764270176,
      "tokens_seen": 1671626752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984152457372116,
      "loss": 1.3306,
      "theoretical_loss": 3.5114558858937586,
      "tokens_seen": 1671692288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049839518555667,
      "loss": 1.4179,
      "theoretical_loss": 3.511444095952138,
      "tokens_seen": 1671757824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983751253761284,
      "loss": 1.4374,
      "theoretical_loss": 3.511432306602102,
      "tokens_seen": 1671823360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983550651955868,
      "loss": 1.4393,
      "theoretical_loss": 3.5114205178435984,
      "tokens_seen": 1671888896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983350050150452,
      "loss": 1.456,
      "theoretical_loss": 3.511408729676575,
      "tokens_seen": 1671954432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983149448345035,
      "loss": 1.2084,
      "theoretical_loss": 3.5113969421009776,
      "tokens_seen": 1672019968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982948846539619,
      "loss": 1.3846,
      "theoretical_loss": 3.511385155116754,
      "tokens_seen": 1672085504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982748244734202,
      "loss": 1.4001,
      "theoretical_loss": 3.5113733687238513,
      "tokens_seen": 1672151040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982547642928786,
      "loss": 1.4125,
      "theoretical_loss": 3.511361582922217,
      "tokens_seen": 1672216576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498234704112337,
      "loss": 1.407,
      "theoretical_loss": 3.5113497977117976,
      "tokens_seen": 1672282112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982146439317953,
      "loss": 1.3244,
      "theoretical_loss": 3.5113380130925407,
      "tokens_seen": 1672347648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981945837512538,
      "loss": 1.3708,
      "theoretical_loss": 3.5113262290643936,
      "tokens_seen": 1672413184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981745235707122,
      "loss": 1.3375,
      "theoretical_loss": 3.5113144456273027,
      "tokens_seen": 1672478720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981544633901705,
      "loss": 1.4217,
      "theoretical_loss": 3.5113026627812163,
      "tokens_seen": 1672544256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981344032096289,
      "loss": 1.2906,
      "theoretical_loss": 3.5112908805260803,
      "tokens_seen": 1672609792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981143430290873,
      "loss": 1.4327,
      "theoretical_loss": 3.5112790988618436,
      "tokens_seen": 1672675328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980942828485456,
      "loss": 1.3056,
      "theoretical_loss": 3.511267317788452,
      "tokens_seen": 1672740864
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498074222668004,
      "loss": 1.3627,
      "theoretical_loss": 3.5112555373058534,
      "tokens_seen": 1672806400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980541624874623,
      "loss": 1.4094,
      "theoretical_loss": 3.511243757413995,
      "tokens_seen": 1672871936
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 611052,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.390806794166565,
      "objective/train/theoretical_loss": 3.511231978112823,
      "objective/train/tokens_used": 43396576,
      "theoretical_loss": 3.511231978112823,
      "tokens_seen": 1672937472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980341023069207,
      "loss": 1.3867,
      "theoretical_loss": 3.511231978112823,
      "tokens_seen": 1672937472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980140421263792,
      "loss": 1.4314,
      "theoretical_loss": 3.5112201994022865,
      "tokens_seen": 1673003008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979939819458375,
      "loss": 1.3568,
      "theoretical_loss": 3.5112084212823316,
      "tokens_seen": 1673068544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979739217652959,
      "loss": 1.4667,
      "theoretical_loss": 3.5111966437529056,
      "tokens_seen": 1673134080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979538615847543,
      "loss": 1.297,
      "theoretical_loss": 3.511184866813956,
      "tokens_seen": 1673199616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979338014042126,
      "loss": 1.5068,
      "theoretical_loss": 3.5111730904654297,
      "tokens_seen": 1673265152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497913741223671,
      "loss": 1.3578,
      "theoretical_loss": 3.5111613147072744,
      "tokens_seen": 1673330688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978936810431293,
      "loss": 1.4304,
      "theoretical_loss": 3.511149539539437,
      "tokens_seen": 1673396224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978736208625877,
      "loss": 1.3073,
      "theoretical_loss": 3.5111377649618656,
      "tokens_seen": 1673461760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978535606820461,
      "loss": 1.4159,
      "theoretical_loss": 3.5111259909745067,
      "tokens_seen": 1673527296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978335005015045,
      "loss": 1.3471,
      "theoretical_loss": 3.5111142175773074,
      "tokens_seen": 1673592832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978134403209629,
      "loss": 1.3852,
      "theoretical_loss": 3.5111024447702155,
      "tokens_seen": 1673658368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977933801404213,
      "loss": 1.2641,
      "theoretical_loss": 3.5110906725531787,
      "tokens_seen": 1673723904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977733199598796,
      "loss": 1.2829,
      "theoretical_loss": 3.5110789009261434,
      "tokens_seen": 1673789440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497753259779338,
      "loss": 1.4219,
      "theoretical_loss": 3.5110671298890574,
      "tokens_seen": 1673854976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977331995987965,
      "loss": 1.4023,
      "theoretical_loss": 3.511055359441868,
      "tokens_seen": 1673920512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977131394182548,
      "loss": 1.3657,
      "theoretical_loss": 3.5110435895845233,
      "tokens_seen": 1673986048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976930792377132,
      "loss": 1.413,
      "theoretical_loss": 3.511031820316969,
      "tokens_seen": 1674051584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976730190571715,
      "loss": 1.3866,
      "theoretical_loss": 3.511020051639154,
      "tokens_seen": 1674117120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976529588766299,
      "loss": 1.3012,
      "theoretical_loss": 3.511008283551025,
      "tokens_seen": 1674182656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976328986960883,
      "loss": 1.246,
      "theoretical_loss": 3.510996516052529,
      "tokens_seen": 1674248192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976128385155466,
      "loss": 1.4133,
      "theoretical_loss": 3.5109847491436144,
      "tokens_seen": 1674313728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497592778335005,
      "loss": 1.4379,
      "theoretical_loss": 3.5109729828242275,
      "tokens_seen": 1674379264
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975727181544635,
      "loss": 1.4259,
      "theoretical_loss": 3.5109612170943163,
      "tokens_seen": 1674444800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975526579739218,
      "loss": 1.2859,
      "theoretical_loss": 3.5109494519538282,
      "tokens_seen": 1674510336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975325977933802,
      "loss": 1.475,
      "theoretical_loss": 3.5109376874027105,
      "tokens_seen": 1674575872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975125376128386,
      "loss": 1.4248,
      "theoretical_loss": 3.5109259234409107,
      "tokens_seen": 1674641408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004974924774322969,
      "loss": 1.497,
      "theoretical_loss": 3.510914160068376,
      "tokens_seen": 1674706944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974724172517553,
      "loss": 1.307,
      "theoretical_loss": 3.5109023972850535,
      "tokens_seen": 1674772480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974523570712136,
      "loss": 1.3168,
      "theoretical_loss": 3.5108906350908917,
      "tokens_seen": 1674838016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497432296890672,
      "loss": 1.355,
      "theoretical_loss": 3.510878873485837,
      "tokens_seen": 1674903552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974122367101304,
      "loss": 1.3286,
      "theoretical_loss": 3.510867112469837,
      "tokens_seen": 1674969088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973921765295888,
      "loss": 1.3835,
      "theoretical_loss": 3.51085535204284,
      "tokens_seen": 1675034624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973721163490472,
      "loss": 1.3501,
      "theoretical_loss": 3.5108435922047927,
      "tokens_seen": 1675100160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973520561685056,
      "loss": 1.4653,
      "theoretical_loss": 3.5108318329556427,
      "tokens_seen": 1675165696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973319959879639,
      "loss": 1.3224,
      "theoretical_loss": 3.5108200742953377,
      "tokens_seen": 1675231232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973119358074223,
      "loss": 1.3738,
      "theoretical_loss": 3.510808316223825,
      "tokens_seen": 1675296768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972918756268806,
      "loss": 1.3234,
      "theoretical_loss": 3.5107965587410517,
      "tokens_seen": 1675362304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497271815446339,
      "loss": 1.1811,
      "theoretical_loss": 3.510784801846966,
      "tokens_seen": 1675427840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972517552657974,
      "loss": 1.2313,
      "theoretical_loss": 3.5107730455415145,
      "tokens_seen": 1675493376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972316950852557,
      "loss": 1.3912,
      "theoretical_loss": 3.5107612898246456,
      "tokens_seen": 1675558912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972116349047142,
      "loss": 1.2141,
      "theoretical_loss": 3.5107495346963065,
      "tokens_seen": 1675624448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971915747241726,
      "loss": 1.3425,
      "theoretical_loss": 3.510737780156445,
      "tokens_seen": 1675689984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971715145436309,
      "loss": 1.2733,
      "theoretical_loss": 3.510726026205008,
      "tokens_seen": 1675755520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971514543630893,
      "loss": 1.3946,
      "theoretical_loss": 3.5107142728419434,
      "tokens_seen": 1675821056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971313941825477,
      "loss": 1.4869,
      "theoretical_loss": 3.5107025200671984,
      "tokens_seen": 1675886592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497111334002006,
      "loss": 1.3577,
      "theoretical_loss": 3.510690767880721,
      "tokens_seen": 1675952128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970912738214644,
      "loss": 1.4416,
      "theoretical_loss": 3.510679016282459,
      "tokens_seen": 1676017664
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970712136409227,
      "loss": 1.2491,
      "theoretical_loss": 3.5106672652723594,
      "tokens_seen": 1676083200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970511534603811,
      "loss": 1.2802,
      "theoretical_loss": 3.51065551485037,
      "tokens_seen": 1676148736
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 612232,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.205704689025879,
      "objective/train/theoretical_loss": 3.5106437650164377,
      "objective/train/tokens_used": 46673376,
      "theoretical_loss": 3.5106437650164377,
      "tokens_seen": 1676214272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970310932798396,
      "loss": 1.3453,
      "theoretical_loss": 3.5106437650164377,
      "tokens_seen": 1676214272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970110330992979,
      "loss": 1.2999,
      "theoretical_loss": 3.5106320157705113,
      "tokens_seen": 1676279808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969909729187563,
      "loss": 1.3165,
      "theoretical_loss": 3.5106202671125377,
      "tokens_seen": 1676345344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969709127382147,
      "loss": 1.3722,
      "theoretical_loss": 3.5106085190424645,
      "tokens_seen": 1676410880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496950852557673,
      "loss": 1.3046,
      "theoretical_loss": 3.5105967715602393,
      "tokens_seen": 1676476416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969307923771314,
      "loss": 1.3223,
      "theoretical_loss": 3.5105850246658097,
      "tokens_seen": 1676541952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969107321965897,
      "loss": 1.288,
      "theoretical_loss": 3.5105732783591237,
      "tokens_seen": 1676607488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968906720160481,
      "loss": 1.3162,
      "theoretical_loss": 3.5105615326401285,
      "tokens_seen": 1676673024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968706118355065,
      "loss": 1.3171,
      "theoretical_loss": 3.5105497875087717,
      "tokens_seen": 1676738560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968505516549649,
      "loss": 1.3306,
      "theoretical_loss": 3.510538042965001,
      "tokens_seen": 1676804096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968304914744233,
      "loss": 1.3371,
      "theoretical_loss": 3.510526299008764,
      "tokens_seen": 1676869632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968104312938817,
      "loss": 1.4059,
      "theoretical_loss": 3.5105145556400092,
      "tokens_seen": 1676935168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049679037111334,
      "loss": 1.2618,
      "theoretical_loss": 3.5105028128586824,
      "tokens_seen": 1677000704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967703109327984,
      "loss": 1.327,
      "theoretical_loss": 3.5104910706647336,
      "tokens_seen": 1677066240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967502507522568,
      "loss": 1.3904,
      "theoretical_loss": 3.5104793290581084,
      "tokens_seen": 1677131776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967301905717151,
      "loss": 1.2958,
      "theoretical_loss": 3.5104675880387552,
      "tokens_seen": 1677197312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967101303911735,
      "loss": 1.3325,
      "theoretical_loss": 3.510455847606622,
      "tokens_seen": 1677262848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496690070210632,
      "loss": 1.4188,
      "theoretical_loss": 3.5104441077616566,
      "tokens_seen": 1677328384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966700100300903,
      "loss": 1.4538,
      "theoretical_loss": 3.510432368503806,
      "tokens_seen": 1677393920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966499498495487,
      "loss": 1.2618,
      "theoretical_loss": 3.5104206298330185,
      "tokens_seen": 1677459456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496629889669007,
      "loss": 1.247,
      "theoretical_loss": 3.5104088917492415,
      "tokens_seen": 1677524992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966098294884654,
      "loss": 1.2767,
      "theoretical_loss": 3.510397154252423,
      "tokens_seen": 1677590528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965897693079238,
      "loss": 1.274,
      "theoretical_loss": 3.51038541734251,
      "tokens_seen": 1677656064
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965697091273821,
      "loss": 1.3477,
      "theoretical_loss": 3.5103736810194515,
      "tokens_seen": 1677721600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965496489468405,
      "loss": 1.261,
      "theoretical_loss": 3.5103619452831936,
      "tokens_seen": 1677787136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965295887662988,
      "loss": 1.4634,
      "theoretical_loss": 3.5103502101336854,
      "tokens_seen": 1677852672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965095285857573,
      "loss": 1.3715,
      "theoretical_loss": 3.510338475570874,
      "tokens_seen": 1677918208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964894684052157,
      "loss": 1.3701,
      "theoretical_loss": 3.510326741594708,
      "tokens_seen": 1677983744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496469408224674,
      "loss": 1.3344,
      "theoretical_loss": 3.5103150082051338,
      "tokens_seen": 1678049280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964493480441324,
      "loss": 1.4372,
      "theoretical_loss": 3.5103032754021,
      "tokens_seen": 1678114816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964292878635908,
      "loss": 1.3899,
      "theoretical_loss": 3.5102915431855544,
      "tokens_seen": 1678180352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964092276830491,
      "loss": 1.2802,
      "theoretical_loss": 3.5102798115554448,
      "tokens_seen": 1678245888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963891675025075,
      "loss": 1.4888,
      "theoretical_loss": 3.510268080511718,
      "tokens_seen": 1678311424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963691073219659,
      "loss": 1.4316,
      "theoretical_loss": 3.5102563500543233,
      "tokens_seen": 1678376960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963490471414242,
      "loss": 1.3811,
      "theoretical_loss": 3.510244620183207,
      "tokens_seen": 1678442496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963289869608827,
      "loss": 1.3533,
      "theoretical_loss": 3.510232890898319,
      "tokens_seen": 1678508032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496308926780341,
      "loss": 1.402,
      "theoretical_loss": 3.510221162199605,
      "tokens_seen": 1678573568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962888665997994,
      "loss": 1.3002,
      "theoretical_loss": 3.5102094340870136,
      "tokens_seen": 1678639104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962688064192578,
      "loss": 1.4567,
      "theoretical_loss": 3.5101977065604926,
      "tokens_seen": 1678704640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962487462387161,
      "loss": 1.3611,
      "theoretical_loss": 3.5101859796199903,
      "tokens_seen": 1678770176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962286860581746,
      "loss": 1.3742,
      "theoretical_loss": 3.510174253265454,
      "tokens_seen": 1678835712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496208625877633,
      "loss": 1.3704,
      "theoretical_loss": 3.5101625274968318,
      "tokens_seen": 1678901248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961885656970913,
      "loss": 1.3149,
      "theoretical_loss": 3.510150802314071,
      "tokens_seen": 1678966784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961685055165497,
      "loss": 1.3109,
      "theoretical_loss": 3.5101390777171204,
      "tokens_seen": 1679032320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961484453360081,
      "loss": 1.4759,
      "theoretical_loss": 3.5101273537059274,
      "tokens_seen": 1679097856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961283851554664,
      "loss": 1.4503,
      "theoretical_loss": 3.5101156302804393,
      "tokens_seen": 1679163392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961083249749248,
      "loss": 1.3239,
      "theoretical_loss": 3.5101039074406053,
      "tokens_seen": 1679228928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960882647943831,
      "loss": 1.4195,
      "theoretical_loss": 3.5100921851863722,
      "tokens_seen": 1679294464
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960682046138415,
      "loss": 1.3329,
      "theoretical_loss": 3.5100804635176877,
      "tokens_seen": 1679360000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960481444333,
      "loss": 1.2157,
      "theoretical_loss": 3.510068742434501,
      "tokens_seen": 1679425536
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 613519,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4023518562316895,
      "objective/train/theoretical_loss": 3.510057021936759,
      "objective/train/tokens_used": 49950176,
      "theoretical_loss": 3.510057021936759,
      "tokens_seen": 1679491072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960280842527583,
      "loss": 1.3508,
      "theoretical_loss": 3.510057021936759,
      "tokens_seen": 1679491072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960080240722167,
      "loss": 1.3704,
      "theoretical_loss": 3.5100453020244093,
      "tokens_seen": 1679556608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959879638916751,
      "loss": 1.2502,
      "theoretical_loss": 3.510033582697401,
      "tokens_seen": 1679622144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959679037111334,
      "loss": 1.3835,
      "theoretical_loss": 3.5100218639556813,
      "tokens_seen": 1679687680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959478435305918,
      "loss": 1.3863,
      "theoretical_loss": 3.510010145799198,
      "tokens_seen": 1679753216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959277833500501,
      "loss": 1.2468,
      "theoretical_loss": 3.5099984282278993,
      "tokens_seen": 1679818752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959077231695085,
      "loss": 1.3369,
      "theoretical_loss": 3.5099867112417336,
      "tokens_seen": 1679884288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958876629889669,
      "loss": 1.3993,
      "theoretical_loss": 3.509974994840648,
      "tokens_seen": 1679949824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958676028084253,
      "loss": 1.2679,
      "theoretical_loss": 3.5099632790245905,
      "tokens_seen": 1680015360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958475426278837,
      "loss": 1.386,
      "theoretical_loss": 3.5099515637935097,
      "tokens_seen": 1680080896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958274824473421,
      "loss": 1.3126,
      "theoretical_loss": 3.5099398491473535,
      "tokens_seen": 1680146432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958074222668004,
      "loss": 1.342,
      "theoretical_loss": 3.5099281350860694,
      "tokens_seen": 1680211968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957873620862588,
      "loss": 1.3853,
      "theoretical_loss": 3.5099164216096055,
      "tokens_seen": 1680277504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957673019057172,
      "loss": 1.2512,
      "theoretical_loss": 3.5099047087179107,
      "tokens_seen": 1680343040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957472417251755,
      "loss": 1.2624,
      "theoretical_loss": 3.5098929964109313,
      "tokens_seen": 1680408576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957271815446339,
      "loss": 1.3223,
      "theoretical_loss": 3.5098812846886167,
      "tokens_seen": 1680474112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957071213640923,
      "loss": 1.2682,
      "theoretical_loss": 3.5098695735509144,
      "tokens_seen": 1680539648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956870611835507,
      "loss": 1.2981,
      "theoretical_loss": 3.509857862997773,
      "tokens_seen": 1680605184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956670010030091,
      "loss": 1.2197,
      "theoretical_loss": 3.509846153029139,
      "tokens_seen": 1680670720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956469408224674,
      "loss": 1.3112,
      "theoretical_loss": 3.509834443644962,
      "tokens_seen": 1680736256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956268806419258,
      "loss": 1.4174,
      "theoretical_loss": 3.50982273484519,
      "tokens_seen": 1680801792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956068204613842,
      "loss": 1.443,
      "theoretical_loss": 3.50981102662977,
      "tokens_seen": 1680867328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955867602808425,
      "loss": 1.4172,
      "theoretical_loss": 3.5097993189986507,
      "tokens_seen": 1680932864
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955667001003009,
      "loss": 1.399,
      "theoretical_loss": 3.50978761195178,
      "tokens_seen": 1680998400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955466399197592,
      "loss": 1.4391,
      "theoretical_loss": 3.5097759054891062,
      "tokens_seen": 1681063936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955265797392177,
      "loss": 1.3421,
      "theoretical_loss": 3.5097641996105766,
      "tokens_seen": 1681129472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955065195586761,
      "loss": 1.3453,
      "theoretical_loss": 3.5097524943161407,
      "tokens_seen": 1681195008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954864593781344,
      "loss": 1.4281,
      "theoretical_loss": 3.5097407896057455,
      "tokens_seen": 1681260544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954663991975928,
      "loss": 1.2651,
      "theoretical_loss": 3.5097290854793393,
      "tokens_seen": 1681326080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954463390170512,
      "loss": 1.4019,
      "theoretical_loss": 3.50971738193687,
      "tokens_seen": 1681391616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954262788365095,
      "loss": 1.2531,
      "theoretical_loss": 3.5097056789782863,
      "tokens_seen": 1681457152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954062186559679,
      "loss": 1.3387,
      "theoretical_loss": 3.509693976603536,
      "tokens_seen": 1681522688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953861584754263,
      "loss": 1.4915,
      "theoretical_loss": 3.509682274812567,
      "tokens_seen": 1681588224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953660982948846,
      "loss": 1.426,
      "theoretical_loss": 3.509670573605328,
      "tokens_seen": 1681653760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953460381143431,
      "loss": 1.2807,
      "theoretical_loss": 3.509658872981767,
      "tokens_seen": 1681719296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953259779338014,
      "loss": 1.3989,
      "theoretical_loss": 3.509647172941831,
      "tokens_seen": 1681784832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953059177532598,
      "loss": 1.2947,
      "theoretical_loss": 3.50963547348547,
      "tokens_seen": 1681850368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952858575727182,
      "loss": 1.3678,
      "theoretical_loss": 3.5096237746126304,
      "tokens_seen": 1681915904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952657973921765,
      "loss": 1.4324,
      "theoretical_loss": 3.509612076323261,
      "tokens_seen": 1681981440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952457372116349,
      "loss": 1.3791,
      "theoretical_loss": 3.5096003786173116,
      "tokens_seen": 1682046976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952256770310933,
      "loss": 1.3376,
      "theoretical_loss": 3.5095886814947277,
      "tokens_seen": 1682112512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952056168505516,
      "loss": 1.3442,
      "theoretical_loss": 3.509576984955459,
      "tokens_seen": 1682178048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049518555667001,
      "loss": 1.376,
      "theoretical_loss": 3.509565288999453,
      "tokens_seen": 1682243584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951654964894685,
      "loss": 1.4491,
      "theoretical_loss": 3.5095535936266593,
      "tokens_seen": 1682309120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951454363089268,
      "loss": 1.3967,
      "theoretical_loss": 3.509541898837024,
      "tokens_seen": 1682374656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951253761283852,
      "loss": 1.3527,
      "theoretical_loss": 3.5095302046304973,
      "tokens_seen": 1682440192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951053159478435,
      "loss": 1.3168,
      "theoretical_loss": 3.509518511007026,
      "tokens_seen": 1682505728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950852557673019,
      "loss": 1.3596,
      "theoretical_loss": 3.5095068179665585,
      "tokens_seen": 1682571264
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950651955867603,
      "loss": 1.3057,
      "theoretical_loss": 3.5094951255090443,
      "tokens_seen": 1682636800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950451354062186,
      "loss": 1.3464,
      "theoretical_loss": 3.50948343363443,
      "tokens_seen": 1682702336
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 614786,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2073283195495605,
      "objective/train/theoretical_loss": 3.5094717423426642,
      "objective/train/tokens_used": 53226976,
      "theoretical_loss": 3.5094717423426642,
      "tokens_seen": 1682767872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495025075225677,
      "loss": 1.3024,
      "theoretical_loss": 3.5094717423426642,
      "tokens_seen": 1682767872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950050150451354,
      "loss": 1.3789,
      "theoretical_loss": 3.5094600516336962,
      "tokens_seen": 1682833408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949849548645938,
      "loss": 1.2626,
      "theoretical_loss": 3.5094483615074736,
      "tokens_seen": 1682898944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949648946840522,
      "loss": 1.3576,
      "theoretical_loss": 3.509436671963944,
      "tokens_seen": 1682964480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949448345035105,
      "loss": 1.415,
      "theoretical_loss": 3.5094249830030564,
      "tokens_seen": 1683030016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949247743229689,
      "loss": 1.3767,
      "theoretical_loss": 3.509413294624759,
      "tokens_seen": 1683095552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949047141424273,
      "loss": 1.3648,
      "theoretical_loss": 3.509401606829,
      "tokens_seen": 1683161088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948846539618856,
      "loss": 1.3451,
      "theoretical_loss": 3.509389919615728,
      "tokens_seen": 1683226624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494864593781344,
      "loss": 1.4432,
      "theoretical_loss": 3.509378232984891,
      "tokens_seen": 1683292160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948445336008024,
      "loss": 1.3289,
      "theoretical_loss": 3.5093665469364366,
      "tokens_seen": 1683357696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948244734202607,
      "loss": 1.3216,
      "theoretical_loss": 3.509354861470314,
      "tokens_seen": 1683423232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948044132397192,
      "loss": 1.3884,
      "theoretical_loss": 3.509343176586472,
      "tokens_seen": 1683488768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947843530591776,
      "loss": 1.337,
      "theoretical_loss": 3.5093314922848577,
      "tokens_seen": 1683554304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947642928786359,
      "loss": 1.3516,
      "theoretical_loss": 3.5093198085654196,
      "tokens_seen": 1683619840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947442326980943,
      "loss": 1.4022,
      "theoretical_loss": 3.509308125428107,
      "tokens_seen": 1683685376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947241725175527,
      "loss": 1.3991,
      "theoretical_loss": 3.5092964428728672,
      "tokens_seen": 1683750912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947041123370111,
      "loss": 1.4298,
      "theoretical_loss": 3.509284760899649,
      "tokens_seen": 1683816448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946840521564695,
      "loss": 1.5053,
      "theoretical_loss": 3.509273079508401,
      "tokens_seen": 1683881984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946639919759278,
      "loss": 1.2488,
      "theoretical_loss": 3.5092613986990715,
      "tokens_seen": 1683947520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946439317953862,
      "loss": 1.5306,
      "theoretical_loss": 3.509249718471608,
      "tokens_seen": 1684013056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946238716148446,
      "loss": 1.3912,
      "theoretical_loss": 3.50923803882596,
      "tokens_seen": 1684078592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946038114343029,
      "loss": 1.3082,
      "theoretical_loss": 3.509226359762075,
      "tokens_seen": 1684144128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945837512537613,
      "loss": 1.4419,
      "theoretical_loss": 3.5092146812799023,
      "tokens_seen": 1684209664
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945636910732196,
      "loss": 1.4211,
      "theoretical_loss": 3.5092030033793895,
      "tokens_seen": 1684275200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945436308926781,
      "loss": 1.399,
      "theoretical_loss": 3.5091913260604857,
      "tokens_seen": 1684340736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945235707121365,
      "loss": 1.2682,
      "theoretical_loss": 3.509179649323138,
      "tokens_seen": 1684406272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945035105315948,
      "loss": 1.3093,
      "theoretical_loss": 3.5091679731672967,
      "tokens_seen": 1684471808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944834503510532,
      "loss": 1.4958,
      "theoretical_loss": 3.5091562975929085,
      "tokens_seen": 1684537344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944633901705116,
      "loss": 1.3395,
      "theoretical_loss": 3.5091446225999228,
      "tokens_seen": 1684602880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944433299899699,
      "loss": 1.3117,
      "theoretical_loss": 3.509132948188287,
      "tokens_seen": 1684668416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944232698094283,
      "loss": 1.4741,
      "theoretical_loss": 3.5091212743579514,
      "tokens_seen": 1684733952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944032096288867,
      "loss": 1.423,
      "theoretical_loss": 3.5091096011088627,
      "tokens_seen": 1684799488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494383149448345,
      "loss": 1.4385,
      "theoretical_loss": 3.5090979284409705,
      "tokens_seen": 1684865024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943630892678035,
      "loss": 1.554,
      "theoretical_loss": 3.5090862563542227,
      "tokens_seen": 1684930560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943430290872618,
      "loss": 1.5009,
      "theoretical_loss": 3.509074584848568,
      "tokens_seen": 1684996096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943229689067202,
      "loss": 1.2959,
      "theoretical_loss": 3.5090629139239535,
      "tokens_seen": 1685061632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943029087261786,
      "loss": 1.3549,
      "theoretical_loss": 3.50905124358033,
      "tokens_seen": 1685127168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942828485456369,
      "loss": 1.3463,
      "theoretical_loss": 3.5090395738176445,
      "tokens_seen": 1685192704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942627883650953,
      "loss": 1.337,
      "theoretical_loss": 3.509027904635846,
      "tokens_seen": 1685258240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942427281845537,
      "loss": 1.2665,
      "theoretical_loss": 3.509016236034883,
      "tokens_seen": 1685323776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494222668004012,
      "loss": 1.2958,
      "theoretical_loss": 3.5090045680147037,
      "tokens_seen": 1685389312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942026078234704,
      "loss": 1.2319,
      "theoretical_loss": 3.5089929005752563,
      "tokens_seen": 1685454848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941825476429289,
      "loss": 1.556,
      "theoretical_loss": 3.50898123371649,
      "tokens_seen": 1685520384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941624874623872,
      "loss": 1.4128,
      "theoretical_loss": 3.5089695674383536,
      "tokens_seen": 1685585920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941424272818456,
      "loss": 1.4533,
      "theoretical_loss": 3.5089579017407946,
      "tokens_seen": 1685651456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941223671013039,
      "loss": 1.3248,
      "theoretical_loss": 3.5089462366237623,
      "tokens_seen": 1685716992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941023069207623,
      "loss": 1.356,
      "theoretical_loss": 3.508934572087205,
      "tokens_seen": 1685782528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940822467402207,
      "loss": 1.3832,
      "theoretical_loss": 3.5089229081310713,
      "tokens_seen": 1685848064
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494062186559679,
      "loss": 1.3853,
      "theoretical_loss": 3.5089112447553097,
      "tokens_seen": 1685913600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940421263791374,
      "loss": 1.4201,
      "theoretical_loss": 3.508899581959869,
      "tokens_seen": 1685979136
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 615881,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4075050354003906,
      "objective/train/theoretical_loss": 3.5088879197446974,
      "objective/train/tokens_used": 56503776,
      "theoretical_loss": 3.5088879197446974,
      "tokens_seen": 1686044672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940220661985958,
      "loss": 1.4282,
      "theoretical_loss": 3.5088879197446974,
      "tokens_seen": 1686044672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940020060180542,
      "loss": 1.3533,
      "theoretical_loss": 3.5088762581097432,
      "tokens_seen": 1686110208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939819458375126,
      "loss": 1.4034,
      "theoretical_loss": 3.5088645970549557,
      "tokens_seen": 1686175744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493961885656971,
      "loss": 1.4432,
      "theoretical_loss": 3.5088529365802836,
      "tokens_seen": 1686241280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939418254764293,
      "loss": 1.3309,
      "theoretical_loss": 3.5088412766856747,
      "tokens_seen": 1686306816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939217652958877,
      "loss": 1.3937,
      "theoretical_loss": 3.5088296173710782,
      "tokens_seen": 1686372352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493901705115346,
      "loss": 1.366,
      "theoretical_loss": 3.508817958636442,
      "tokens_seen": 1686437888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938816449348044,
      "loss": 1.3288,
      "theoretical_loss": 3.508806300481716,
      "tokens_seen": 1686503424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938615847542628,
      "loss": 1.4774,
      "theoretical_loss": 3.5087946429068477,
      "tokens_seen": 1686568960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938415245737211,
      "loss": 1.4116,
      "theoretical_loss": 3.5087829859117856,
      "tokens_seen": 1686634496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938214643931796,
      "loss": 1.4398,
      "theoretical_loss": 3.5087713294964793,
      "tokens_seen": 1686700032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493801404212638,
      "loss": 1.3074,
      "theoretical_loss": 3.508759673660877,
      "tokens_seen": 1686765568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937813440320963,
      "loss": 1.2896,
      "theoretical_loss": 3.508748018404927,
      "tokens_seen": 1686831104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937612838515547,
      "loss": 1.5198,
      "theoretical_loss": 3.5087363637285787,
      "tokens_seen": 1686896640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493741223671013,
      "loss": 1.3359,
      "theoretical_loss": 3.50872470963178,
      "tokens_seen": 1686962176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937211634904714,
      "loss": 1.3666,
      "theoretical_loss": 3.5087130561144795,
      "tokens_seen": 1687027712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937011033099298,
      "loss": 1.2792,
      "theoretical_loss": 3.5087014031766266,
      "tokens_seen": 1687093248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936810431293881,
      "loss": 1.2595,
      "theoretical_loss": 3.50868975081817,
      "tokens_seen": 1687158784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936609829488465,
      "loss": 1.3133,
      "theoretical_loss": 3.5086780990390576,
      "tokens_seen": 1687224320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493640922768305,
      "loss": 1.3364,
      "theoretical_loss": 3.5086664478392384,
      "tokens_seen": 1687289856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936208625877633,
      "loss": 1.2262,
      "theoretical_loss": 3.5086547972186617,
      "tokens_seen": 1687355392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936008024072217,
      "loss": 1.3995,
      "theoretical_loss": 3.5086431471772754,
      "tokens_seen": 1687420928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049358074222668,
      "loss": 1.4076,
      "theoretical_loss": 3.5086314977150286,
      "tokens_seen": 1687486464
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935606820461384,
      "loss": 1.3261,
      "theoretical_loss": 3.50861984883187,
      "tokens_seen": 1687552000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935406218655968,
      "loss": 1.3928,
      "theoretical_loss": 3.508608200527749,
      "tokens_seen": 1687617536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935205616850551,
      "loss": 1.3427,
      "theoretical_loss": 3.5085965528026124,
      "tokens_seen": 1687683072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935005015045135,
      "loss": 1.4534,
      "theoretical_loss": 3.5085849056564107,
      "tokens_seen": 1687748608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934804413239719,
      "loss": 1.3425,
      "theoretical_loss": 3.508573259089092,
      "tokens_seen": 1687814144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934603811434303,
      "loss": 1.3352,
      "theoretical_loss": 3.508561613100605,
      "tokens_seen": 1687879680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934403209628887,
      "loss": 1.4193,
      "theoretical_loss": 3.5085499676908993,
      "tokens_seen": 1687945216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493420260782347,
      "loss": 1.4887,
      "theoretical_loss": 3.5085383228599225,
      "tokens_seen": 1688010752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934002006018054,
      "loss": 1.3698,
      "theoretical_loss": 3.5085266786076237,
      "tokens_seen": 1688076288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933801404212638,
      "loss": 1.551,
      "theoretical_loss": 3.5085150349339522,
      "tokens_seen": 1688141824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933600802407221,
      "loss": 1.3699,
      "theoretical_loss": 3.508503391838856,
      "tokens_seen": 1688207360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933400200601805,
      "loss": 1.4732,
      "theoretical_loss": 3.5084917493222845,
      "tokens_seen": 1688272896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933199598796389,
      "loss": 1.4211,
      "theoretical_loss": 3.5084801073841865,
      "tokens_seen": 1688338432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932998996990972,
      "loss": 1.3232,
      "theoretical_loss": 3.5084684660245107,
      "tokens_seen": 1688403968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932798395185557,
      "loss": 1.2814,
      "theoretical_loss": 3.5084568252432056,
      "tokens_seen": 1688469504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932597793380141,
      "loss": 1.355,
      "theoretical_loss": 3.5084451850402205,
      "tokens_seen": 1688535040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932397191574724,
      "loss": 1.3167,
      "theoretical_loss": 3.5084335454155036,
      "tokens_seen": 1688600576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932196589769308,
      "loss": 1.2502,
      "theoretical_loss": 3.508421906369004,
      "tokens_seen": 1688666112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931995987963893,
      "loss": 1.2834,
      "theoretical_loss": 3.508410267900671,
      "tokens_seen": 1688731648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931795386158476,
      "loss": 1.3767,
      "theoretical_loss": 3.5083986300104533,
      "tokens_seen": 1688797184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493159478435306,
      "loss": 1.3759,
      "theoretical_loss": 3.508386992698299,
      "tokens_seen": 1688862720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931394182547643,
      "loss": 1.297,
      "theoretical_loss": 3.5083753559641577,
      "tokens_seen": 1688928256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931193580742227,
      "loss": 1.46,
      "theoretical_loss": 3.508363719807978,
      "tokens_seen": 1688993792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930992978936811,
      "loss": 1.2731,
      "theoretical_loss": 3.508352084229709,
      "tokens_seen": 1689059328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930792377131394,
      "loss": 1.3691,
      "theoretical_loss": 3.5083404492292996,
      "tokens_seen": 1689124864
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930591775325978,
      "loss": 1.4393,
      "theoretical_loss": 3.508328814806698,
      "tokens_seen": 1689190400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930391173520562,
      "loss": 1.3628,
      "theoretical_loss": 3.5083171809618534,
      "tokens_seen": 1689255936
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 617011,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4949119091033936,
      "objective/train/theoretical_loss": 3.5083055476947154,
      "objective/train/tokens_used": 59780576,
      "theoretical_loss": 3.5083055476947154,
      "tokens_seen": 1689321472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930190571715146,
      "loss": 1.507,
      "theoretical_loss": 3.5083055476947154,
      "tokens_seen": 1689321472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492998996990973,
      "loss": 1.3042,
      "theoretical_loss": 3.508293915005232,
      "tokens_seen": 1689387008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929789368104313,
      "loss": 1.2017,
      "theoretical_loss": 3.5082822828933526,
      "tokens_seen": 1689452544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929588766298897,
      "loss": 1.2654,
      "theoretical_loss": 3.5082706513590263,
      "tokens_seen": 1689518080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929388164493481,
      "loss": 1.3117,
      "theoretical_loss": 3.5082590204022015,
      "tokens_seen": 1689583616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929187562688064,
      "loss": 1.3386,
      "theoretical_loss": 3.5082473900228273,
      "tokens_seen": 1689649152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928986960882648,
      "loss": 1.4195,
      "theoretical_loss": 3.5082357602208525,
      "tokens_seen": 1689714688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928786359077232,
      "loss": 1.4252,
      "theoretical_loss": 3.5082241309962265,
      "tokens_seen": 1689780224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928585757271815,
      "loss": 1.2852,
      "theoretical_loss": 3.508212502348898,
      "tokens_seen": 1689845760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049283851554664,
      "loss": 1.3667,
      "theoretical_loss": 3.508200874278816,
      "tokens_seen": 1689911296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928184553660984,
      "loss": 1.3233,
      "theoretical_loss": 3.5081892467859292,
      "tokens_seen": 1689976832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927983951855567,
      "loss": 1.3363,
      "theoretical_loss": 3.5081776198701866,
      "tokens_seen": 1690042368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927783350050151,
      "loss": 1.3989,
      "theoretical_loss": 3.5081659935315375,
      "tokens_seen": 1690107904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927582748244734,
      "loss": 1.3096,
      "theoretical_loss": 3.5081543677699307,
      "tokens_seen": 1690173440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927382146439318,
      "loss": 1.2772,
      "theoretical_loss": 3.5081427425853153,
      "tokens_seen": 1690238976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927181544633902,
      "loss": 1.3291,
      "theoretical_loss": 3.5081311179776398,
      "tokens_seen": 1690304512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926980942828485,
      "loss": 1.3775,
      "theoretical_loss": 3.508119493946854,
      "tokens_seen": 1690370048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926780341023069,
      "loss": 1.4747,
      "theoretical_loss": 3.5081078704929065,
      "tokens_seen": 1690435584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926579739217654,
      "loss": 1.3508,
      "theoretical_loss": 3.5080962476157462,
      "tokens_seen": 1690501120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926379137412237,
      "loss": 1.4896,
      "theoretical_loss": 3.5080846253153224,
      "tokens_seen": 1690566656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926178535606821,
      "loss": 1.5249,
      "theoretical_loss": 3.508073003591584,
      "tokens_seen": 1690632192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925977933801404,
      "loss": 1.3723,
      "theoretical_loss": 3.5080613824444793,
      "tokens_seen": 1690697728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925777331995988,
      "loss": 1.3005,
      "theoretical_loss": 3.508049761873959,
      "tokens_seen": 1690763264
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925576730190572,
      "loss": 1.463,
      "theoretical_loss": 3.5080381418799704,
      "tokens_seen": 1690828800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925376128385155,
      "loss": 1.3577,
      "theoretical_loss": 3.5080265224624636,
      "tokens_seen": 1690894336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925175526579739,
      "loss": 1.3403,
      "theoretical_loss": 3.5080149036213877,
      "tokens_seen": 1690959872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924974924774323,
      "loss": 1.3324,
      "theoretical_loss": 3.508003285356691,
      "tokens_seen": 1691025408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924774322968907,
      "loss": 1.4333,
      "theoretical_loss": 3.5079916676683234,
      "tokens_seen": 1691090944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924573721163491,
      "loss": 1.3066,
      "theoretical_loss": 3.507980050556233,
      "tokens_seen": 1691156480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924373119358075,
      "loss": 1.3331,
      "theoretical_loss": 3.50796843402037,
      "tokens_seen": 1691222016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004924172517552658,
      "loss": 1.3279,
      "theoretical_loss": 3.507956818060683,
      "tokens_seen": 1691287552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923971915747242,
      "loss": 1.4721,
      "theoretical_loss": 3.5079452026771207,
      "tokens_seen": 1691353088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923771313941825,
      "loss": 1.3156,
      "theoretical_loss": 3.507933587869633,
      "tokens_seen": 1691418624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923570712136409,
      "loss": 1.5105,
      "theoretical_loss": 3.5079219736381684,
      "tokens_seen": 1691484160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923370110330993,
      "loss": 1.4068,
      "theoretical_loss": 3.507910359982676,
      "tokens_seen": 1691549696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923169508525576,
      "loss": 1.2907,
      "theoretical_loss": 3.5078987469031055,
      "tokens_seen": 1691615232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922968906720161,
      "loss": 1.3368,
      "theoretical_loss": 3.507887134399405,
      "tokens_seen": 1691680768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922768304914745,
      "loss": 1.4479,
      "theoretical_loss": 3.507875522471525,
      "tokens_seen": 1691746304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922567703109328,
      "loss": 1.3563,
      "theoretical_loss": 3.5078639111194136,
      "tokens_seen": 1691811840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922367101303912,
      "loss": 1.4157,
      "theoretical_loss": 3.5078523003430204,
      "tokens_seen": 1691877376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922166499498495,
      "loss": 1.2107,
      "theoretical_loss": 3.5078406901422943,
      "tokens_seen": 1691942912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921965897693079,
      "loss": 1.2928,
      "theoretical_loss": 3.5078290805171846,
      "tokens_seen": 1692008448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921765295887663,
      "loss": 1.4624,
      "theoretical_loss": 3.5078174714676407,
      "tokens_seen": 1692073984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921564694082246,
      "loss": 1.3345,
      "theoretical_loss": 3.507805862993611,
      "tokens_seen": 1692139520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492136409227683,
      "loss": 1.3296,
      "theoretical_loss": 3.507794255095046,
      "tokens_seen": 1692205056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921163490471415,
      "loss": 1.3384,
      "theoretical_loss": 3.5077826477718936,
      "tokens_seen": 1692270592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920962888665998,
      "loss": 1.4161,
      "theoretical_loss": 3.5077710410241036,
      "tokens_seen": 1692336128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920762286860582,
      "loss": 1.3567,
      "theoretical_loss": 3.507759434851625,
      "tokens_seen": 1692401664
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920561685055166,
      "loss": 1.4252,
      "theoretical_loss": 3.5077478292544075,
      "tokens_seen": 1692467200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920361083249749,
      "loss": 1.4003,
      "theoretical_loss": 3.5077362242323993,
      "tokens_seen": 1692532736
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 618329,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.301423192024231,
      "objective/train/theoretical_loss": 3.5077246197855505,
      "objective/train/tokens_used": 63057376,
      "theoretical_loss": 3.5077246197855505,
      "tokens_seen": 1692598272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920160481444333,
      "loss": 1.2787,
      "theoretical_loss": 3.5077246197855505,
      "tokens_seen": 1692598272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919959879638916,
      "loss": 1.4332,
      "theoretical_loss": 3.5077130159138106,
      "tokens_seen": 1692663808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049197592778335,
      "loss": 1.3741,
      "theoretical_loss": 3.507701412617128,
      "tokens_seen": 1692729344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919558676028084,
      "loss": 1.4026,
      "theoretical_loss": 3.507689809895452,
      "tokens_seen": 1692794880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919358074222668,
      "loss": 1.4147,
      "theoretical_loss": 3.5076782077487323,
      "tokens_seen": 1692860416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919157472417252,
      "loss": 1.3272,
      "theoretical_loss": 3.507666606176918,
      "tokens_seen": 1692925952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918956870611836,
      "loss": 1.41,
      "theoretical_loss": 3.5076550051799575,
      "tokens_seen": 1692991488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918756268806419,
      "loss": 1.3977,
      "theoretical_loss": 3.507643404757802,
      "tokens_seen": 1693057024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918555667001003,
      "loss": 1.3732,
      "theoretical_loss": 3.5076318049103987,
      "tokens_seen": 1693122560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918355065195586,
      "loss": 1.471,
      "theoretical_loss": 3.507620205637698,
      "tokens_seen": 1693188096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491815446339017,
      "loss": 1.2343,
      "theoretical_loss": 3.50760860693965,
      "tokens_seen": 1693253632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917953861584754,
      "loss": 1.4344,
      "theoretical_loss": 3.507597008816202,
      "tokens_seen": 1693319168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917753259779337,
      "loss": 1.3792,
      "theoretical_loss": 3.5075854112673044,
      "tokens_seen": 1693384704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917552657973922,
      "loss": 1.2945,
      "theoretical_loss": 3.5075738142929067,
      "tokens_seen": 1693450240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917352056168506,
      "loss": 1.3872,
      "theoretical_loss": 3.507562217892958,
      "tokens_seen": 1693515776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917151454363089,
      "loss": 1.2753,
      "theoretical_loss": 3.5075506220674066,
      "tokens_seen": 1693581312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916950852557673,
      "loss": 1.3919,
      "theoretical_loss": 3.507539026816204,
      "tokens_seen": 1693646848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916750250752258,
      "loss": 1.4268,
      "theoretical_loss": 3.5075274321392973,
      "tokens_seen": 1693712384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916549648946841,
      "loss": 1.3616,
      "theoretical_loss": 3.507515838036637,
      "tokens_seen": 1693777920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916349047141425,
      "loss": 1.468,
      "theoretical_loss": 3.5075042445081728,
      "tokens_seen": 1693843456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916148445336008,
      "loss": 1.4037,
      "theoretical_loss": 3.5074926515538527,
      "tokens_seen": 1693908992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915947843530592,
      "loss": 1.3366,
      "theoretical_loss": 3.507481059173627,
      "tokens_seen": 1693974528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915747241725176,
      "loss": 1.3876,
      "theoretical_loss": 3.5074694673674456,
      "tokens_seen": 1694040064
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915546639919759,
      "loss": 1.4898,
      "theoretical_loss": 3.507457876135257,
      "tokens_seen": 1694105600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915346038114343,
      "loss": 1.4334,
      "theoretical_loss": 3.5074462854770103,
      "tokens_seen": 1694171136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915145436308927,
      "loss": 1.3677,
      "theoretical_loss": 3.5074346953926554,
      "tokens_seen": 1694236672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914944834503511,
      "loss": 1.2385,
      "theoretical_loss": 3.507423105882142,
      "tokens_seen": 1694302208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914744232698095,
      "loss": 1.4356,
      "theoretical_loss": 3.5074115169454188,
      "tokens_seen": 1694367744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914543630892679,
      "loss": 1.4225,
      "theoretical_loss": 3.5073999285824353,
      "tokens_seen": 1694433280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914343029087262,
      "loss": 1.3465,
      "theoretical_loss": 3.507388340793142,
      "tokens_seen": 1694498816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914142427281846,
      "loss": 1.3653,
      "theoretical_loss": 3.5073767535774865,
      "tokens_seen": 1694564352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913941825476429,
      "loss": 1.4899,
      "theoretical_loss": 3.5073651669354193,
      "tokens_seen": 1694629888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913741223671013,
      "loss": 1.3602,
      "theoretical_loss": 3.50735358086689,
      "tokens_seen": 1694695424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913540621865597,
      "loss": 1.3722,
      "theoretical_loss": 3.5073419953718474,
      "tokens_seen": 1694760960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491334002006018,
      "loss": 1.3544,
      "theoretical_loss": 3.507330410450242,
      "tokens_seen": 1694826496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913139418254765,
      "loss": 1.4186,
      "theoretical_loss": 3.507318826102022,
      "tokens_seen": 1694892032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912938816449349,
      "loss": 1.4275,
      "theoretical_loss": 3.507307242327137,
      "tokens_seen": 1694957568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912738214643932,
      "loss": 1.4761,
      "theoretical_loss": 3.5072956591255373,
      "tokens_seen": 1695023104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912537612838516,
      "loss": 1.2628,
      "theoretical_loss": 3.5072840764971716,
      "tokens_seen": 1695088640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049123370110331,
      "loss": 1.3808,
      "theoretical_loss": 3.50727249444199,
      "tokens_seen": 1695154176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912136409227683,
      "loss": 1.344,
      "theoretical_loss": 3.507260912959941,
      "tokens_seen": 1695219712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911935807422267,
      "loss": 1.3068,
      "theoretical_loss": 3.507249332050975,
      "tokens_seen": 1695285248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491173520561685,
      "loss": 1.3323,
      "theoretical_loss": 3.5072377517150413,
      "tokens_seen": 1695350784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911534603811434,
      "loss": 1.2759,
      "theoretical_loss": 3.507226171952089,
      "tokens_seen": 1695416320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911334002006019,
      "loss": 1.4506,
      "theoretical_loss": 3.507214592762068,
      "tokens_seen": 1695481856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911133400200602,
      "loss": 1.3608,
      "theoretical_loss": 3.5072030141449284,
      "tokens_seen": 1695547392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910932798395186,
      "loss": 1.319,
      "theoretical_loss": 3.507191436100618,
      "tokens_seen": 1695612928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491073219658977,
      "loss": 1.3914,
      "theoretical_loss": 3.5071798586290877,
      "tokens_seen": 1695678464
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910531594784353,
      "loss": 1.3683,
      "theoretical_loss": 3.507168281730287,
      "tokens_seen": 1695744000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910330992978937,
      "loss": 1.3267,
      "theoretical_loss": 3.5071567054041646,
      "tokens_seen": 1695809536
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 619583,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.226139783859253,
      "objective/train/theoretical_loss": 3.507145129650671,
      "objective/train/tokens_used": 66334176,
      "theoretical_loss": 3.507145129650671,
      "tokens_seen": 1695875072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491013039117352,
      "loss": 1.3059,
      "theoretical_loss": 3.507145129650671,
      "tokens_seen": 1695875072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909929789368104,
      "loss": 1.4155,
      "theoretical_loss": 3.507133554469755,
      "tokens_seen": 1695940608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909729187562688,
      "loss": 1.4839,
      "theoretical_loss": 3.507121979861367,
      "tokens_seen": 1696006144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909528585757272,
      "loss": 1.336,
      "theoretical_loss": 3.507110405825455,
      "tokens_seen": 1696071680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909327983951856,
      "loss": 1.3128,
      "theoretical_loss": 3.50709883236197,
      "tokens_seen": 1696137216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490912738214644,
      "loss": 1.3394,
      "theoretical_loss": 3.507087259470861,
      "tokens_seen": 1696202752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908926780341023,
      "loss": 1.4315,
      "theoretical_loss": 3.5070756871520783,
      "tokens_seen": 1696268288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908726178535607,
      "loss": 1.4136,
      "theoretical_loss": 3.50706411540557,
      "tokens_seen": 1696333824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490852557673019,
      "loss": 1.4009,
      "theoretical_loss": 3.5070525442312874,
      "tokens_seen": 1696399360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908324974924774,
      "loss": 1.3483,
      "theoretical_loss": 3.507040973629179,
      "tokens_seen": 1696464896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908124373119358,
      "loss": 1.5329,
      "theoretical_loss": 3.507029403599195,
      "tokens_seen": 1696530432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907923771313941,
      "loss": 1.3032,
      "theoretical_loss": 3.5070178341412843,
      "tokens_seen": 1696595968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907723169508526,
      "loss": 1.393,
      "theoretical_loss": 3.5070062652553977,
      "tokens_seen": 1696661504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490752256770311,
      "loss": 1.2946,
      "theoretical_loss": 3.506994696941483,
      "tokens_seen": 1696727040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907321965897693,
      "loss": 1.2365,
      "theoretical_loss": 3.5069831291994915,
      "tokens_seen": 1696792576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907121364092277,
      "loss": 1.3889,
      "theoretical_loss": 3.506971562029372,
      "tokens_seen": 1696858112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490692076228686,
      "loss": 1.3606,
      "theoretical_loss": 3.5069599954310746,
      "tokens_seen": 1696923648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906720160481444,
      "loss": 1.2779,
      "theoretical_loss": 3.5069484294045483,
      "tokens_seen": 1696989184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906519558676028,
      "loss": 1.2498,
      "theoretical_loss": 3.5069368639497434,
      "tokens_seen": 1697054720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906318956870611,
      "loss": 1.3094,
      "theoretical_loss": 3.5069252990666095,
      "tokens_seen": 1697120256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906118355065195,
      "loss": 1.5626,
      "theoretical_loss": 3.506913734755096,
      "tokens_seen": 1697185792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490591775325978,
      "loss": 1.3086,
      "theoretical_loss": 3.5069021710151524,
      "tokens_seen": 1697251328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905717151454363,
      "loss": 1.4392,
      "theoretical_loss": 3.506890607846729,
      "tokens_seen": 1697316864
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905516549648947,
      "loss": 1.4032,
      "theoretical_loss": 3.506879045249775,
      "tokens_seen": 1697382400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905315947843531,
      "loss": 1.0952,
      "theoretical_loss": 3.5068674832242404,
      "tokens_seen": 1697447936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905115346038114,
      "loss": 1.3024,
      "theoretical_loss": 3.5068559217700748,
      "tokens_seen": 1697513472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904914744232698,
      "loss": 1.4544,
      "theoretical_loss": 3.5068443608872277,
      "tokens_seen": 1697579008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904714142427281,
      "loss": 1.3557,
      "theoretical_loss": 3.5068328005756486,
      "tokens_seen": 1697644544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904513540621865,
      "loss": 1.3338,
      "theoretical_loss": 3.506821240835288,
      "tokens_seen": 1697710080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904312938816449,
      "loss": 1.4531,
      "theoretical_loss": 3.5068096816660956,
      "tokens_seen": 1697775616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904112337011033,
      "loss": 1.351,
      "theoretical_loss": 3.5067981230680196,
      "tokens_seen": 1697841152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903911735205617,
      "loss": 1.3425,
      "theoretical_loss": 3.506786565041012,
      "tokens_seen": 1697906688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903711133400201,
      "loss": 1.3297,
      "theoretical_loss": 3.506775007585021,
      "tokens_seen": 1697972224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903510531594784,
      "loss": 1.3589,
      "theoretical_loss": 3.5067634506999967,
      "tokens_seen": 1698037760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903309929789368,
      "loss": 1.2881,
      "theoretical_loss": 3.5067518943858893,
      "tokens_seen": 1698103296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903109327983952,
      "loss": 1.4676,
      "theoretical_loss": 3.506740338642648,
      "tokens_seen": 1698168832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902908726178535,
      "loss": 1.3875,
      "theoretical_loss": 3.5067287834702223,
      "tokens_seen": 1698234368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902708124373119,
      "loss": 1.4076,
      "theoretical_loss": 3.506717228868563,
      "tokens_seen": 1698299904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902507522567703,
      "loss": 1.367,
      "theoretical_loss": 3.506705674837619,
      "tokens_seen": 1698365440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902306920762287,
      "loss": 1.5286,
      "theoretical_loss": 3.5066941213773406,
      "tokens_seen": 1698430976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902106318956871,
      "loss": 1.4007,
      "theoretical_loss": 3.5066825684876775,
      "tokens_seen": 1698496512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901905717151454,
      "loss": 1.2717,
      "theoretical_loss": 3.5066710161685792,
      "tokens_seen": 1698562048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901705115346038,
      "loss": 1.4175,
      "theoretical_loss": 3.5066594644199958,
      "tokens_seen": 1698627584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901504513540623,
      "loss": 1.4019,
      "theoretical_loss": 3.506647913241877,
      "tokens_seen": 1698693120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901303911735206,
      "loss": 1.4839,
      "theoretical_loss": 3.5066363626341728,
      "tokens_seen": 1698758656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490110330992979,
      "loss": 1.3258,
      "theoretical_loss": 3.506624812596833,
      "tokens_seen": 1698824192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900902708124374,
      "loss": 1.2434,
      "theoretical_loss": 3.5066132631298075,
      "tokens_seen": 1698889728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900702106318957,
      "loss": 1.3587,
      "theoretical_loss": 3.506601714233045,
      "tokens_seen": 1698955264
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900501504513541,
      "loss": 1.3124,
      "theoretical_loss": 3.5065901659064975,
      "tokens_seen": 1699020800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900300902708124,
      "loss": 1.423,
      "theoretical_loss": 3.5065786181501126,
      "tokens_seen": 1699086336
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 620634,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3232345581054688,
      "objective/train/theoretical_loss": 3.5065670709638423,
      "objective/train/tokens_used": 69610976,
      "theoretical_loss": 3.5065670709638423,
      "tokens_seen": 1699151872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900100300902708,
      "loss": 1.3566,
      "theoretical_loss": 3.5065670709638423,
      "tokens_seen": 1699151872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899899699097292,
      "loss": 1.2686,
      "theoretical_loss": 3.5065555243476347,
      "tokens_seen": 1699217408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899699097291876,
      "loss": 1.2973,
      "theoretical_loss": 3.5065439783014405,
      "tokens_seen": 1699282944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489949849548646,
      "loss": 1.2781,
      "theoretical_loss": 3.5065324328252094,
      "tokens_seen": 1699348480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899297893681044,
      "loss": 1.3311,
      "theoretical_loss": 3.5065208879188914,
      "tokens_seen": 1699414016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899097291875627,
      "loss": 1.3235,
      "theoretical_loss": 3.506509343582436,
      "tokens_seen": 1699479552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898896690070211,
      "loss": 1.3341,
      "theoretical_loss": 3.506497799815794,
      "tokens_seen": 1699545088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898696088264794,
      "loss": 1.2423,
      "theoretical_loss": 3.5064862566189143,
      "tokens_seen": 1699610624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898495486459378,
      "loss": 1.3397,
      "theoretical_loss": 3.5064747139917474,
      "tokens_seen": 1699676160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898294884653962,
      "loss": 1.2578,
      "theoretical_loss": 3.506463171934243,
      "tokens_seen": 1699741696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898094282848545,
      "loss": 1.3833,
      "theoretical_loss": 3.5064516304463513,
      "tokens_seen": 1699807232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489789368104313,
      "loss": 1.2578,
      "theoretical_loss": 3.506440089528022,
      "tokens_seen": 1699872768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897693079237714,
      "loss": 1.4603,
      "theoretical_loss": 3.5064285491792044,
      "tokens_seen": 1699938304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897492477432297,
      "loss": 1.2667,
      "theoretical_loss": 3.5064170093998497,
      "tokens_seen": 1700003840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897291875626881,
      "loss": 1.2995,
      "theoretical_loss": 3.506405470189907,
      "tokens_seen": 1700069376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897091273821465,
      "loss": 1.2886,
      "theoretical_loss": 3.5063939315493267,
      "tokens_seen": 1700134912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896890672016048,
      "loss": 1.3382,
      "theoretical_loss": 3.5063823934780585,
      "tokens_seen": 1700200448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896690070210632,
      "loss": 1.3299,
      "theoretical_loss": 3.506370855976052,
      "tokens_seen": 1700265984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896489468405215,
      "loss": 1.2978,
      "theoretical_loss": 3.5063593190432574,
      "tokens_seen": 1700331520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896288866599799,
      "loss": 1.2663,
      "theoretical_loss": 3.506347782679626,
      "tokens_seen": 1700397056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896088264794384,
      "loss": 1.2953,
      "theoretical_loss": 3.5063362468851054,
      "tokens_seen": 1700462592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895887662988967,
      "loss": 1.3638,
      "theoretical_loss": 3.5063247116596474,
      "tokens_seen": 1700528128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895687061183551,
      "loss": 1.2951,
      "theoretical_loss": 3.506313177003202,
      "tokens_seen": 1700593664
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895486459378135,
      "loss": 1.4755,
      "theoretical_loss": 3.5063016429157177,
      "tokens_seen": 1700659200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895285857572718,
      "loss": 1.4268,
      "theoretical_loss": 3.506290109397146,
      "tokens_seen": 1700724736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895085255767302,
      "loss": 1.296,
      "theoretical_loss": 3.5062785764474365,
      "tokens_seen": 1700790272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894884653961885,
      "loss": 1.3229,
      "theoretical_loss": 3.506267044066538,
      "tokens_seen": 1700855808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894684052156469,
      "loss": 1.2576,
      "theoretical_loss": 3.506255512254403,
      "tokens_seen": 1700921344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894483450351053,
      "loss": 1.3534,
      "theoretical_loss": 3.5062439810109796,
      "tokens_seen": 1700986880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894282848545637,
      "loss": 1.4438,
      "theoretical_loss": 3.5062324503362183,
      "tokens_seen": 1701052416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894082246740221,
      "loss": 1.5134,
      "theoretical_loss": 3.5062209202300694,
      "tokens_seen": 1701117952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893881644934805,
      "loss": 1.4263,
      "theoretical_loss": 3.506209390692483,
      "tokens_seen": 1701183488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893681043129388,
      "loss": 1.3592,
      "theoretical_loss": 3.5061978617234084,
      "tokens_seen": 1701249024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893480441323972,
      "loss": 1.418,
      "theoretical_loss": 3.5061863333227965,
      "tokens_seen": 1701314560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893279839518556,
      "loss": 1.3557,
      "theoretical_loss": 3.5061748054905975,
      "tokens_seen": 1701380096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893079237713139,
      "loss": 1.2816,
      "theoretical_loss": 3.5061632782267607,
      "tokens_seen": 1701445632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892878635907723,
      "loss": 1.3639,
      "theoretical_loss": 3.506151751531237,
      "tokens_seen": 1701511168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892678034102307,
      "loss": 1.4028,
      "theoretical_loss": 3.5061402254039757,
      "tokens_seen": 1701576704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892477432296891,
      "loss": 1.3536,
      "theoretical_loss": 3.5061286998449273,
      "tokens_seen": 1701642240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892276830491475,
      "loss": 1.3034,
      "theoretical_loss": 3.506117174854042,
      "tokens_seen": 1701707776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892076228686058,
      "loss": 1.2685,
      "theoretical_loss": 3.50610565043127,
      "tokens_seen": 1701773312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891875626880642,
      "loss": 1.3486,
      "theoretical_loss": 3.506094126576561,
      "tokens_seen": 1701838848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891675025075226,
      "loss": 1.3492,
      "theoretical_loss": 3.5060826032898653,
      "tokens_seen": 1701904384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891474423269809,
      "loss": 1.5114,
      "theoretical_loss": 3.5060710805711333,
      "tokens_seen": 1701969920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891273821464393,
      "loss": 1.3318,
      "theoretical_loss": 3.506059558420314,
      "tokens_seen": 1702035456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891073219658976,
      "loss": 1.3173,
      "theoretical_loss": 3.5060480368373597,
      "tokens_seen": 1702100992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890872617853561,
      "loss": 1.4413,
      "theoretical_loss": 3.5060365158222186,
      "tokens_seen": 1702166528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890672016048145,
      "loss": 1.3575,
      "theoretical_loss": 3.5060249953748417,
      "tokens_seen": 1702232064
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890471414242728,
      "loss": 1.2915,
      "theoretical_loss": 3.506013475495179,
      "tokens_seen": 1702297600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890270812437312,
      "loss": 1.345,
      "theoretical_loss": 3.506001956183181,
      "tokens_seen": 1702363136
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 621885,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3620576858520508,
      "objective/train/theoretical_loss": 3.5059904374387973,
      "objective/train/tokens_used": 72887776,
      "theoretical_loss": 3.5059904374387973,
      "tokens_seen": 1702428672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890070210631896,
      "loss": 1.3207,
      "theoretical_loss": 3.5059904374387973,
      "tokens_seen": 1702428672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889869608826479,
      "loss": 1.2116,
      "theoretical_loss": 3.5059789192619784,
      "tokens_seen": 1702494208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889669007021063,
      "loss": 1.3397,
      "theoretical_loss": 3.5059674016526747,
      "tokens_seen": 1702559744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889468405215647,
      "loss": 1.2122,
      "theoretical_loss": 3.5059558846108354,
      "tokens_seen": 1702625280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488926780341023,
      "loss": 1.4189,
      "theoretical_loss": 3.5059443681364124,
      "tokens_seen": 1702690816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889067201604815,
      "loss": 1.3155,
      "theoretical_loss": 3.5059328522293542,
      "tokens_seen": 1702756352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888866599799398,
      "loss": 1.4403,
      "theoretical_loss": 3.505921336889612,
      "tokens_seen": 1702821888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888665997993982,
      "loss": 1.4239,
      "theoretical_loss": 3.505909822117136,
      "tokens_seen": 1702887424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888465396188566,
      "loss": 1.3553,
      "theoretical_loss": 3.5058983079118757,
      "tokens_seen": 1702952960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888264794383149,
      "loss": 1.3793,
      "theoretical_loss": 3.505886794273782,
      "tokens_seen": 1703018496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888064192577733,
      "loss": 1.375,
      "theoretical_loss": 3.505875281202805,
      "tokens_seen": 1703084032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887863590772317,
      "loss": 1.3817,
      "theoretical_loss": 3.5058637686988954,
      "tokens_seen": 1703149568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048876629889669,
      "loss": 1.3402,
      "theoretical_loss": 3.5058522567620023,
      "tokens_seen": 1703215104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887462387161484,
      "loss": 1.3355,
      "theoretical_loss": 3.505840745392077,
      "tokens_seen": 1703280640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887261785356069,
      "loss": 1.3791,
      "theoretical_loss": 3.5058292345890694,
      "tokens_seen": 1703346176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887061183550652,
      "loss": 1.326,
      "theoretical_loss": 3.505817724352929,
      "tokens_seen": 1703411712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886860581745236,
      "loss": 1.292,
      "theoretical_loss": 3.505806214683608,
      "tokens_seen": 1703477248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886659979939819,
      "loss": 1.4999,
      "theoretical_loss": 3.5057947055810548,
      "tokens_seen": 1703542784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886459378134403,
      "loss": 1.4128,
      "theoretical_loss": 3.5057831970452207,
      "tokens_seen": 1703608320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886258776328988,
      "loss": 1.3775,
      "theoretical_loss": 3.5057716890760555,
      "tokens_seen": 1703673856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886058174523571,
      "loss": 1.4581,
      "theoretical_loss": 3.5057601816735096,
      "tokens_seen": 1703739392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885857572718155,
      "loss": 1.2917,
      "theoretical_loss": 3.505748674837534,
      "tokens_seen": 1703804928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885656970912739,
      "loss": 1.3024,
      "theoretical_loss": 3.5057371685680776,
      "tokens_seen": 1703870464
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885456369107322,
      "loss": 1.2958,
      "theoretical_loss": 3.505725662865092,
      "tokens_seen": 1703936000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885255767301906,
      "loss": 1.3506,
      "theoretical_loss": 3.5057141577285273,
      "tokens_seen": 1704001536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488505516549649,
      "loss": 1.2579,
      "theoretical_loss": 3.505702653158333,
      "tokens_seen": 1704067072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884854563691073,
      "loss": 1.3549,
      "theoretical_loss": 3.5056911491544605,
      "tokens_seen": 1704132608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884653961885657,
      "loss": 1.3368,
      "theoretical_loss": 3.5056796457168593,
      "tokens_seen": 1704198144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884453360080241,
      "loss": 1.4262,
      "theoretical_loss": 3.5056681428454803,
      "tokens_seen": 1704263680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884252758274825,
      "loss": 1.4651,
      "theoretical_loss": 3.5056566405402743,
      "tokens_seen": 1704329216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884052156469409,
      "loss": 1.4421,
      "theoretical_loss": 3.5056451388011904,
      "tokens_seen": 1704394752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883851554663992,
      "loss": 1.2227,
      "theoretical_loss": 3.5056336376281796,
      "tokens_seen": 1704460288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883650952858576,
      "loss": 1.4228,
      "theoretical_loss": 3.5056221370211924,
      "tokens_seen": 1704525824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488345035105316,
      "loss": 1.2801,
      "theoretical_loss": 3.505610636980179,
      "tokens_seen": 1704591360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883249749247743,
      "loss": 1.4061,
      "theoretical_loss": 3.50559913750509,
      "tokens_seen": 1704656896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883049147442327,
      "loss": 1.4225,
      "theoretical_loss": 3.505587638595876,
      "tokens_seen": 1704722432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882848545636911,
      "loss": 1.4614,
      "theoretical_loss": 3.5055761402524865,
      "tokens_seen": 1704787968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048826479438314946,
      "loss": 1.3331,
      "theoretical_loss": 3.5055646424748725,
      "tokens_seen": 1704853504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882447342026078,
      "loss": 1.3728,
      "theoretical_loss": 3.5055531452629847,
      "tokens_seen": 1704919040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048822467402206624,
      "loss": 1.3744,
      "theoretical_loss": 3.505541648616773,
      "tokens_seen": 1704984576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882046138415246,
      "loss": 1.4004,
      "theoretical_loss": 3.5055301525361884,
      "tokens_seen": 1705050112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048818455366098296,
      "loss": 1.3302,
      "theoretical_loss": 3.5055186570211805,
      "tokens_seen": 1705115648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881644934804413,
      "loss": 1.3399,
      "theoretical_loss": 3.5055071620717007,
      "tokens_seen": 1705181184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048814443329989974,
      "loss": 1.439,
      "theoretical_loss": 3.5054956676876987,
      "tokens_seen": 1705246720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881243731193581,
      "loss": 1.4232,
      "theoretical_loss": 3.505484173869125,
      "tokens_seen": 1705312256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048810431293881646,
      "loss": 1.3036,
      "theoretical_loss": 3.5054726806159304,
      "tokens_seen": 1705377792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880842527582748,
      "loss": 1.4074,
      "theoretical_loss": 3.505461187928065,
      "tokens_seen": 1705443328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880641925777332,
      "loss": 1.3335,
      "theoretical_loss": 3.50544969580548,
      "tokens_seen": 1705508864
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880441323971916,
      "loss": 1.3578,
      "theoretical_loss": 3.505438204248125,
      "tokens_seen": 1705574400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048802407221664997,
      "loss": 1.2749,
      "theoretical_loss": 3.505426713255951,
      "tokens_seen": 1705639936
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 623014,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2867580652236938,
      "objective/train/theoretical_loss": 3.5054152228289084,
      "objective/train/tokens_used": 76164576,
      "theoretical_loss": 3.5054152228289084,
      "tokens_seen": 1705705472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048800401203610833,
      "loss": 1.408,
      "theoretical_loss": 3.5054152228289084,
      "tokens_seen": 1705705472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879839518555667,
      "loss": 1.3039,
      "theoretical_loss": 3.5054037329669474,
      "tokens_seen": 1705771008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879638916750251,
      "loss": 1.4341,
      "theoretical_loss": 3.505392243670019,
      "tokens_seen": 1705836544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048794383149448347,
      "loss": 1.4318,
      "theoretical_loss": 3.5053807549380736,
      "tokens_seen": 1705902080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048792377131394183,
      "loss": 1.3985,
      "theoretical_loss": 3.5053692667710608,
      "tokens_seen": 1705967616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879037111334002,
      "loss": 1.4275,
      "theoretical_loss": 3.5053577791689325,
      "tokens_seen": 1706033152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048788365095285856,
      "loss": 1.2929,
      "theoretical_loss": 3.5053462921316383,
      "tokens_seen": 1706098688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048786359077231697,
      "loss": 1.2899,
      "theoretical_loss": 3.505334805659129,
      "tokens_seen": 1706164224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048784353059177533,
      "loss": 1.3528,
      "theoretical_loss": 3.505323319751356,
      "tokens_seen": 1706229760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878234704112337,
      "loss": 1.3581,
      "theoretical_loss": 3.505311834408268,
      "tokens_seen": 1706295296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048780341023069206,
      "loss": 1.3595,
      "theoretical_loss": 3.505300349629817,
      "tokens_seen": 1706360832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877833500501505,
      "loss": 1.1994,
      "theoretical_loss": 3.5052888654159533,
      "tokens_seen": 1706426368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048776328986960884,
      "loss": 1.3109,
      "theoretical_loss": 3.5052773817666267,
      "tokens_seen": 1706491904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877432296890672,
      "loss": 1.379,
      "theoretical_loss": 3.505265898681789,
      "tokens_seen": 1706557440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048772316950852556,
      "loss": 1.3817,
      "theoretical_loss": 3.5052544161613897,
      "tokens_seen": 1706622976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877031093279839,
      "loss": 1.4001,
      "theoretical_loss": 3.50524293420538,
      "tokens_seen": 1706688512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048768304914744234,
      "loss": 1.4308,
      "theoretical_loss": 3.5052314528137103,
      "tokens_seen": 1706754048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876629889669007,
      "loss": 1.4139,
      "theoretical_loss": 3.5052199719863317,
      "tokens_seen": 1706819584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048764292878635907,
      "loss": 1.4029,
      "theoretical_loss": 3.5052084917231934,
      "tokens_seen": 1706885120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048762286860581743,
      "loss": 1.3839,
      "theoretical_loss": 3.5051970120242473,
      "tokens_seen": 1706950656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048760280842527584,
      "loss": 1.3622,
      "theoretical_loss": 3.5051855328894437,
      "tokens_seen": 1707016192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875827482447342,
      "loss": 1.1993,
      "theoretical_loss": 3.5051740543187333,
      "tokens_seen": 1707081728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048756268806419257,
      "loss": 1.3528,
      "theoretical_loss": 3.5051625763120664,
      "tokens_seen": 1707147264
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048754262788365093,
      "loss": 1.2254,
      "theoretical_loss": 3.5051510988693937,
      "tokens_seen": 1707212800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048752256770310935,
      "loss": 1.3726,
      "theoretical_loss": 3.505139621990666,
      "tokens_seen": 1707278336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875025075225677,
      "loss": 1.2674,
      "theoretical_loss": 3.505128145675834,
      "tokens_seen": 1707343872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048748244734202607,
      "loss": 1.4298,
      "theoretical_loss": 3.505116669924848,
      "tokens_seen": 1707409408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048746238716148443,
      "loss": 1.3453,
      "theoretical_loss": 3.5051051947376592,
      "tokens_seen": 1707474944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874423269809428,
      "loss": 1.3674,
      "theoretical_loss": 3.5050937201142176,
      "tokens_seen": 1707540480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874222668004012,
      "loss": 1.4249,
      "theoretical_loss": 3.5050822460544744,
      "tokens_seen": 1707606016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874022066198596,
      "loss": 1.3695,
      "theoretical_loss": 3.5050707725583803,
      "tokens_seen": 1707671552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048738214643931794,
      "loss": 1.2728,
      "theoretical_loss": 3.5050592996258856,
      "tokens_seen": 1707737088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873620862587763,
      "loss": 1.2962,
      "theoretical_loss": 3.5050478272569414,
      "tokens_seen": 1707802624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873420260782347,
      "loss": 1.3047,
      "theoretical_loss": 3.505036355451498,
      "tokens_seen": 1707868160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873219658976931,
      "loss": 1.4268,
      "theoretical_loss": 3.505024884209506,
      "tokens_seen": 1707933696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048730190571715144,
      "loss": 1.2318,
      "theoretical_loss": 3.5050134135309166,
      "tokens_seen": 1707999232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872818455366098,
      "loss": 1.332,
      "theoretical_loss": 3.50500194341568,
      "tokens_seen": 1708064768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048726178535606816,
      "loss": 1.3295,
      "theoretical_loss": 3.5049904738637476,
      "tokens_seen": 1708130304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872417251755266,
      "loss": 1.2741,
      "theoretical_loss": 3.5049790048750697,
      "tokens_seen": 1708195840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000487221664994985,
      "loss": 1.4285,
      "theoretical_loss": 3.504967536449597,
      "tokens_seen": 1708261376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048720160481444336,
      "loss": 1.3439,
      "theoretical_loss": 3.50495606858728,
      "tokens_seen": 1708326912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871815446339017,
      "loss": 1.3742,
      "theoretical_loss": 3.50494460128807,
      "tokens_seen": 1708392448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048716148445336014,
      "loss": 1.4036,
      "theoretical_loss": 3.5049331345519175,
      "tokens_seen": 1708457984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871414242728185,
      "loss": 1.2504,
      "theoretical_loss": 3.504921668378773,
      "tokens_seen": 1708523520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048712136409227686,
      "loss": 1.3234,
      "theoretical_loss": 3.504910202768588,
      "tokens_seen": 1708589056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871013039117352,
      "loss": 1.2954,
      "theoretical_loss": 3.5048987377213123,
      "tokens_seen": 1708654592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870812437311936,
      "loss": 1.3709,
      "theoretical_loss": 3.5048872732368976,
      "tokens_seen": 1708720128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000487061183550652,
      "loss": 1.3592,
      "theoretical_loss": 3.504875809315294,
      "tokens_seen": 1708785664
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048704112337011037,
      "loss": 1.3201,
      "theoretical_loss": 3.5048643459564524,
      "tokens_seen": 1708851200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048702106318956873,
      "loss": 1.2678,
      "theoretical_loss": 3.5048528831603236,
      "tokens_seen": 1708916736
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 624223,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4896975755691528,
      "objective/train/theoretical_loss": 3.5048414209268586,
      "objective/train/tokens_used": 79441376,
      "theoretical_loss": 3.5048414209268586,
      "tokens_seen": 1708982272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870010030090271,
      "loss": 1.3843,
      "theoretical_loss": 3.5048414209268586,
      "tokens_seen": 1708982272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869809428284855,
      "loss": 1.331,
      "theoretical_loss": 3.504829959256009,
      "tokens_seen": 1709047808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048696088264794387,
      "loss": 1.227,
      "theoretical_loss": 3.5048184981477233,
      "tokens_seen": 1709113344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048694082246740223,
      "loss": 1.3152,
      "theoretical_loss": 3.504807037601955,
      "tokens_seen": 1709178880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869207622868606,
      "loss": 1.3221,
      "theoretical_loss": 3.504795577618653,
      "tokens_seen": 1709244416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048690070210631896,
      "loss": 1.2703,
      "theoretical_loss": 3.504784118197769,
      "tokens_seen": 1709309952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048688064192577737,
      "loss": 1.3755,
      "theoretical_loss": 3.5047726593392534,
      "tokens_seen": 1709375488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048686058174523573,
      "loss": 1.3504,
      "theoretical_loss": 3.504761201043057,
      "tokens_seen": 1709441024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868405215646941,
      "loss": 1.3398,
      "theoretical_loss": 3.5047497433091315,
      "tokens_seen": 1709506560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048682046138415246,
      "loss": 1.2513,
      "theoretical_loss": 3.504738286137427,
      "tokens_seen": 1709572096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868004012036109,
      "loss": 1.3342,
      "theoretical_loss": 3.5047268295278946,
      "tokens_seen": 1709637632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048678034102306924,
      "loss": 1.444,
      "theoretical_loss": 3.504715373480485,
      "tokens_seen": 1709703168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867602808425276,
      "loss": 1.3006,
      "theoretical_loss": 3.5047039179951494,
      "tokens_seen": 1709768704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048674022066198596,
      "loss": 1.4039,
      "theoretical_loss": 3.504692463071838,
      "tokens_seen": 1709834240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867201604814443,
      "loss": 1.3323,
      "theoretical_loss": 3.5046810087105027,
      "tokens_seen": 1709899776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048670010030090274,
      "loss": 1.3463,
      "theoretical_loss": 3.5046695549110938,
      "tokens_seen": 1709965312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866800401203611,
      "loss": 1.3981,
      "theoretical_loss": 3.504658101673562,
      "tokens_seen": 1710030848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048665997993981947,
      "loss": 1.4365,
      "theoretical_loss": 3.5046466489978583,
      "tokens_seen": 1710096384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048663991975927783,
      "loss": 1.4214,
      "theoretical_loss": 3.504635196883934,
      "tokens_seen": 1710161920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048661985957873624,
      "loss": 1.3643,
      "theoretical_loss": 3.50462374533174,
      "tokens_seen": 1710227456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865997993981946,
      "loss": 1.3654,
      "theoretical_loss": 3.504612294341227,
      "tokens_seen": 1710292992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048657973921765297,
      "loss": 1.3913,
      "theoretical_loss": 3.5046008439123453,
      "tokens_seen": 1710358528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048655967903711133,
      "loss": 1.408,
      "theoretical_loss": 3.5045893940450465,
      "tokens_seen": 1710424064
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048653961885656975,
      "loss": 1.383,
      "theoretical_loss": 3.504577944739282,
      "tokens_seen": 1710489600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865195586760281,
      "loss": 1.3865,
      "theoretical_loss": 3.504566495995002,
      "tokens_seen": 1710555136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048649949849548647,
      "loss": 1.3072,
      "theoretical_loss": 3.504555047812158,
      "tokens_seen": 1710620672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048647943831494483,
      "loss": 1.4701,
      "theoretical_loss": 3.5045436001907007,
      "tokens_seen": 1710686208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864593781344032,
      "loss": 1.4357,
      "theoretical_loss": 3.5045321531305804,
      "tokens_seen": 1710751744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864393179538616,
      "loss": 1.3332,
      "theoretical_loss": 3.5045207066317494,
      "tokens_seen": 1710817280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048641925777332,
      "loss": 1.3726,
      "theoretical_loss": 3.5045092606941575,
      "tokens_seen": 1710882816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048639919759277834,
      "loss": 1.2636,
      "theoretical_loss": 3.5044978153177566,
      "tokens_seen": 1710948352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863791374122367,
      "loss": 1.2637,
      "theoretical_loss": 3.504486370502497,
      "tokens_seen": 1711013888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863590772316951,
      "loss": 1.5042,
      "theoretical_loss": 3.5044749262483297,
      "tokens_seen": 1711079424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863390170511535,
      "loss": 1.3395,
      "theoretical_loss": 3.504463482555207,
      "tokens_seen": 1711144960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048631895687061184,
      "loss": 1.2796,
      "theoretical_loss": 3.504452039423078,
      "tokens_seen": 1711210496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862988966900702,
      "loss": 1.185,
      "theoretical_loss": 3.5044405968518944,
      "tokens_seen": 1711276032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048627883650952857,
      "loss": 1.3957,
      "theoretical_loss": 3.5044291548416076,
      "tokens_seen": 1711341568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486258776328987,
      "loss": 1.3358,
      "theoretical_loss": 3.5044177133921686,
      "tokens_seen": 1711407104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048623871614844534,
      "loss": 1.3734,
      "theoretical_loss": 3.504406272503528,
      "tokens_seen": 1711472640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862186559679037,
      "loss": 1.3995,
      "theoretical_loss": 3.5043948321756373,
      "tokens_seen": 1711538176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048619859578736207,
      "loss": 1.2497,
      "theoretical_loss": 3.5043833924084473,
      "tokens_seen": 1711603712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861785356068205,
      "loss": 1.4278,
      "theoretical_loss": 3.5043719532019093,
      "tokens_seen": 1711669248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048615847542627885,
      "loss": 1.3903,
      "theoretical_loss": 3.504360514555974,
      "tokens_seen": 1711734784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861384152457372,
      "loss": 1.3357,
      "theoretical_loss": 3.504349076470593,
      "tokens_seen": 1711800320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048611835506519557,
      "loss": 1.4167,
      "theoretical_loss": 3.5043376389457164,
      "tokens_seen": 1711865856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048609829488465393,
      "loss": 1.4046,
      "theoretical_loss": 3.504326201981296,
      "tokens_seen": 1711931392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048607823470411235,
      "loss": 1.4093,
      "theoretical_loss": 3.504314765577283,
      "tokens_seen": 1711996928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860581745235707,
      "loss": 1.405,
      "theoretical_loss": 3.504303329733628,
      "tokens_seen": 1712062464
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860381143430291,
      "loss": 1.3021,
      "theoretical_loss": 3.5042918944502826,
      "tokens_seen": 1712128000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048601805416248744,
      "loss": 1.3389,
      "theoretical_loss": 3.5042804597271973,
      "tokens_seen": 1712193536
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 625444,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4292044639587402,
      "objective/train/theoretical_loss": 3.504269025564324,
      "objective/train/tokens_used": 82718176,
      "theoretical_loss": 3.504269025564324,
      "tokens_seen": 1712259072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048599799398194585,
      "loss": 1.3556,
      "theoretical_loss": 3.504269025564324,
      "tokens_seen": 1712259072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859779338014042,
      "loss": 1.3043,
      "theoretical_loss": 3.504257591961613,
      "tokens_seen": 1712324608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859578736208626,
      "loss": 1.3576,
      "theoretical_loss": 3.504246158919016,
      "tokens_seen": 1712390144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048593781344032094,
      "loss": 1.2208,
      "theoretical_loss": 3.5042347264364837,
      "tokens_seen": 1712455680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859177532597793,
      "loss": 1.4373,
      "theoretical_loss": 3.5042232945139675,
      "tokens_seen": 1712521216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858976930792377,
      "loss": 1.217,
      "theoretical_loss": 3.5042118631514185,
      "tokens_seen": 1712586752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858776328986961,
      "loss": 1.3577,
      "theoretical_loss": 3.504200432348788,
      "tokens_seen": 1712652288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048585757271815444,
      "loss": 1.3441,
      "theoretical_loss": 3.5041890021060267,
      "tokens_seen": 1712717824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858375125376128,
      "loss": 1.3392,
      "theoretical_loss": 3.504177572423086,
      "tokens_seen": 1712783360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858174523570712,
      "loss": 1.2145,
      "theoretical_loss": 3.5041661432999174,
      "tokens_seen": 1712848896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857973921765296,
      "loss": 1.4326,
      "theoretical_loss": 3.504154714736472,
      "tokens_seen": 1712914432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048577733199598795,
      "loss": 1.2853,
      "theoretical_loss": 3.5041432867327007,
      "tokens_seen": 1712979968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857572718154463,
      "loss": 1.3071,
      "theoretical_loss": 3.504131859288554,
      "tokens_seen": 1713045504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048573721163490467,
      "loss": 1.3519,
      "theoretical_loss": 3.504120432403984,
      "tokens_seen": 1713111040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048571715145436314,
      "loss": 1.3422,
      "theoretical_loss": 3.5041090060789424,
      "tokens_seen": 1713176576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856970912738215,
      "loss": 1.3945,
      "theoretical_loss": 3.504097580313379,
      "tokens_seen": 1713242112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048567703109327987,
      "loss": 1.4079,
      "theoretical_loss": 3.504086155107246,
      "tokens_seen": 1713307648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048565697091273823,
      "loss": 1.3968,
      "theoretical_loss": 3.504074730460494,
      "tokens_seen": 1713373184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048563691073219664,
      "loss": 1.3522,
      "theoretical_loss": 3.504063306373075,
      "tokens_seen": 1713438720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485616850551655,
      "loss": 1.3877,
      "theoretical_loss": 3.50405188284494,
      "tokens_seen": 1713504256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048559679037111337,
      "loss": 1.222,
      "theoretical_loss": 3.5040404598760393,
      "tokens_seen": 1713569792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048557673019057173,
      "loss": 1.3387,
      "theoretical_loss": 3.504029037466325,
      "tokens_seen": 1713635328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048555667001003015,
      "loss": 1.3555,
      "theoretical_loss": 3.504017615615748,
      "tokens_seen": 1713700864
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855366098294885,
      "loss": 1.3985,
      "theoretical_loss": 3.50400619432426,
      "tokens_seen": 1713766400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048551654964894687,
      "loss": 1.357,
      "theoretical_loss": 3.5039947735918124,
      "tokens_seen": 1713831936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048549648946840523,
      "loss": 1.3677,
      "theoretical_loss": 3.5039833534183553,
      "tokens_seen": 1713897472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854764292878636,
      "loss": 1.4683,
      "theoretical_loss": 3.5039719338038413,
      "tokens_seen": 1713963008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485456369107322,
      "loss": 1.3917,
      "theoretical_loss": 3.5039605147482202,
      "tokens_seen": 1714028544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854363089267804,
      "loss": 1.4392,
      "theoretical_loss": 3.503949096251445,
      "tokens_seen": 1714094080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048541624874623874,
      "loss": 1.2508,
      "theoretical_loss": 3.5039376783134655,
      "tokens_seen": 1714159616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853961885656971,
      "loss": 1.3869,
      "theoretical_loss": 3.503926260934234,
      "tokens_seen": 1714225152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853761283851555,
      "loss": 1.3878,
      "theoretical_loss": 3.5039148441137016,
      "tokens_seen": 1714290688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853560682046139,
      "loss": 1.3385,
      "theoretical_loss": 3.503903427851819,
      "tokens_seen": 1714356224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048533600802407224,
      "loss": 1.3191,
      "theoretical_loss": 3.5038920121485386,
      "tokens_seen": 1714421760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853159478435306,
      "loss": 1.3496,
      "theoretical_loss": 3.5038805970038105,
      "tokens_seen": 1714487296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048529588766298897,
      "loss": 1.3658,
      "theoretical_loss": 3.5038691824175867,
      "tokens_seen": 1714552832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852758274824474,
      "loss": 1.3793,
      "theoretical_loss": 3.5038577683898184,
      "tokens_seen": 1714618368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048525576730190574,
      "loss": 1.4162,
      "theoretical_loss": 3.503846354920457,
      "tokens_seen": 1714683904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852357071213641,
      "loss": 1.3681,
      "theoretical_loss": 3.503834942009454,
      "tokens_seen": 1714749440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048521564694082247,
      "loss": 1.5114,
      "theoretical_loss": 3.50382352965676,
      "tokens_seen": 1714814976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851955867602809,
      "loss": 1.3666,
      "theoretical_loss": 3.5038121178623274,
      "tokens_seen": 1714880512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048517552657973925,
      "loss": 1.2265,
      "theoretical_loss": 3.503800706626107,
      "tokens_seen": 1714946048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851554663991976,
      "loss": 1.356,
      "theoretical_loss": 3.5037892959480494,
      "tokens_seen": 1715011584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048513540621865597,
      "loss": 1.3333,
      "theoretical_loss": 3.5037778858281077,
      "tokens_seen": 1715077120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048511534603811433,
      "loss": 1.3985,
      "theoretical_loss": 3.503766476266232,
      "tokens_seen": 1715142656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048509528585757275,
      "loss": 1.4035,
      "theoretical_loss": 3.503755067262374,
      "tokens_seen": 1715208192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850752256770311,
      "loss": 1.3749,
      "theoretical_loss": 3.503743658816485,
      "tokens_seen": 1715273728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850551654964895,
      "loss": 1.4813,
      "theoretical_loss": 3.503732250928517,
      "tokens_seen": 1715339264
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048503510531594784,
      "loss": 1.3899,
      "theoretical_loss": 3.5037208435984204,
      "tokens_seen": 1715404800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048501504513540625,
      "loss": 1.3947,
      "theoretical_loss": 3.5037094368261474,
      "tokens_seen": 1715470336
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 626732,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3518413305282593,
      "objective/train/theoretical_loss": 3.503698030611649,
      "objective/train/tokens_used": 85994976,
      "theoretical_loss": 3.503698030611649,
      "tokens_seen": 1715535872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849949849548646,
      "loss": 1.2875,
      "theoretical_loss": 3.503698030611649,
      "tokens_seen": 1715535872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484974924774323,
      "loss": 1.3052,
      "theoretical_loss": 3.503686624954877,
      "tokens_seen": 1715601408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048495486459378134,
      "loss": 1.3779,
      "theoretical_loss": 3.503675219855782,
      "tokens_seen": 1715666944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849348044132397,
      "loss": 1.3558,
      "theoretical_loss": 3.503663815314317,
      "tokens_seen": 1715732480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849147442326981,
      "loss": 1.4397,
      "theoretical_loss": 3.5036524113304317,
      "tokens_seen": 1715798016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848946840521565,
      "loss": 1.3833,
      "theoretical_loss": 3.5036410079040783,
      "tokens_seen": 1715863552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048487462387161484,
      "loss": 1.4652,
      "theoretical_loss": 3.5036296050352087,
      "tokens_seen": 1715929088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848545636910732,
      "loss": 1.4268,
      "theoretical_loss": 3.5036182027237732,
      "tokens_seen": 1715994624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848345035105316,
      "loss": 1.3413,
      "theoretical_loss": 3.5036068009697248,
      "tokens_seen": 1716060160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048481444332999,
      "loss": 1.518,
      "theoretical_loss": 3.5035953997730136,
      "tokens_seen": 1716125696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048479438314944835,
      "loss": 1.4168,
      "theoretical_loss": 3.5035839991335918,
      "tokens_seen": 1716191232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847743229689067,
      "loss": 1.4196,
      "theoretical_loss": 3.5035725990514104,
      "tokens_seen": 1716256768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048475426278836507,
      "loss": 1.2809,
      "theoretical_loss": 3.5035611995264215,
      "tokens_seen": 1716322304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847342026078235,
      "loss": 1.3607,
      "theoretical_loss": 3.503549800558576,
      "tokens_seen": 1716387840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048471414242728185,
      "loss": 1.3481,
      "theoretical_loss": 3.503538402147826,
      "tokens_seen": 1716453376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846940822467402,
      "loss": 1.3558,
      "theoretical_loss": 3.5035270042941224,
      "tokens_seen": 1716518912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846740220661986,
      "loss": 1.3838,
      "theoretical_loss": 3.503515606997417,
      "tokens_seen": 1716584448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484653961885657,
      "loss": 1.4044,
      "theoretical_loss": 3.5035042102576615,
      "tokens_seen": 1716649984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048463390170511535,
      "loss": 1.4107,
      "theoretical_loss": 3.5034928140748067,
      "tokens_seen": 1716715520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846138415245737,
      "loss": 1.3627,
      "theoretical_loss": 3.5034814184488052,
      "tokens_seen": 1716781056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845937813440321,
      "loss": 1.3421,
      "theoretical_loss": 3.5034700233796077,
      "tokens_seen": 1716846592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048457372116349044,
      "loss": 1.4453,
      "theoretical_loss": 3.5034586288671656,
      "tokens_seen": 1716912128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048455366098294886,
      "loss": 1.4054,
      "theoretical_loss": 3.5034472349114316,
      "tokens_seen": 1716977664
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845336008024072,
      "loss": 1.34,
      "theoretical_loss": 3.5034358415123563,
      "tokens_seen": 1717043200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845135406218656,
      "loss": 1.3819,
      "theoretical_loss": 3.5034244486698913,
      "tokens_seen": 1717108736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048449348044132394,
      "loss": 1.405,
      "theoretical_loss": 3.5034130563839887,
      "tokens_seen": 1717174272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048447342026078236,
      "loss": 1.3497,
      "theoretical_loss": 3.5034016646545996,
      "tokens_seen": 1717239808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844533600802407,
      "loss": 1.3935,
      "theoretical_loss": 3.503390273481676,
      "tokens_seen": 1717305344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844332998996991,
      "loss": 1.4651,
      "theoretical_loss": 3.503378882865168,
      "tokens_seen": 1717370880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048441323971915745,
      "loss": 1.2457,
      "theoretical_loss": 3.5033674928050296,
      "tokens_seen": 1717436416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048439317953861586,
      "loss": 1.3766,
      "theoretical_loss": 3.5033561033012104,
      "tokens_seen": 1717501952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843731193580742,
      "loss": 1.3019,
      "theoretical_loss": 3.5033447143536636,
      "tokens_seen": 1717567488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843530591775326,
      "loss": 1.4416,
      "theoretical_loss": 3.5033333259623394,
      "tokens_seen": 1717633024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048433299899699095,
      "loss": 1.3507,
      "theoretical_loss": 3.50332193812719,
      "tokens_seen": 1717698560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843129388164493,
      "loss": 1.3759,
      "theoretical_loss": 3.503310550848167,
      "tokens_seen": 1717764096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048429287863590773,
      "loss": 1.4532,
      "theoretical_loss": 3.503299164125222,
      "tokens_seen": 1717829632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842728184553661,
      "loss": 1.401,
      "theoretical_loss": 3.503287777958307,
      "tokens_seen": 1717895168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048425275827482445,
      "loss": 1.3517,
      "theoretical_loss": 3.503276392347373,
      "tokens_seen": 1717960704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048423269809428287,
      "loss": 1.3692,
      "theoretical_loss": 3.503265007292372,
      "tokens_seen": 1718026240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842126379137413,
      "loss": 1.2823,
      "theoretical_loss": 3.5032536227932556,
      "tokens_seen": 1718091776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048419257773319965,
      "loss": 1.3413,
      "theoretical_loss": 3.5032422388499755,
      "tokens_seen": 1718157312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484172517552658,
      "loss": 1.3555,
      "theoretical_loss": 3.503230855462483,
      "tokens_seen": 1718222848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048415245737211637,
      "loss": 1.4239,
      "theoretical_loss": 3.5032194726307306,
      "tokens_seen": 1718288384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048413239719157473,
      "loss": 1.2772,
      "theoretical_loss": 3.503208090354669,
      "tokens_seen": 1718353920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048411233701103315,
      "loss": 1.3275,
      "theoretical_loss": 3.5031967086342504,
      "tokens_seen": 1718419456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840922768304915,
      "loss": 1.397,
      "theoretical_loss": 3.5031853274694265,
      "tokens_seen": 1718484992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840722166499499,
      "loss": 1.3658,
      "theoretical_loss": 3.503173946860149,
      "tokens_seen": 1718550528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048405215646940824,
      "loss": 1.2916,
      "theoretical_loss": 3.5031625668063695,
      "tokens_seen": 1718616064
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048403209628886665,
      "loss": 1.3655,
      "theoretical_loss": 3.5031511873080396,
      "tokens_seen": 1718681600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484012036108325,
      "loss": 1.3443,
      "theoretical_loss": 3.5031398083651113,
      "tokens_seen": 1718747136
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 627826,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4458614587783813,
      "objective/train/theoretical_loss": 3.503128429977536,
      "objective/train/tokens_used": 89271776,
      "theoretical_loss": 3.503128429977536,
      "tokens_seen": 1718812672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839919759277834,
      "loss": 1.4504,
      "theoretical_loss": 3.503128429977536,
      "tokens_seen": 1718812672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048397191574724174,
      "loss": 1.4615,
      "theoretical_loss": 3.5031170521452655,
      "tokens_seen": 1718878208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839518555667001,
      "loss": 1.3933,
      "theoretical_loss": 3.5031056748682516,
      "tokens_seen": 1718943744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839317953861585,
      "loss": 1.3026,
      "theoretical_loss": 3.503094298146446,
      "tokens_seen": 1719009280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839117352056169,
      "loss": 1.3633,
      "theoretical_loss": 3.503082921979801,
      "tokens_seen": 1719074816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048389167502507524,
      "loss": 1.2215,
      "theoretical_loss": 3.5030715463682673,
      "tokens_seen": 1719140352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838716148445336,
      "loss": 1.3471,
      "theoretical_loss": 3.5030601713117973,
      "tokens_seen": 1719205888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483851554663992,
      "loss": 1.3934,
      "theoretical_loss": 3.5030487968103428,
      "tokens_seen": 1719271424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838314944834504,
      "loss": 1.3297,
      "theoretical_loss": 3.503037422863855,
      "tokens_seen": 1719336960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048381143430290875,
      "loss": 1.4072,
      "theoretical_loss": 3.5030260494722865,
      "tokens_seen": 1719402496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837913741223671,
      "loss": 1.4298,
      "theoretical_loss": 3.5030146766355887,
      "tokens_seen": 1719468032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048377131394182547,
      "loss": 1.437,
      "theoretical_loss": 3.5030033043537125,
      "tokens_seen": 1719533568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837512537612839,
      "loss": 1.4109,
      "theoretical_loss": 3.5029919326266112,
      "tokens_seen": 1719599104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048373119358074225,
      "loss": 1.333,
      "theoretical_loss": 3.502980561454236,
      "tokens_seen": 1719664640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837111334002006,
      "loss": 1.3641,
      "theoretical_loss": 3.502969190836538,
      "tokens_seen": 1719730176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483691073219659,
      "loss": 1.4565,
      "theoretical_loss": 3.50295782077347,
      "tokens_seen": 1719795712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836710130391174,
      "loss": 1.428,
      "theoretical_loss": 3.5029464512649833,
      "tokens_seen": 1719861248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048365095285857575,
      "loss": 1.3461,
      "theoretical_loss": 3.50293508231103,
      "tokens_seen": 1719926784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836308926780341,
      "loss": 1.3766,
      "theoretical_loss": 3.502923713911562,
      "tokens_seen": 1719992320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836108324974925,
      "loss": 1.3903,
      "theoretical_loss": 3.5029123460665303,
      "tokens_seen": 1720057856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048359077231695084,
      "loss": 1.4031,
      "theoretical_loss": 3.502900978775888,
      "tokens_seen": 1720123392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048357071213640926,
      "loss": 1.3642,
      "theoretical_loss": 3.502889612039586,
      "tokens_seen": 1720188928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835506519558676,
      "loss": 1.3222,
      "theoretical_loss": 3.502878245857576,
      "tokens_seen": 1720254464
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483530591775326,
      "loss": 1.3895,
      "theoretical_loss": 3.502866880229811,
      "tokens_seen": 1720320000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048351053159478434,
      "loss": 1.3602,
      "theoretical_loss": 3.5028555151562415,
      "tokens_seen": 1720385536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048349047141424276,
      "loss": 1.3234,
      "theoretical_loss": 3.5028441506368204,
      "tokens_seen": 1720451072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834704112337011,
      "loss": 1.3973,
      "theoretical_loss": 3.502832786671499,
      "tokens_seen": 1720516608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834503510531595,
      "loss": 1.3449,
      "theoretical_loss": 3.5028214232602295,
      "tokens_seen": 1720582144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048343029087261785,
      "loss": 1.3186,
      "theoretical_loss": 3.5028100604029637,
      "tokens_seen": 1720647680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048341023069207626,
      "loss": 1.4168,
      "theoretical_loss": 3.5027986980996535,
      "tokens_seen": 1720713216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833901705115346,
      "loss": 1.4152,
      "theoretical_loss": 3.5027873363502504,
      "tokens_seen": 1720778752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483370110330993,
      "loss": 1.4194,
      "theoretical_loss": 3.502775975154707,
      "tokens_seen": 1720844288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048335005015045135,
      "loss": 1.3233,
      "theoretical_loss": 3.5027646145129747,
      "tokens_seen": 1720909824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833299899699097,
      "loss": 1.4071,
      "theoretical_loss": 3.5027532544250057,
      "tokens_seen": 1720975360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048330992978936813,
      "loss": 1.5229,
      "theoretical_loss": 3.502741894890751,
      "tokens_seen": 1721040896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832898696088265,
      "loss": 1.463,
      "theoretical_loss": 3.5027305359101644,
      "tokens_seen": 1721106432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048326980942828485,
      "loss": 1.4471,
      "theoretical_loss": 3.502719177483196,
      "tokens_seen": 1721171968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832497492477432,
      "loss": 1.4777,
      "theoretical_loss": 3.502707819609799,
      "tokens_seen": 1721237504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048322968906720163,
      "loss": 1.5016,
      "theoretical_loss": 3.502696462289925,
      "tokens_seen": 1721303040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048320962888666,
      "loss": 1.3798,
      "theoretical_loss": 3.502685105523525,
      "tokens_seen": 1721368576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048318956870611836,
      "loss": 1.3582,
      "theoretical_loss": 3.5026737493105524,
      "tokens_seen": 1721434112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831695085255767,
      "loss": 1.3654,
      "theoretical_loss": 3.502662393650958,
      "tokens_seen": 1721499648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831494483450351,
      "loss": 1.231,
      "theoretical_loss": 3.5026510385446947,
      "tokens_seen": 1721565184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831293881644935,
      "loss": 1.3885,
      "theoretical_loss": 3.502639683991714,
      "tokens_seen": 1721630720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048310932798395186,
      "loss": 1.4827,
      "theoretical_loss": 3.5026283299919676,
      "tokens_seen": 1721696256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830892678034102,
      "loss": 1.3953,
      "theoretical_loss": 3.5026169765454083,
      "tokens_seen": 1721761792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830692076228686,
      "loss": 1.2148,
      "theoretical_loss": 3.5026056236519874,
      "tokens_seen": 1721827328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483049147442327,
      "loss": 1.404,
      "theoretical_loss": 3.502594271311657,
      "tokens_seen": 1721892864
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048302908726178536,
      "loss": 1.3541,
      "theoretical_loss": 3.502582919524369,
      "tokens_seen": 1721958400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830090270812437,
      "loss": 1.3703,
      "theoretical_loss": 3.502571568290076,
      "tokens_seen": 1722023936
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 629016,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3245419263839722,
      "objective/train/theoretical_loss": 3.50256021760873,
      "objective/train/tokens_used": 92548576,
      "theoretical_loss": 3.50256021760873,
      "tokens_seen": 1722089472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829889669007021,
      "loss": 1.3607,
      "theoretical_loss": 3.50256021760873,
      "tokens_seen": 1722089472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048296890672016045,
      "loss": 1.2648,
      "theoretical_loss": 3.502548867480282,
      "tokens_seen": 1722155008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048294884653961887,
      "loss": 1.3391,
      "theoretical_loss": 3.5025375179046847,
      "tokens_seen": 1722220544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048292878635907723,
      "loss": 1.3666,
      "theoretical_loss": 3.5025261688818903,
      "tokens_seen": 1722286080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829087261785356,
      "loss": 1.4693,
      "theoretical_loss": 3.502514820411851,
      "tokens_seen": 1722351616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048288866599799395,
      "loss": 1.358,
      "theoretical_loss": 3.502503472494518,
      "tokens_seen": 1722417152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048286860581745237,
      "loss": 1.3005,
      "theoretical_loss": 3.5024921251298444,
      "tokens_seen": 1722482688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048284854563691073,
      "loss": 1.5898,
      "theoretical_loss": 3.502480778317781,
      "tokens_seen": 1722548224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828284854563691,
      "loss": 1.4709,
      "theoretical_loss": 3.502469432058281,
      "tokens_seen": 1722613760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048280842527582746,
      "loss": 1.3743,
      "theoretical_loss": 3.5024580863512966,
      "tokens_seen": 1722679296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827883650952858,
      "loss": 1.3715,
      "theoretical_loss": 3.5024467411967786,
      "tokens_seen": 1722744832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048276830491474423,
      "loss": 1.5312,
      "theoretical_loss": 3.50243539659468,
      "tokens_seen": 1722810368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827482447342026,
      "loss": 1.4025,
      "theoretical_loss": 3.502424052544953,
      "tokens_seen": 1722875904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482728184553661,
      "loss": 1.3609,
      "theoretical_loss": 3.502412709047549,
      "tokens_seen": 1722941440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827081243731194,
      "loss": 1.3602,
      "theoretical_loss": 3.502401366102421,
      "tokens_seen": 1723006976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826880641925778,
      "loss": 1.3103,
      "theoretical_loss": 3.5023900237095202,
      "tokens_seen": 1723072512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048266800401203615,
      "loss": 1.4195,
      "theoretical_loss": 3.502378681868799,
      "tokens_seen": 1723138048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826479438314945,
      "loss": 1.4357,
      "theoretical_loss": 3.50236734058021,
      "tokens_seen": 1723203584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826278836509529,
      "loss": 1.3346,
      "theoretical_loss": 3.5023559998437053,
      "tokens_seen": 1723269120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048260782347041124,
      "loss": 1.4596,
      "theoretical_loss": 3.5023446596592365,
      "tokens_seen": 1723334656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048258776328986966,
      "loss": 1.2937,
      "theoretical_loss": 3.5023333200267555,
      "tokens_seen": 1723400192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482567703109328,
      "loss": 1.2587,
      "theoretical_loss": 3.502321980946215,
      "tokens_seen": 1723465728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825476429287864,
      "loss": 1.4569,
      "theoretical_loss": 3.502310642417567,
      "tokens_seen": 1723531264
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048252758274824474,
      "loss": 1.4138,
      "theoretical_loss": 3.502299304440764,
      "tokens_seen": 1723596800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048250752256770316,
      "loss": 1.354,
      "theoretical_loss": 3.5022879670157576,
      "tokens_seen": 1723662336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824874623871615,
      "loss": 1.4462,
      "theoretical_loss": 3.5022766301425,
      "tokens_seen": 1723727872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824674022066199,
      "loss": 1.3097,
      "theoretical_loss": 3.502265293820944,
      "tokens_seen": 1723793408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048244734202607825,
      "loss": 1.3613,
      "theoretical_loss": 3.5022539580510417,
      "tokens_seen": 1723858944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048242728184553666,
      "loss": 1.4121,
      "theoretical_loss": 3.5022426228327443,
      "tokens_seen": 1723924480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482407221664995,
      "loss": 1.2609,
      "theoretical_loss": 3.502231288166005,
      "tokens_seen": 1723990016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823871614844534,
      "loss": 1.3218,
      "theoretical_loss": 3.5022199540507755,
      "tokens_seen": 1724055552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048236710130391175,
      "loss": 1.3748,
      "theoretical_loss": 3.502208620487008,
      "tokens_seen": 1724121088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823470411233701,
      "loss": 1.2851,
      "theoretical_loss": 3.5021972874746554,
      "tokens_seen": 1724186624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048232698094282853,
      "loss": 1.4481,
      "theoretical_loss": 3.502185955013669,
      "tokens_seen": 1724252160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004823069207622869,
      "loss": 1.4375,
      "theoretical_loss": 3.5021746231040014,
      "tokens_seen": 1724317696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048228686058174525,
      "loss": 1.3691,
      "theoretical_loss": 3.5021632917456045,
      "tokens_seen": 1724383232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822668004012036,
      "loss": 1.4047,
      "theoretical_loss": 3.502151960938431,
      "tokens_seen": 1724448768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048224674022066203,
      "loss": 1.3064,
      "theoretical_loss": 3.502140630682433,
      "tokens_seen": 1724514304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822266800401204,
      "loss": 1.3829,
      "theoretical_loss": 3.502129300977563,
      "tokens_seen": 1724579840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048220661985957876,
      "loss": 1.4479,
      "theoretical_loss": 3.5021179718237727,
      "tokens_seen": 1724645376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821865596790371,
      "loss": 1.3896,
      "theoretical_loss": 3.5021066432210146,
      "tokens_seen": 1724710912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821664994984955,
      "loss": 1.4652,
      "theoretical_loss": 3.5020953151692416,
      "tokens_seen": 1724776448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821464393179539,
      "loss": 1.3009,
      "theoretical_loss": 3.5020839876684047,
      "tokens_seen": 1724841984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048212637913741226,
      "loss": 1.4205,
      "theoretical_loss": 3.502072660718457,
      "tokens_seen": 1724907520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821063189568706,
      "loss": 1.4712,
      "theoretical_loss": 3.5020613343193507,
      "tokens_seen": 1724973056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000482086258776329,
      "loss": 1.4408,
      "theoretical_loss": 3.5020500084710378,
      "tokens_seen": 1725038592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820661985957874,
      "loss": 1.3621,
      "theoretical_loss": 3.502038683173471,
      "tokens_seen": 1725104128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048204613841524576,
      "loss": 1.4467,
      "theoretical_loss": 3.502027358426602,
      "tokens_seen": 1725169664
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820260782347041,
      "loss": 1.35,
      "theoretical_loss": 3.502016034230384,
      "tokens_seen": 1725235200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820060180541625,
      "loss": 1.4556,
      "theoretical_loss": 3.5020047105847687,
      "tokens_seen": 1725300736
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 630228,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3771824836730957,
      "objective/train/theoretical_loss": 3.5019933874897085,
      "objective/train/tokens_used": 95825376,
      "theoretical_loss": 3.5019933874897085,
      "tokens_seen": 1725366272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048198595787362085,
      "loss": 1.3527,
      "theoretical_loss": 3.5019933874897085,
      "tokens_seen": 1725366272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048196589769307927,
      "loss": 1.4271,
      "theoretical_loss": 3.5019820649451554,
      "tokens_seen": 1725431808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048194583751253763,
      "loss": 1.3421,
      "theoretical_loss": 3.5019707429510625,
      "tokens_seen": 1725497344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481925777331996,
      "loss": 1.378,
      "theoretical_loss": 3.5019594215073813,
      "tokens_seen": 1725562880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048190571715145435,
      "loss": 1.3448,
      "theoretical_loss": 3.501948100614065,
      "tokens_seen": 1725628416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048188565697091277,
      "loss": 1.3443,
      "theoretical_loss": 3.5019367802710653,
      "tokens_seen": 1725693952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048186559679037113,
      "loss": 1.3398,
      "theoretical_loss": 3.5019254604783345,
      "tokens_seen": 1725759488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818455366098295,
      "loss": 1.4324,
      "theoretical_loss": 3.5019141412358255,
      "tokens_seen": 1725825024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048182547642928786,
      "loss": 1.4844,
      "theoretical_loss": 3.5019028225434905,
      "tokens_seen": 1725890560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818054162487462,
      "loss": 1.4636,
      "theoretical_loss": 3.5018915044012813,
      "tokens_seen": 1725956096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048178535606820463,
      "loss": 1.4421,
      "theoretical_loss": 3.5018801868091507,
      "tokens_seen": 1726021632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481765295887663,
      "loss": 1.4172,
      "theoretical_loss": 3.5018688697670513,
      "tokens_seen": 1726087168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048174523570712136,
      "loss": 1.4705,
      "theoretical_loss": 3.501857553274935,
      "tokens_seen": 1726152704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817251755265797,
      "loss": 1.3969,
      "theoretical_loss": 3.501846237332755,
      "tokens_seen": 1726218240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048170511534603814,
      "loss": 1.3441,
      "theoretical_loss": 3.5018349219404628,
      "tokens_seen": 1726283776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816850551654965,
      "loss": 1.4434,
      "theoretical_loss": 3.501823607098011,
      "tokens_seen": 1726349312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048166499498495486,
      "loss": 1.342,
      "theoretical_loss": 3.5018122928053526,
      "tokens_seen": 1726414848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816449348044132,
      "loss": 1.2928,
      "theoretical_loss": 3.5018009790624394,
      "tokens_seen": 1726480384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816248746238716,
      "loss": 1.3533,
      "theoretical_loss": 3.5017896658692234,
      "tokens_seen": 1726545920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048160481444333,
      "loss": 1.3742,
      "theoretical_loss": 3.5017783532256583,
      "tokens_seen": 1726611456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048158475426278837,
      "loss": 1.5971,
      "theoretical_loss": 3.501767041131696,
      "tokens_seen": 1726676992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048156469408224673,
      "loss": 1.38,
      "theoretical_loss": 3.5017557295872885,
      "tokens_seen": 1726742528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815446339017051,
      "loss": 1.321,
      "theoretical_loss": 3.5017444185923887,
      "tokens_seen": 1726808064
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815245737211635,
      "loss": 1.2945,
      "theoretical_loss": 3.501733108146949,
      "tokens_seen": 1726873600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048150451354062187,
      "loss": 1.3117,
      "theoretical_loss": 3.5017217982509212,
      "tokens_seen": 1726939136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048148445336008023,
      "loss": 1.4456,
      "theoretical_loss": 3.5017104889042585,
      "tokens_seen": 1727004672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004814643931795386,
      "loss": 1.3466,
      "theoretical_loss": 3.5016991801069137,
      "tokens_seen": 1727070208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048144433299899696,
      "loss": 1.3321,
      "theoretical_loss": 3.501687871858839,
      "tokens_seen": 1727135744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048142427281845537,
      "loss": 1.2479,
      "theoretical_loss": 3.501676564159986,
      "tokens_seen": 1727201280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048140421263791373,
      "loss": 1.3769,
      "theoretical_loss": 3.501665257010308,
      "tokens_seen": 1727266816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813841524573721,
      "loss": 1.4701,
      "theoretical_loss": 3.5016539504097572,
      "tokens_seen": 1727332352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048136409227683046,
      "loss": 1.4552,
      "theoretical_loss": 3.501642644358286,
      "tokens_seen": 1727397888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813440320962889,
      "loss": 1.3864,
      "theoretical_loss": 3.501631338855848,
      "tokens_seen": 1727463424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048132397191574724,
      "loss": 1.37,
      "theoretical_loss": 3.5016200339023946,
      "tokens_seen": 1727528960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813039117352056,
      "loss": 1.336,
      "theoretical_loss": 3.501608729497878,
      "tokens_seen": 1727594496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048128385155466396,
      "loss": 1.3311,
      "theoretical_loss": 3.5015974256422515,
      "tokens_seen": 1727660032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812637913741223,
      "loss": 1.411,
      "theoretical_loss": 3.501586122335468,
      "tokens_seen": 1727725568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048124373119358074,
      "loss": 1.3059,
      "theoretical_loss": 3.501574819577479,
      "tokens_seen": 1727791104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048122367101303916,
      "loss": 1.4278,
      "theoretical_loss": 3.5015635173682376,
      "tokens_seen": 1727856640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812036108324975,
      "loss": 1.4128,
      "theoretical_loss": 3.5015522157076964,
      "tokens_seen": 1727922176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811835506519559,
      "loss": 1.4261,
      "theoretical_loss": 3.5015409145958074,
      "tokens_seen": 1727987712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811634904714143,
      "loss": 1.4068,
      "theoretical_loss": 3.5015296140325236,
      "tokens_seen": 1728053248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048114343029087266,
      "loss": 1.3934,
      "theoretical_loss": 3.5015183140177975,
      "tokens_seen": 1728118784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481123370110331,
      "loss": 1.3494,
      "theoretical_loss": 3.5015070145515823,
      "tokens_seen": 1728184320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811033099297894,
      "loss": 1.366,
      "theoretical_loss": 3.5014957156338293,
      "tokens_seen": 1728249856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810832497492478,
      "loss": 1.3325,
      "theoretical_loss": 3.501484417264492,
      "tokens_seen": 1728315392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048106318956870616,
      "loss": 1.3124,
      "theoretical_loss": 3.5014731194435225,
      "tokens_seen": 1728380928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810431293881645,
      "loss": 1.3541,
      "theoretical_loss": 3.501461822170874,
      "tokens_seen": 1728446464
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810230692076229,
      "loss": 1.347,
      "theoretical_loss": 3.5014505254464985,
      "tokens_seen": 1728512000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048100300902708125,
      "loss": 1.3788,
      "theoretical_loss": 3.501439229270349,
      "tokens_seen": 1728577536
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 631392,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.398906946182251,
      "objective/train/theoretical_loss": 3.5014279336423777,
      "objective/train/tokens_used": 99102176,
      "theoretical_loss": 3.5014279336423777,
      "tokens_seen": 1728643072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048098294884653967,
      "loss": 1.3433,
      "theoretical_loss": 3.5014279336423777,
      "tokens_seen": 1728643072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048096288866599803,
      "loss": 1.3671,
      "theoretical_loss": 3.501416638562538,
      "tokens_seen": 1728708608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004809428284854564,
      "loss": 1.4467,
      "theoretical_loss": 3.501405344030781,
      "tokens_seen": 1728774144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048092276830491475,
      "loss": 1.3963,
      "theoretical_loss": 3.501394050047061,
      "tokens_seen": 1728839680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048090270812437317,
      "loss": 1.3759,
      "theoretical_loss": 3.5013827566113296,
      "tokens_seen": 1728905216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048088264794383153,
      "loss": 1.261,
      "theoretical_loss": 3.5013714637235402,
      "tokens_seen": 1728970752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808625877632899,
      "loss": 1.3582,
      "theoretical_loss": 3.501360171383645,
      "tokens_seen": 1729036288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048084252758274826,
      "loss": 1.3878,
      "theoretical_loss": 3.5013488795915966,
      "tokens_seen": 1729101824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808224674022066,
      "loss": 1.2432,
      "theoretical_loss": 3.5013375883473477,
      "tokens_seen": 1729167360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048080240722166503,
      "loss": 1.359,
      "theoretical_loss": 3.501326297650851,
      "tokens_seen": 1729232896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807823470411234,
      "loss": 1.4311,
      "theoretical_loss": 3.5013150075020594,
      "tokens_seen": 1729298432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048076228686058176,
      "loss": 1.403,
      "theoretical_loss": 3.5013037179009245,
      "tokens_seen": 1729363968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807422266800401,
      "loss": 1.4158,
      "theoretical_loss": 3.5012924288474006,
      "tokens_seen": 1729429504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048072216649949854,
      "loss": 1.3937,
      "theoretical_loss": 3.5012811403414394,
      "tokens_seen": 1729495040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807021063189569,
      "loss": 1.2947,
      "theoretical_loss": 3.501269852382994,
      "tokens_seen": 1729560576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048068204613841526,
      "loss": 1.4609,
      "theoretical_loss": 3.501258564972017,
      "tokens_seen": 1729626112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806619859578736,
      "loss": 1.413,
      "theoretical_loss": 3.501247278108461,
      "tokens_seen": 1729691648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480641925777332,
      "loss": 1.4556,
      "theoretical_loss": 3.501235991792279,
      "tokens_seen": 1729757184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806218655967904,
      "loss": 1.3312,
      "theoretical_loss": 3.5012247060234225,
      "tokens_seen": 1729822720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048060180541624877,
      "loss": 1.3286,
      "theoretical_loss": 3.5012134208018457,
      "tokens_seen": 1729888256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048058174523570713,
      "loss": 1.3236,
      "theoretical_loss": 3.501202136127501,
      "tokens_seen": 1729953792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805616850551655,
      "loss": 1.3335,
      "theoretical_loss": 3.5011908520003407,
      "tokens_seen": 1730019328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805416248746239,
      "loss": 1.3722,
      "theoretical_loss": 3.501179568420318,
      "tokens_seen": 1730084864
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048052156469408227,
      "loss": 1.5094,
      "theoretical_loss": 3.5011682853873856,
      "tokens_seen": 1730150400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048050150451354063,
      "loss": 1.4556,
      "theoretical_loss": 3.501157002901495,
      "tokens_seen": 1730215936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480481444332999,
      "loss": 1.3884,
      "theoretical_loss": 3.501145720962601,
      "tokens_seen": 1730281472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048046138415245736,
      "loss": 1.278,
      "theoretical_loss": 3.5011344395706554,
      "tokens_seen": 1730347008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048044132397191577,
      "loss": 1.352,
      "theoretical_loss": 3.501123158725611,
      "tokens_seen": 1730412544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048042126379137413,
      "loss": 1.3292,
      "theoretical_loss": 3.5011118784274204,
      "tokens_seen": 1730478080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804012036108325,
      "loss": 1.37,
      "theoretical_loss": 3.501100598676036,
      "tokens_seen": 1730543616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048038114343029086,
      "loss": 1.3221,
      "theoretical_loss": 3.501089319471412,
      "tokens_seen": 1730609152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803610832497493,
      "loss": 1.4695,
      "theoretical_loss": 3.5010780408135,
      "tokens_seen": 1730674688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048034102306920764,
      "loss": 1.3348,
      "theoretical_loss": 3.501066762702253,
      "tokens_seen": 1730740224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480320962888666,
      "loss": 1.3532,
      "theoretical_loss": 3.501055485137624,
      "tokens_seen": 1730805760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048030090270812436,
      "loss": 1.2408,
      "theoretical_loss": 3.5010442081195654,
      "tokens_seen": 1730871296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802808425275827,
      "loss": 1.2848,
      "theoretical_loss": 3.5010329316480306,
      "tokens_seen": 1730936832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048026078234704114,
      "loss": 1.3686,
      "theoretical_loss": 3.501021655722972,
      "tokens_seen": 1731002368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802407221664995,
      "loss": 1.3355,
      "theoretical_loss": 3.501010380344343,
      "tokens_seen": 1731067904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048022066198595786,
      "loss": 1.4343,
      "theoretical_loss": 3.5009991055120957,
      "tokens_seen": 1731133440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048020060180541623,
      "loss": 1.4387,
      "theoretical_loss": 3.5009878312261833,
      "tokens_seen": 1731198976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048018054162487464,
      "loss": 1.4586,
      "theoretical_loss": 3.500976557486559,
      "tokens_seen": 1731264512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480160481444333,
      "loss": 1.3992,
      "theoretical_loss": 3.5009652842931747,
      "tokens_seen": 1731330048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048014042126379137,
      "loss": 1.4603,
      "theoretical_loss": 3.5009540116459843,
      "tokens_seen": 1731395584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048012036108324973,
      "loss": 1.3138,
      "theoretical_loss": 3.5009427395449397,
      "tokens_seen": 1731461120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048010030090270815,
      "loss": 1.2857,
      "theoretical_loss": 3.5009314679899948,
      "tokens_seen": 1731526656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800802407221665,
      "loss": 1.2783,
      "theoretical_loss": 3.5009201969811015,
      "tokens_seen": 1731592192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048006018054162487,
      "loss": 1.3111,
      "theoretical_loss": 3.500908926518213,
      "tokens_seen": 1731657728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048004012036108323,
      "loss": 1.3926,
      "theoretical_loss": 3.500897656601283,
      "tokens_seen": 1731723264
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800200601805416,
      "loss": 1.3168,
      "theoretical_loss": 3.5008863872302634,
      "tokens_seen": 1731788800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048,
      "loss": 1.4196,
      "theoretical_loss": 3.5008751184051072,
      "tokens_seen": 1731854336
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 632543,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4212391376495361,
      "objective/train/theoretical_loss": 3.500863850125768,
      "objective/train/tokens_used": 102378976,
      "theoretical_loss": 3.500863850125768,
      "tokens_seen": 1731919872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799799398194584,
      "loss": 1.3597,
      "theoretical_loss": 3.500863850125768,
      "tokens_seen": 1731919872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047995987963891674,
      "loss": 1.4753,
      "theoretical_loss": 3.5008525823921977,
      "tokens_seen": 1731985408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799398194583751,
      "loss": 1.3734,
      "theoretical_loss": 3.5008413152043496,
      "tokens_seen": 1732050944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799197592778335,
      "loss": 1.3823,
      "theoretical_loss": 3.500830048562177,
      "tokens_seen": 1732116480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798996990972919,
      "loss": 1.3213,
      "theoretical_loss": 3.5008187824656325,
      "tokens_seen": 1732182016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047987963891675024,
      "loss": 1.356,
      "theoretical_loss": 3.5008075169146693,
      "tokens_seen": 1732247552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798595787362086,
      "loss": 1.4404,
      "theoretical_loss": 3.50079625190924,
      "tokens_seen": 1732313088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047983951855566696,
      "loss": 1.4043,
      "theoretical_loss": 3.500784987449298,
      "tokens_seen": 1732378624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798194583751254,
      "loss": 1.4762,
      "theoretical_loss": 3.5007737235347953,
      "tokens_seen": 1732444160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047979939819458374,
      "loss": 1.4518,
      "theoretical_loss": 3.500762460165686,
      "tokens_seen": 1732509696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797793380140421,
      "loss": 1.1211,
      "theoretical_loss": 3.5007511973419225,
      "tokens_seen": 1732575232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047975927783350047,
      "loss": 1.3428,
      "theoretical_loss": 3.5007399350634576,
      "tokens_seen": 1732640768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797392176529589,
      "loss": 1.3472,
      "theoretical_loss": 3.5007286733302445,
      "tokens_seen": 1732706304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797191574724173,
      "loss": 1.2619,
      "theoretical_loss": 3.5007174121422366,
      "tokens_seen": 1732771840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047969909729187566,
      "loss": 1.5019,
      "theoretical_loss": 3.5007061514993865,
      "tokens_seen": 1732837376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479679037111334,
      "loss": 1.3659,
      "theoretical_loss": 3.500694891401647,
      "tokens_seen": 1732902912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796589769307924,
      "loss": 1.4453,
      "theoretical_loss": 3.500683631848971,
      "tokens_seen": 1732968448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796389167502508,
      "loss": 1.3936,
      "theoretical_loss": 3.5006723728413123,
      "tokens_seen": 1733033984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047961885656970917,
      "loss": 1.4513,
      "theoretical_loss": 3.500661114378623,
      "tokens_seen": 1733099520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047959879638916753,
      "loss": 1.4493,
      "theoretical_loss": 3.500649856460856,
      "tokens_seen": 1733165056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795787362086259,
      "loss": 1.3189,
      "theoretical_loss": 3.5006385990879663,
      "tokens_seen": 1733230592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795586760280843,
      "loss": 1.3582,
      "theoretical_loss": 3.500627342259904,
      "tokens_seen": 1733296128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047953861584754267,
      "loss": 1.4295,
      "theoretical_loss": 3.5006160859766244,
      "tokens_seen": 1733361664
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047951855566700103,
      "loss": 1.3399,
      "theoretical_loss": 3.5006048302380792,
      "tokens_seen": 1733427200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794984954864594,
      "loss": 1.4912,
      "theoretical_loss": 3.5005935750442223,
      "tokens_seen": 1733492736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047947843530591776,
      "loss": 1.3689,
      "theoretical_loss": 3.5005823203950066,
      "tokens_seen": 1733558272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047945837512537617,
      "loss": 1.4018,
      "theoretical_loss": 3.5005710662903846,
      "tokens_seen": 1733623808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047943831494483453,
      "loss": 1.4469,
      "theoretical_loss": 3.50055981273031,
      "tokens_seen": 1733689344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794182547642929,
      "loss": 1.4087,
      "theoretical_loss": 3.500548559714735,
      "tokens_seen": 1733754880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047939819458375126,
      "loss": 1.3906,
      "theoretical_loss": 3.500537307243614,
      "tokens_seen": 1733820416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793781344032097,
      "loss": 1.3194,
      "theoretical_loss": 3.5005260553168993,
      "tokens_seen": 1733885952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047935807422266804,
      "loss": 1.3877,
      "theoretical_loss": 3.5005148039345437,
      "tokens_seen": 1733951488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793380140421264,
      "loss": 1.4208,
      "theoretical_loss": 3.500503553096501,
      "tokens_seen": 1734017024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047931795386158476,
      "loss": 1.4336,
      "theoretical_loss": 3.5004923028027237,
      "tokens_seen": 1734082560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792978936810431,
      "loss": 1.3924,
      "theoretical_loss": 3.500481053053165,
      "tokens_seen": 1734148096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047927783350050154,
      "loss": 1.3703,
      "theoretical_loss": 3.5004698038477784,
      "tokens_seen": 1734213632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792577733199599,
      "loss": 1.3303,
      "theoretical_loss": 3.500458555186517,
      "tokens_seen": 1734279168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047923771313941827,
      "loss": 1.3545,
      "theoretical_loss": 3.500447307069333,
      "tokens_seen": 1734344704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047921765295887663,
      "loss": 1.3707,
      "theoretical_loss": 3.500436059496181,
      "tokens_seen": 1734410240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047919759277833504,
      "loss": 1.3634,
      "theoretical_loss": 3.5004248124670125,
      "tokens_seen": 1734475776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791775325977934,
      "loss": 1.3367,
      "theoretical_loss": 3.5004135659817823,
      "tokens_seen": 1734541312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047915747241725177,
      "loss": 1.3838,
      "theoretical_loss": 3.500402320040442,
      "tokens_seen": 1734606848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047913741223671013,
      "loss": 1.4155,
      "theoretical_loss": 3.5003910746429456,
      "tokens_seen": 1734672384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047911735205616855,
      "loss": 1.5309,
      "theoretical_loss": 3.5003798297892468,
      "tokens_seen": 1734737920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790972918756269,
      "loss": 1.5158,
      "theoretical_loss": 3.500368585479297,
      "tokens_seen": 1734803456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047907723169508527,
      "loss": 1.4727,
      "theoretical_loss": 3.500357341713051,
      "tokens_seen": 1734868992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047905717151454363,
      "loss": 1.3598,
      "theoretical_loss": 3.5003460984904615,
      "tokens_seen": 1734934528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479037111334002,
      "loss": 1.3277,
      "theoretical_loss": 3.5003348558114817,
      "tokens_seen": 1735000064
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790170511534604,
      "loss": 1.4234,
      "theoretical_loss": 3.5003236136760645,
      "tokens_seen": 1735065600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789969909729188,
      "loss": 1.404,
      "theoretical_loss": 3.5003123720841636,
      "tokens_seen": 1735131136
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 633815,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3423707485198975,
      "objective/train/theoretical_loss": 3.500301131035731,
      "objective/train/tokens_used": 105655776,
      "theoretical_loss": 3.500301131035731,
      "tokens_seen": 1735196672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047897693079237714,
      "loss": 1.4197,
      "theoretical_loss": 3.500301131035731,
      "tokens_seen": 1735196672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789568706118355,
      "loss": 1.3932,
      "theoretical_loss": 3.5002898905307216,
      "tokens_seen": 1735262208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789368104312939,
      "loss": 1.4013,
      "theoretical_loss": 3.5002786505690873,
      "tokens_seen": 1735327744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789167502507523,
      "loss": 1.3346,
      "theoretical_loss": 3.5002674111507823,
      "tokens_seen": 1735393280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047889669007021064,
      "loss": 1.3498,
      "theoretical_loss": 3.5002561722757592,
      "tokens_seen": 1735458816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478876629889669,
      "loss": 1.4119,
      "theoretical_loss": 3.500244933943971,
      "tokens_seen": 1735524352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047885656970912736,
      "loss": 1.282,
      "theoretical_loss": 3.5002336961553713,
      "tokens_seen": 1735589888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788365095285858,
      "loss": 1.4133,
      "theoretical_loss": 3.5002224589099136,
      "tokens_seen": 1735655424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047881644934804414,
      "loss": 1.4711,
      "theoretical_loss": 3.5002112222075508,
      "tokens_seen": 1735720960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787963891675025,
      "loss": 1.2595,
      "theoretical_loss": 3.5001999860482362,
      "tokens_seen": 1735786496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047877632898696087,
      "loss": 1.3564,
      "theoretical_loss": 3.500188750431923,
      "tokens_seen": 1735852032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787562688064193,
      "loss": 1.4286,
      "theoretical_loss": 3.5001775153585646,
      "tokens_seen": 1735917568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047873620862587765,
      "loss": 1.3318,
      "theoretical_loss": 3.5001662808281138,
      "tokens_seen": 1735983104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478716148445336,
      "loss": 1.3822,
      "theoretical_loss": 3.5001550468405247,
      "tokens_seen": 1736048640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047869608826479437,
      "loss": 1.4199,
      "theoretical_loss": 3.5001438133957494,
      "tokens_seen": 1736114176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047867602808425273,
      "loss": 1.3039,
      "theoretical_loss": 3.5001325804937427,
      "tokens_seen": 1736179712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047865596790371115,
      "loss": 1.381,
      "theoretical_loss": 3.500121348134457,
      "tokens_seen": 1736245248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786359077231695,
      "loss": 1.4469,
      "theoretical_loss": 3.5001101163178454,
      "tokens_seen": 1736310784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786158475426279,
      "loss": 1.508,
      "theoretical_loss": 3.5000988850438617,
      "tokens_seen": 1736376320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047859578736208624,
      "loss": 1.338,
      "theoretical_loss": 3.500087654312459,
      "tokens_seen": 1736441856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047857572718154465,
      "loss": 1.444,
      "theoretical_loss": 3.5000764241235904,
      "tokens_seen": 1736507392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478555667001003,
      "loss": 1.3398,
      "theoretical_loss": 3.500065194477209,
      "tokens_seen": 1736572928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785356068204614,
      "loss": 1.471,
      "theoretical_loss": 3.5000539653732696,
      "tokens_seen": 1736638464
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047851554663991974,
      "loss": 1.4408,
      "theoretical_loss": 3.5000427368117233,
      "tokens_seen": 1736704000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784954864593781,
      "loss": 1.4286,
      "theoretical_loss": 3.5000315087925253,
      "tokens_seen": 1736769536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784754262788365,
      "loss": 1.3556,
      "theoretical_loss": 3.5000202813156287,
      "tokens_seen": 1736835072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784553660982949,
      "loss": 1.2849,
      "theoretical_loss": 3.5000090543809854,
      "tokens_seen": 1736900608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047843530591775324,
      "loss": 1.3791,
      "theoretical_loss": 3.4999978279885506,
      "tokens_seen": 1736966144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784152457372116,
      "loss": 1.3739,
      "theoretical_loss": 3.4999866021382764,
      "tokens_seen": 1737031680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047839518555667,
      "loss": 1.4587,
      "theoretical_loss": 3.4999753768301165,
      "tokens_seen": 1737097216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783751253761284,
      "loss": 1.4282,
      "theoretical_loss": 3.499964152064025,
      "tokens_seen": 1737162752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047835506519558675,
      "loss": 1.3448,
      "theoretical_loss": 3.4999529278399537,
      "tokens_seen": 1737228288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783350050150451,
      "loss": 1.4438,
      "theoretical_loss": 3.499941704157857,
      "tokens_seen": 1737293824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047831494483450347,
      "loss": 1.5265,
      "theoretical_loss": 3.4999304810176888,
      "tokens_seen": 1737359360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782948846539619,
      "loss": 1.32,
      "theoretical_loss": 3.499919258419401,
      "tokens_seen": 1737424896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047827482447342025,
      "loss": 1.3087,
      "theoretical_loss": 3.4999080363629487,
      "tokens_seen": 1737490432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782547642928786,
      "loss": 1.3742,
      "theoretical_loss": 3.4998968148482845,
      "tokens_seen": 1737555968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478234704112337,
      "loss": 1.373,
      "theoretical_loss": 3.4998855938753612,
      "tokens_seen": 1737621504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047821464393179544,
      "loss": 1.4785,
      "theoretical_loss": 3.499874373444133,
      "tokens_seen": 1737687040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781945837512538,
      "loss": 1.382,
      "theoretical_loss": 3.4998631535545535,
      "tokens_seen": 1737752576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047817452357071217,
      "loss": 1.2707,
      "theoretical_loss": 3.4998519342065753,
      "tokens_seen": 1737818112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047815446339017053,
      "loss": 1.3979,
      "theoretical_loss": 3.4998407154001523,
      "tokens_seen": 1737883648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047813440320962895,
      "loss": 1.4073,
      "theoretical_loss": 3.499829497135238,
      "tokens_seen": 1737949184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781143430290873,
      "loss": 1.3476,
      "theoretical_loss": 3.4998182794117856,
      "tokens_seen": 1738014720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047809428284854567,
      "loss": 1.2186,
      "theoretical_loss": 3.499807062229749,
      "tokens_seen": 1738080256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047807422266800403,
      "loss": 1.5325,
      "theoretical_loss": 3.4997958455890816,
      "tokens_seen": 1738145792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780541624874624,
      "loss": 1.5235,
      "theoretical_loss": 3.4997846294897363,
      "tokens_seen": 1738211328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780341023069208,
      "loss": 1.342,
      "theoretical_loss": 3.499773413931667,
      "tokens_seen": 1738276864
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780140421263792,
      "loss": 1.378,
      "theoretical_loss": 3.4997621989148273,
      "tokens_seen": 1738342400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047799398194583754,
      "loss": 1.3128,
      "theoretical_loss": 3.49975098443917,
      "tokens_seen": 1738407936
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 635089,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5819060802459717,
      "objective/train/theoretical_loss": 3.4997397705046493,
      "objective/train/tokens_used": 108932576,
      "theoretical_loss": 3.4997397705046493,
      "tokens_seen": 1738473472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779739217652959,
      "loss": 1.421,
      "theoretical_loss": 3.4997397705046493,
      "tokens_seen": 1738473472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779538615847543,
      "loss": 1.3741,
      "theoretical_loss": 3.4997285571112178,
      "tokens_seen": 1738539008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779338014042127,
      "loss": 1.4196,
      "theoretical_loss": 3.4997173442588307,
      "tokens_seen": 1738604544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047791374122367104,
      "loss": 1.3153,
      "theoretical_loss": 3.4997061319474394,
      "tokens_seen": 1738670080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778936810431294,
      "loss": 1.3222,
      "theoretical_loss": 3.4996949201769993,
      "tokens_seen": 1738735616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047787362086258776,
      "loss": 1.3881,
      "theoretical_loss": 3.4996837089474626,
      "tokens_seen": 1738801152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778535606820462,
      "loss": 1.3621,
      "theoretical_loss": 3.4996724982587835,
      "tokens_seen": 1738866688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047783350050150454,
      "loss": 1.3975,
      "theoretical_loss": 3.4996612881109153,
      "tokens_seen": 1738932224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778134403209629,
      "loss": 1.3734,
      "theoretical_loss": 3.4996500785038114,
      "tokens_seen": 1738997760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047779338014042127,
      "loss": 1.4063,
      "theoretical_loss": 3.499638869437425,
      "tokens_seen": 1739063296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777733199598797,
      "loss": 1.4812,
      "theoretical_loss": 3.4996276609117105,
      "tokens_seen": 1739128832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047775325977933805,
      "loss": 1.422,
      "theoretical_loss": 3.499616452926621,
      "tokens_seen": 1739194368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777331995987964,
      "loss": 1.2819,
      "theoretical_loss": 3.4996052454821105,
      "tokens_seen": 1739259904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047771313941825477,
      "loss": 1.4095,
      "theoretical_loss": 3.499594038578132,
      "tokens_seen": 1739325440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047769307923771313,
      "loss": 1.3387,
      "theoretical_loss": 3.4995828322146387,
      "tokens_seen": 1739390976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047767301905717155,
      "loss": 1.4609,
      "theoretical_loss": 3.4995716263915853,
      "tokens_seen": 1739456512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776529588766299,
      "loss": 1.3585,
      "theoretical_loss": 3.4995604211089244,
      "tokens_seen": 1739522048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776328986960883,
      "loss": 1.2814,
      "theoretical_loss": 3.49954921636661,
      "tokens_seen": 1739587584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047761283851554664,
      "loss": 1.3707,
      "theoretical_loss": 3.499538012164596,
      "tokens_seen": 1739653120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047759277833500505,
      "loss": 1.3301,
      "theoretical_loss": 3.499526808502835,
      "tokens_seen": 1739718656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775727181544634,
      "loss": 1.5041,
      "theoretical_loss": 3.4995156053812817,
      "tokens_seen": 1739784192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775526579739218,
      "loss": 1.302,
      "theoretical_loss": 3.4995044027998894,
      "tokens_seen": 1739849728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047753259779338014,
      "loss": 1.4329,
      "theoretical_loss": 3.499493200758611,
      "tokens_seen": 1739915264
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775125376128385,
      "loss": 1.44,
      "theoretical_loss": 3.499481999257401,
      "tokens_seen": 1739980800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774924774322969,
      "loss": 1.3856,
      "theoretical_loss": 3.4994707982962128,
      "tokens_seen": 1740046336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774724172517553,
      "loss": 1.4457,
      "theoretical_loss": 3.499459597875,
      "tokens_seen": 1740111872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047745235707121364,
      "loss": 1.4626,
      "theoretical_loss": 3.4994483979937154,
      "tokens_seen": 1740177408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000477432296890672,
      "loss": 1.3459,
      "theoretical_loss": 3.4994371986523145,
      "tokens_seen": 1740242944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774122367101304,
      "loss": 1.2833,
      "theoretical_loss": 3.499425999850749,
      "tokens_seen": 1740308480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773921765295888,
      "loss": 1.3325,
      "theoretical_loss": 3.4994148015889737,
      "tokens_seen": 1740374016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047737211634904715,
      "loss": 1.4385,
      "theoretical_loss": 3.499403603866942,
      "tokens_seen": 1740439552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773520561685055,
      "loss": 1.3918,
      "theoretical_loss": 3.499392406684607,
      "tokens_seen": 1740505088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047733199598796387,
      "loss": 1.359,
      "theoretical_loss": 3.4993812100419235,
      "tokens_seen": 1740570624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773119358074223,
      "loss": 1.2948,
      "theoretical_loss": 3.4993700139388446,
      "tokens_seen": 1740636160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047729187562688065,
      "loss": 1.4623,
      "theoretical_loss": 3.4993588183753235,
      "tokens_seen": 1740701696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477271815446339,
      "loss": 1.4959,
      "theoretical_loss": 3.4993476233513148,
      "tokens_seen": 1740767232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772517552657974,
      "loss": 1.4357,
      "theoretical_loss": 3.499336428866771,
      "tokens_seen": 1740832768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772316950852558,
      "loss": 1.367,
      "theoretical_loss": 3.4993252349216473,
      "tokens_seen": 1740898304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047721163490471415,
      "loss": 1.4615,
      "theoretical_loss": 3.4993140415158956,
      "tokens_seen": 1740963840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771915747241725,
      "loss": 1.4295,
      "theoretical_loss": 3.4993028486494717,
      "tokens_seen": 1741029376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771715145436309,
      "loss": 1.3367,
      "theoretical_loss": 3.4992916563223275,
      "tokens_seen": 1741094912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047715145436308924,
      "loss": 1.3049,
      "theoretical_loss": 3.4992804645344178,
      "tokens_seen": 1741160448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047713139418254766,
      "loss": 1.4341,
      "theoretical_loss": 3.499269273285696,
      "tokens_seen": 1741225984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477111334002006,
      "loss": 1.3332,
      "theoretical_loss": 3.499258082576116,
      "tokens_seen": 1741291520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770912738214644,
      "loss": 1.3339,
      "theoretical_loss": 3.4992468924056306,
      "tokens_seen": 1741357056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047707121364092274,
      "loss": 1.3419,
      "theoretical_loss": 3.499235702774195,
      "tokens_seen": 1741422592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047705115346038116,
      "loss": 1.3719,
      "theoretical_loss": 3.499224513681762,
      "tokens_seen": 1741488128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770310932798395,
      "loss": 1.3671,
      "theoretical_loss": 3.4992133251282853,
      "tokens_seen": 1741553664
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770110330992979,
      "loss": 1.3919,
      "theoretical_loss": 3.499202137113719,
      "tokens_seen": 1741619200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047699097291875625,
      "loss": 1.4125,
      "theoretical_loss": 3.499190949638017,
      "tokens_seen": 1741684736
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 636395,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3723151683807373,
      "objective/train/theoretical_loss": 3.499179762701133,
      "objective/train/tokens_used": 112209376,
      "theoretical_loss": 3.499179762701133,
      "tokens_seen": 1741750272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047697091273821466,
      "loss": 1.3412,
      "theoretical_loss": 3.499179762701133,
      "tokens_seen": 1741750272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476950852557673,
      "loss": 1.3023,
      "theoretical_loss": 3.4991685763030205,
      "tokens_seen": 1741815808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769307923771314,
      "loss": 1.347,
      "theoretical_loss": 3.4991573904436333,
      "tokens_seen": 1741881344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047691073219658975,
      "loss": 1.4426,
      "theoretical_loss": 3.499146205122926,
      "tokens_seen": 1741946880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768906720160481,
      "loss": 1.3363,
      "theoretical_loss": 3.4991350203408507,
      "tokens_seen": 1742012416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047687061183550653,
      "loss": 1.3294,
      "theoretical_loss": 3.499123836097363,
      "tokens_seen": 1742077952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768505516549649,
      "loss": 1.4609,
      "theoretical_loss": 3.499112652392416,
      "tokens_seen": 1742143488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047683049147442325,
      "loss": 1.3622,
      "theoretical_loss": 3.499101469225963,
      "tokens_seen": 1742209024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768104312938816,
      "loss": 1.2921,
      "theoretical_loss": 3.4990902865979585,
      "tokens_seen": 1742274560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047679037111334003,
      "loss": 1.3533,
      "theoretical_loss": 3.499079104508356,
      "tokens_seen": 1742340096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767703109327984,
      "loss": 1.3406,
      "theoretical_loss": 3.499067922957109,
      "tokens_seen": 1742405632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047675025075225676,
      "loss": 1.4049,
      "theoretical_loss": 3.4990567419441723,
      "tokens_seen": 1742471168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767301905717151,
      "loss": 1.4386,
      "theoretical_loss": 3.499045561469499,
      "tokens_seen": 1742536704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047671013039117353,
      "loss": 1.3344,
      "theoretical_loss": 3.4990343815330434,
      "tokens_seen": 1742602240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047669007021063195,
      "loss": 1.3613,
      "theoretical_loss": 3.4990232021347585,
      "tokens_seen": 1742667776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766700100300903,
      "loss": 1.427,
      "theoretical_loss": 3.4990120232745987,
      "tokens_seen": 1742733312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766499498495487,
      "loss": 1.4005,
      "theoretical_loss": 3.4990008449525183,
      "tokens_seen": 1742798848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047662988966900704,
      "loss": 1.2999,
      "theoretical_loss": 3.4989896671684706,
      "tokens_seen": 1742864384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047660982948846545,
      "loss": 1.4541,
      "theoretical_loss": 3.49897848992241,
      "tokens_seen": 1742929920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765897693079238,
      "loss": 1.4126,
      "theoretical_loss": 3.4989673132142896,
      "tokens_seen": 1742995456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765697091273822,
      "loss": 1.4026,
      "theoretical_loss": 3.498956137044064,
      "tokens_seen": 1743060992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047654964894684054,
      "loss": 1.4616,
      "theoretical_loss": 3.498944961411687,
      "tokens_seen": 1743126528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765295887662989,
      "loss": 1.4183,
      "theoretical_loss": 3.4989337863171115,
      "tokens_seen": 1743192064
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765095285857573,
      "loss": 1.3282,
      "theoretical_loss": 3.4989226117602925,
      "tokens_seen": 1743257600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764894684052157,
      "loss": 1.5221,
      "theoretical_loss": 3.498911437741184,
      "tokens_seen": 1743323136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047646940822467404,
      "loss": 1.4154,
      "theoretical_loss": 3.4989002642597393,
      "tokens_seen": 1743388672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764493480441324,
      "loss": 1.4993,
      "theoretical_loss": 3.4988890913159123,
      "tokens_seen": 1743454208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764292878635908,
      "loss": 1.3779,
      "theoretical_loss": 3.498877918909657,
      "tokens_seen": 1743519744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764092276830492,
      "loss": 1.432,
      "theoretical_loss": 3.4988667470409283,
      "tokens_seen": 1743585280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047638916750250755,
      "loss": 1.5108,
      "theoretical_loss": 3.498855575709679,
      "tokens_seen": 1743650816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763691073219659,
      "loss": 1.3134,
      "theoretical_loss": 3.4988444049158627,
      "tokens_seen": 1743716352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047634904714142427,
      "loss": 1.3464,
      "theoretical_loss": 3.4988332346594344,
      "tokens_seen": 1743781888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763289869608827,
      "loss": 1.3202,
      "theoretical_loss": 3.498822064940348,
      "tokens_seen": 1743847424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047630892678034105,
      "loss": 1.4056,
      "theoretical_loss": 3.498810895758557,
      "tokens_seen": 1743912960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762888665997994,
      "loss": 1.417,
      "theoretical_loss": 3.498799727114015,
      "tokens_seen": 1743978496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762688064192578,
      "loss": 1.4262,
      "theoretical_loss": 3.4987885590066767,
      "tokens_seen": 1744044032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762487462387162,
      "loss": 1.3868,
      "theoretical_loss": 3.498777391436496,
      "tokens_seen": 1744109568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047622868605817455,
      "loss": 1.3644,
      "theoretical_loss": 3.4987662244034268,
      "tokens_seen": 1744175104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762086258776329,
      "loss": 1.4063,
      "theoretical_loss": 3.498755057907423,
      "tokens_seen": 1744240640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761885656970913,
      "loss": 1.426,
      "theoretical_loss": 3.4987438919484384,
      "tokens_seen": 1744306176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047616850551654964,
      "loss": 1.3395,
      "theoretical_loss": 3.498732726526427,
      "tokens_seen": 1744371712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047614844533600806,
      "loss": 1.3326,
      "theoretical_loss": 3.4987215616413434,
      "tokens_seen": 1744437248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761283851554664,
      "loss": 1.3448,
      "theoretical_loss": 3.498710397293141,
      "tokens_seen": 1744502784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761083249749248,
      "loss": 1.3836,
      "theoretical_loss": 3.498699233481774,
      "tokens_seen": 1744568320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047608826479438314,
      "loss": 1.4019,
      "theoretical_loss": 3.498688070207197,
      "tokens_seen": 1744633856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047606820461384156,
      "loss": 1.4755,
      "theoretical_loss": 3.4986769074693624,
      "tokens_seen": 1744699392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760481444332999,
      "loss": 1.4124,
      "theoretical_loss": 3.498665745268226,
      "tokens_seen": 1744764928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760280842527583,
      "loss": 1.4381,
      "theoretical_loss": 3.498654583603741,
      "tokens_seen": 1744830464
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047600802407221665,
      "loss": 1.395,
      "theoretical_loss": 3.498643422475862,
      "tokens_seen": 1744896000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047598796389167506,
      "loss": 1.3212,
      "theoretical_loss": 3.4986322618845422,
      "tokens_seen": 1744961536
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 637583,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.238525152206421,
      "objective/train/theoretical_loss": 3.4986211018297357,
      "objective/train/tokens_used": 115486176,
      "theoretical_loss": 3.4986211018297357,
      "tokens_seen": 1745027072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759679037111334,
      "loss": 1.2066,
      "theoretical_loss": 3.4986211018297357,
      "tokens_seen": 1745027072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759478435305918,
      "loss": 1.4097,
      "theoretical_loss": 3.498609942311398,
      "tokens_seen": 1745092608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047592778335005015,
      "loss": 1.2245,
      "theoretical_loss": 3.4985987833294816,
      "tokens_seen": 1745158144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759077231695085,
      "loss": 1.3761,
      "theoretical_loss": 3.4985876248839407,
      "tokens_seen": 1745223680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047588766298896693,
      "loss": 1.4018,
      "theoretical_loss": 3.49857646697473,
      "tokens_seen": 1745289216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758676028084253,
      "loss": 1.4224,
      "theoretical_loss": 3.498565309601804,
      "tokens_seen": 1745354752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047584754262788365,
      "loss": 1.3306,
      "theoretical_loss": 3.498554152765115,
      "tokens_seen": 1745420288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475827482447342,
      "loss": 1.4405,
      "theoretical_loss": 3.498542996464619,
      "tokens_seen": 1745485824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047580742226680043,
      "loss": 1.3593,
      "theoretical_loss": 3.4985318407002692,
      "tokens_seen": 1745551360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757873620862588,
      "loss": 1.367,
      "theoretical_loss": 3.4985206854720197,
      "tokens_seen": 1745616896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047576730190571716,
      "loss": 1.4049,
      "theoretical_loss": 3.498509530779825,
      "tokens_seen": 1745682432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757472417251755,
      "loss": 1.4652,
      "theoretical_loss": 3.4984983766236386,
      "tokens_seen": 1745747968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757271815446339,
      "loss": 1.3683,
      "theoretical_loss": 3.498487223003415,
      "tokens_seen": 1745813504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757071213640923,
      "loss": 1.3818,
      "theoretical_loss": 3.498476069919109,
      "tokens_seen": 1745879040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047568706118355066,
      "loss": 1.3599,
      "theoretical_loss": 3.498464917370673,
      "tokens_seen": 1745944576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475667001003009,
      "loss": 1.4126,
      "theoretical_loss": 3.498453765358063,
      "tokens_seen": 1746010112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756469408224674,
      "loss": 1.5066,
      "theoretical_loss": 3.498442613881232,
      "tokens_seen": 1746075648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756268806419258,
      "loss": 1.4062,
      "theoretical_loss": 3.4984314629401347,
      "tokens_seen": 1746141184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047560682046138416,
      "loss": 1.4216,
      "theoretical_loss": 3.498420312534725,
      "tokens_seen": 1746206720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755867602808425,
      "loss": 1.3739,
      "theoretical_loss": 3.498409162664957,
      "tokens_seen": 1746272256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755667001003009,
      "loss": 1.3176,
      "theoretical_loss": 3.4983980133307853,
      "tokens_seen": 1746337792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047554663991975925,
      "loss": 1.3238,
      "theoretical_loss": 3.498386864532163,
      "tokens_seen": 1746403328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047552657973921766,
      "loss": 1.348,
      "theoretical_loss": 3.498375716269046,
      "tokens_seen": 1746468864
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047550651955867603,
      "loss": 1.3714,
      "theoretical_loss": 3.498364568541387,
      "tokens_seen": 1746534400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754864593781344,
      "loss": 1.3557,
      "theoretical_loss": 3.4983534213491407,
      "tokens_seen": 1746599936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047546639919759275,
      "loss": 1.3873,
      "theoretical_loss": 3.4983422746922614,
      "tokens_seen": 1746665472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047544633901705117,
      "loss": 1.442,
      "theoretical_loss": 3.4983311285707033,
      "tokens_seen": 1746731008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047542627883650953,
      "loss": 1.3813,
      "theoretical_loss": 3.4983199829844205,
      "tokens_seen": 1746796544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754062186559679,
      "loss": 1.26,
      "theoretical_loss": 3.498308837933367,
      "tokens_seen": 1746862080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047538615847542625,
      "loss": 1.3855,
      "theoretical_loss": 3.4982976934174976,
      "tokens_seen": 1746927616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753660982948846,
      "loss": 1.3722,
      "theoretical_loss": 3.4982865494367656,
      "tokens_seen": 1746993152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047534603811434303,
      "loss": 1.4065,
      "theoretical_loss": 3.4982754059911265,
      "tokens_seen": 1747058688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753259779338014,
      "loss": 1.309,
      "theoretical_loss": 3.498264263080533,
      "tokens_seen": 1747124224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047530591775325976,
      "loss": 1.5057,
      "theoretical_loss": 3.498253120704941,
      "tokens_seen": 1747189760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752858575727181,
      "loss": 1.4015,
      "theoretical_loss": 3.4982419788643035,
      "tokens_seen": 1747255296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047526579739217654,
      "loss": 1.5058,
      "theoretical_loss": 3.498230837558576,
      "tokens_seen": 1747320832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752457372116349,
      "loss": 1.433,
      "theoretical_loss": 3.498219696787711,
      "tokens_seen": 1747386368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047522567703109326,
      "loss": 1.2959,
      "theoretical_loss": 3.498208556551664,
      "tokens_seen": 1747451904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752056168505517,
      "loss": 1.4287,
      "theoretical_loss": 3.4981974168503887,
      "tokens_seen": 1747517440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047518555667001004,
      "loss": 1.3092,
      "theoretical_loss": 3.49818627768384,
      "tokens_seen": 1747582976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047516549648946846,
      "loss": 1.4277,
      "theoretical_loss": 3.4981751390519715,
      "tokens_seen": 1747648512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751454363089268,
      "loss": 1.4031,
      "theoretical_loss": 3.498164000954738,
      "tokens_seen": 1747714048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751253761283852,
      "loss": 1.3672,
      "theoretical_loss": 3.498152863392094,
      "tokens_seen": 1747779584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047510531594784354,
      "loss": 1.3672,
      "theoretical_loss": 3.498141726363993,
      "tokens_seen": 1747845120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047508525576730196,
      "loss": 1.3441,
      "theoretical_loss": 3.4981305898703896,
      "tokens_seen": 1747910656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750651955867603,
      "loss": 1.4591,
      "theoretical_loss": 3.4981194539112384,
      "tokens_seen": 1747976192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750451354062187,
      "loss": 1.4302,
      "theoretical_loss": 3.4981083184864934,
      "tokens_seen": 1748041728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047502507522567705,
      "loss": 1.3723,
      "theoretical_loss": 3.4980971835961094,
      "tokens_seen": 1748107264
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047500501504513546,
      "loss": 1.3964,
      "theoretical_loss": 3.49808604924004,
      "tokens_seen": 1748172800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749849548645938,
      "loss": 1.3725,
      "theoretical_loss": 3.49807491541824,
      "tokens_seen": 1748238336
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 638722,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1972347497940063,
      "objective/train/theoretical_loss": 3.4980637821306635,
      "objective/train/tokens_used": 118762976,
      "theoretical_loss": 3.4980637821306635,
      "tokens_seen": 1748303872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749648946840522,
      "loss": 1.2796,
      "theoretical_loss": 3.4980637821306635,
      "tokens_seen": 1748303872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047494483450351055,
      "loss": 1.3572,
      "theoretical_loss": 3.4980526493772652,
      "tokens_seen": 1748369408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749247743229689,
      "loss": 1.3555,
      "theoretical_loss": 3.498041517157999,
      "tokens_seen": 1748434944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047490471414242733,
      "loss": 1.3755,
      "theoretical_loss": 3.49803038547282,
      "tokens_seen": 1748500480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748846539618857,
      "loss": 1.3053,
      "theoretical_loss": 3.4980192543216813,
      "tokens_seen": 1748566016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047486459378134405,
      "loss": 1.3668,
      "theoretical_loss": 3.4980081237045386,
      "tokens_seen": 1748631552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748445336008024,
      "loss": 1.2702,
      "theoretical_loss": 3.4979969936213458,
      "tokens_seen": 1748697088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047482447342026083,
      "loss": 1.4395,
      "theoretical_loss": 3.497985864072057,
      "tokens_seen": 1748762624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748044132397192,
      "loss": 1.341,
      "theoretical_loss": 3.4979747350566264,
      "tokens_seen": 1748828160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047478435305917756,
      "loss": 1.3135,
      "theoretical_loss": 3.4979636065750093,
      "tokens_seen": 1748893696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747642928786359,
      "loss": 1.4415,
      "theoretical_loss": 3.497952478627159,
      "tokens_seen": 1748959232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747442326980943,
      "loss": 1.3595,
      "theoretical_loss": 3.4979413512130306,
      "tokens_seen": 1749024768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747241725175527,
      "loss": 1.4819,
      "theoretical_loss": 3.4979302243325785,
      "tokens_seen": 1749090304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047470411233701106,
      "loss": 1.4374,
      "theoretical_loss": 3.497919097985757,
      "tokens_seen": 1749155840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746840521564694,
      "loss": 1.3396,
      "theoretical_loss": 3.49790797217252,
      "tokens_seen": 1749221376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746639919759278,
      "loss": 1.3657,
      "theoretical_loss": 3.4978968468928224,
      "tokens_seen": 1749286912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746439317953862,
      "loss": 1.5177,
      "theoretical_loss": 3.497885722146619,
      "tokens_seen": 1749352448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047462387161484456,
      "loss": 1.3515,
      "theoretical_loss": 3.4978745979338637,
      "tokens_seen": 1749417984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746038114343029,
      "loss": 1.3026,
      "theoretical_loss": 3.497863474254511,
      "tokens_seen": 1749483520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745837512537613,
      "loss": 1.4378,
      "theoretical_loss": 3.4978523511085156,
      "tokens_seen": 1749549056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047456369107321965,
      "loss": 1.3463,
      "theoretical_loss": 3.4978412284958313,
      "tokens_seen": 1749614592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047454363089267807,
      "loss": 1.3778,
      "theoretical_loss": 3.497830106416414,
      "tokens_seen": 1749680128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047452357071213643,
      "loss": 1.4166,
      "theoretical_loss": 3.497818984870216,
      "tokens_seen": 1749745664
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745035105315948,
      "loss": 1.4328,
      "theoretical_loss": 3.4978078638571937,
      "tokens_seen": 1749811200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047448345035105315,
      "loss": 1.4418,
      "theoretical_loss": 3.4977967433773003,
      "tokens_seen": 1749876736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047446339017051157,
      "loss": 1.406,
      "theoretical_loss": 3.497785623430491,
      "tokens_seen": 1749942272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047444332998996993,
      "loss": 1.4632,
      "theoretical_loss": 3.4977745040167205,
      "tokens_seen": 1750007808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744232698094283,
      "loss": 1.3397,
      "theoretical_loss": 3.4977633851359426,
      "tokens_seen": 1750073344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047440320962888666,
      "loss": 1.3761,
      "theoretical_loss": 3.497752266788112,
      "tokens_seen": 1750138880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474383149448345,
      "loss": 1.4273,
      "theoretical_loss": 3.497741148973183,
      "tokens_seen": 1750204416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047436308926780343,
      "loss": 1.3322,
      "theoretical_loss": 3.497730031691111,
      "tokens_seen": 1750269952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743430290872618,
      "loss": 1.3073,
      "theoretical_loss": 3.4977189149418493,
      "tokens_seen": 1750335488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047432296890672016,
      "loss": 1.4261,
      "theoretical_loss": 3.497707798725353,
      "tokens_seen": 1750401024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743029087261785,
      "loss": 1.3149,
      "theoretical_loss": 3.497696683041577,
      "tokens_seen": 1750466560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047428284854563694,
      "loss": 1.3635,
      "theoretical_loss": 3.4976855678904752,
      "tokens_seen": 1750532096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742627883650953,
      "loss": 1.4252,
      "theoretical_loss": 3.497674453272002,
      "tokens_seen": 1750597632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047424272818455366,
      "loss": 1.4385,
      "theoretical_loss": 3.497663339186113,
      "tokens_seen": 1750663168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474222668004012,
      "loss": 1.3353,
      "theoretical_loss": 3.497652225632762,
      "tokens_seen": 1750728704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742026078234704,
      "loss": 1.4279,
      "theoretical_loss": 3.4976411126119036,
      "tokens_seen": 1750794240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741825476429288,
      "loss": 1.3647,
      "theoretical_loss": 3.497630000123492,
      "tokens_seen": 1750859776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047416248746238716,
      "loss": 1.4138,
      "theoretical_loss": 3.497618888167482,
      "tokens_seen": 1750925312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741424272818455,
      "loss": 1.3681,
      "theoretical_loss": 3.4976077767438287,
      "tokens_seen": 1750990848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741223671013039,
      "loss": 1.3681,
      "theoretical_loss": 3.497596665852486,
      "tokens_seen": 1751056384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741023069207623,
      "loss": 1.4527,
      "theoretical_loss": 3.497585555493409,
      "tokens_seen": 1751121920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047408224674022067,
      "loss": 1.4242,
      "theoretical_loss": 3.497574445666552,
      "tokens_seen": 1751187456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047406218655967903,
      "loss": 1.4123,
      "theoretical_loss": 3.4975633363718694,
      "tokens_seen": 1751252992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004740421263791374,
      "loss": 1.4234,
      "theoretical_loss": 3.497552227609316,
      "tokens_seen": 1751318528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047402206619859575,
      "loss": 1.4451,
      "theoretical_loss": 3.4975411193788464,
      "tokens_seen": 1751384064
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047400200601805417,
      "loss": 1.3812,
      "theoretical_loss": 3.4975300116804156,
      "tokens_seen": 1751449600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047398194583751253,
      "loss": 1.3308,
      "theoretical_loss": 3.497518904513978,
      "tokens_seen": 1751515136
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 639884,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3613932132720947,
      "objective/train/theoretical_loss": 3.4975077978794875,
      "objective/train/tokens_used": 122039776,
      "theoretical_loss": 3.4975077978794875,
      "tokens_seen": 1751580672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739618856569709,
      "loss": 1.3875,
      "theoretical_loss": 3.4975077978794875,
      "tokens_seen": 1751580672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047394182547642926,
      "loss": 1.3947,
      "theoretical_loss": 3.497496691776899,
      "tokens_seen": 1751646208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739217652958877,
      "loss": 1.3981,
      "theoretical_loss": 3.497485586206168,
      "tokens_seen": 1751711744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047390170511534604,
      "loss": 1.4743,
      "theoretical_loss": 3.497474481167248,
      "tokens_seen": 1751777280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738816449348044,
      "loss": 1.4963,
      "theoretical_loss": 3.4974633766600944,
      "tokens_seen": 1751842816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047386158475426276,
      "loss": 1.3637,
      "theoretical_loss": 3.4974522726846615,
      "tokens_seen": 1751908352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738415245737211,
      "loss": 1.3598,
      "theoretical_loss": 3.497441169240904,
      "tokens_seen": 1751973888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047382146439317954,
      "loss": 1.3887,
      "theoretical_loss": 3.497430066328777,
      "tokens_seen": 1752039424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738014042126379,
      "loss": 1.4186,
      "theoretical_loss": 3.4974189639482347,
      "tokens_seen": 1752104960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047378134403209626,
      "loss": 1.4263,
      "theoretical_loss": 3.497407862099232,
      "tokens_seen": 1752170496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737612838515546,
      "loss": 1.3414,
      "theoretical_loss": 3.497396760781723,
      "tokens_seen": 1752236032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047374122367101304,
      "loss": 1.3561,
      "theoretical_loss": 3.497385659995663,
      "tokens_seen": 1752301568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047372116349047146,
      "loss": 1.3877,
      "theoretical_loss": 3.497374559741006,
      "tokens_seen": 1752367104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737011033099298,
      "loss": 1.3865,
      "theoretical_loss": 3.4973634600177075,
      "tokens_seen": 1752432640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736810431293882,
      "loss": 1.5212,
      "theoretical_loss": 3.497352360825722,
      "tokens_seen": 1752498176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736609829488466,
      "loss": 1.4743,
      "theoretical_loss": 3.497341262165004,
      "tokens_seen": 1752563712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047364092276830496,
      "loss": 1.4386,
      "theoretical_loss": 3.4973301640355086,
      "tokens_seen": 1752629248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736208625877633,
      "loss": 1.3088,
      "theoretical_loss": 3.497319066437189,
      "tokens_seen": 1752694784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736008024072217,
      "loss": 1.378,
      "theoretical_loss": 3.497307969370002,
      "tokens_seen": 1752760320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047358074222668005,
      "loss": 1.3621,
      "theoretical_loss": 3.4972968728339016,
      "tokens_seen": 1752825856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047356068204613847,
      "loss": 1.344,
      "theoretical_loss": 3.497285776828842,
      "tokens_seen": 1752891392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047354062186559683,
      "loss": 1.4438,
      "theoretical_loss": 3.4972746813547784,
      "tokens_seen": 1752956928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735205616850552,
      "loss": 1.5445,
      "theoretical_loss": 3.4972635864116652,
      "tokens_seen": 1753022464
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047350050150451355,
      "loss": 1.3919,
      "theoretical_loss": 3.4972524919994576,
      "tokens_seen": 1753088000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047348044132397197,
      "loss": 1.29,
      "theoretical_loss": 3.49724139811811,
      "tokens_seen": 1753153536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047346038114343033,
      "loss": 1.3707,
      "theoretical_loss": 3.4972303047675775,
      "tokens_seen": 1753219072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734403209628887,
      "loss": 1.4446,
      "theoretical_loss": 3.4972192119478143,
      "tokens_seen": 1753284608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047342026078234706,
      "loss": 1.3172,
      "theoretical_loss": 3.497208119658775,
      "tokens_seen": 1753350144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734002006018054,
      "loss": 1.3996,
      "theoretical_loss": 3.4971970279004156,
      "tokens_seen": 1753415680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047338014042126383,
      "loss": 1.3655,
      "theoretical_loss": 3.4971859366726905,
      "tokens_seen": 1753481216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733600802407222,
      "loss": 1.4249,
      "theoretical_loss": 3.497174845975553,
      "tokens_seen": 1753546752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047334002006018056,
      "loss": 1.487,
      "theoretical_loss": 3.49716375580896,
      "tokens_seen": 1753612288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733199598796389,
      "loss": 1.4138,
      "theoretical_loss": 3.4971526661728647,
      "tokens_seen": 1753677824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047329989969909734,
      "loss": 1.3696,
      "theoretical_loss": 3.4971415770672225,
      "tokens_seen": 1753743360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732798395185557,
      "loss": 1.4004,
      "theoretical_loss": 3.4971304884919885,
      "tokens_seen": 1753808896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047325977933801406,
      "loss": 1.3973,
      "theoretical_loss": 3.4971194004471173,
      "tokens_seen": 1753874432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732397191574724,
      "loss": 1.4326,
      "theoretical_loss": 3.4971083129325633,
      "tokens_seen": 1753939968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732196589769308,
      "loss": 1.3029,
      "theoretical_loss": 3.4970972259482824,
      "tokens_seen": 1754005504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731995987963892,
      "loss": 1.4806,
      "theoretical_loss": 3.497086139494228,
      "tokens_seen": 1754071040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047317953861584757,
      "loss": 1.4118,
      "theoretical_loss": 3.497075053570356,
      "tokens_seen": 1754136576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047315947843530593,
      "loss": 1.4521,
      "theoretical_loss": 3.4970639681766205,
      "tokens_seen": 1754202112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731394182547643,
      "loss": 1.3827,
      "theoretical_loss": 3.4970528833129766,
      "tokens_seen": 1754267648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731193580742227,
      "loss": 1.4467,
      "theoretical_loss": 3.49704179897938,
      "tokens_seen": 1754333184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047309929789368107,
      "loss": 1.437,
      "theoretical_loss": 3.497030715175784,
      "tokens_seen": 1754398720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047307923771313943,
      "loss": 1.4981,
      "theoretical_loss": 3.4970196319021447,
      "tokens_seen": 1754464256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730591775325978,
      "loss": 1.4951,
      "theoretical_loss": 3.4970085491584166,
      "tokens_seen": 1754529792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047303911735205615,
      "loss": 1.4327,
      "theoretical_loss": 3.496997466944554,
      "tokens_seen": 1754595328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047301905717151457,
      "loss": 1.3146,
      "theoretical_loss": 3.496986385260513,
      "tokens_seen": 1754660864
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047299899699097293,
      "loss": 1.3945,
      "theoretical_loss": 3.496975304106247,
      "tokens_seen": 1754726400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729789368104313,
      "loss": 1.4097,
      "theoretical_loss": 3.4969642234817124,
      "tokens_seen": 1754791936
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 641116,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3436782360076904,
      "objective/train/theoretical_loss": 3.496953143386863,
      "objective/train/tokens_used": 125316576,
      "theoretical_loss": 3.496953143386863,
      "tokens_seen": 1754857472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047295887662988966,
      "loss": 1.373,
      "theoretical_loss": 3.496953143386863,
      "tokens_seen": 1754857472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729388164493481,
      "loss": 1.3959,
      "theoretical_loss": 3.4969420638216535,
      "tokens_seen": 1754923008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047291875626880644,
      "loss": 1.3779,
      "theoretical_loss": 3.49693098478604,
      "tokens_seen": 1754988544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728986960882648,
      "loss": 1.3608,
      "theoretical_loss": 3.496919906279977,
      "tokens_seen": 1755054080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047287863590772316,
      "loss": 1.4135,
      "theoretical_loss": 3.496908828303419,
      "tokens_seen": 1755119616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728585757271815,
      "loss": 1.3838,
      "theoretical_loss": 3.496897750856321,
      "tokens_seen": 1755185152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047283851554663994,
      "loss": 1.3669,
      "theoretical_loss": 3.496886673938638,
      "tokens_seen": 1755250688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728184553660983,
      "loss": 1.4353,
      "theoretical_loss": 3.496875597550325,
      "tokens_seen": 1755316224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047279839518555666,
      "loss": 1.3751,
      "theoretical_loss": 3.4968645216913368,
      "tokens_seen": 1755381760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472778335005015,
      "loss": 1.323,
      "theoretical_loss": 3.4968534463616283,
      "tokens_seen": 1755447296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047275827482447344,
      "loss": 1.4803,
      "theoretical_loss": 3.496842371561155,
      "tokens_seen": 1755512832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727382146439318,
      "loss": 1.3426,
      "theoretical_loss": 3.496831297289871,
      "tokens_seen": 1755578368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047271815446339017,
      "loss": 1.3724,
      "theoretical_loss": 3.4968202235477324,
      "tokens_seen": 1755643904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047269809428284853,
      "loss": 1.3883,
      "theoretical_loss": 3.496809150334693,
      "tokens_seen": 1755709440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047267803410230695,
      "loss": 1.3686,
      "theoretical_loss": 3.4967980776507086,
      "tokens_seen": 1755774976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726579739217653,
      "loss": 1.3677,
      "theoretical_loss": 3.4967870054957335,
      "tokens_seen": 1755840512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047263791374122367,
      "loss": 1.3445,
      "theoretical_loss": 3.4967759338697233,
      "tokens_seen": 1755906048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047261785356068203,
      "loss": 1.3832,
      "theoretical_loss": 3.496764862772632,
      "tokens_seen": 1755971584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725977933801404,
      "loss": 1.3187,
      "theoretical_loss": 3.496753792204416,
      "tokens_seen": 1756037120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725777331995988,
      "loss": 1.253,
      "theoretical_loss": 3.4967427221650293,
      "tokens_seen": 1756102656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725576730190572,
      "loss": 1.3705,
      "theoretical_loss": 3.4967316526544274,
      "tokens_seen": 1756168192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047253761283851554,
      "loss": 1.3044,
      "theoretical_loss": 3.496720583672565,
      "tokens_seen": 1756233728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725175526579739,
      "loss": 1.4567,
      "theoretical_loss": 3.496709515219397,
      "tokens_seen": 1756299264
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724974924774323,
      "loss": 1.4206,
      "theoretical_loss": 3.4966984472948788,
      "tokens_seen": 1756364800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724774322968907,
      "loss": 1.4351,
      "theoretical_loss": 3.4966873798989653,
      "tokens_seen": 1756430336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047245737211634904,
      "loss": 1.4001,
      "theoretical_loss": 3.4966763130316116,
      "tokens_seen": 1756495872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724373119358074,
      "loss": 1.4328,
      "theoretical_loss": 3.4966652466927726,
      "tokens_seen": 1756561408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047241725175526576,
      "loss": 1.4283,
      "theoretical_loss": 3.496654180882403,
      "tokens_seen": 1756626944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723971915747242,
      "loss": 1.3985,
      "theoretical_loss": 3.4966431156004587,
      "tokens_seen": 1756692480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047237713139418254,
      "loss": 1.2532,
      "theoretical_loss": 3.496632050846894,
      "tokens_seen": 1756758016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723570712136409,
      "loss": 1.4727,
      "theoretical_loss": 3.4966209866216644,
      "tokens_seen": 1756823552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047233701103309927,
      "loss": 1.4377,
      "theoretical_loss": 3.496609922924725,
      "tokens_seen": 1756889088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723169508525577,
      "loss": 1.4425,
      "theoretical_loss": 3.4965988597560305,
      "tokens_seen": 1756954624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047229689067201605,
      "loss": 1.4734,
      "theoretical_loss": 3.496587797115536,
      "tokens_seen": 1757020160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004722768304914744,
      "loss": 1.4572,
      "theoretical_loss": 3.496576735003197,
      "tokens_seen": 1757085696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047225677031093277,
      "loss": 1.4508,
      "theoretical_loss": 3.496565673418968,
      "tokens_seen": 1757151232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047223671013039113,
      "loss": 1.4301,
      "theoretical_loss": 3.4965546123628046,
      "tokens_seen": 1757216768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004722166499498496,
      "loss": 1.2344,
      "theoretical_loss": 3.4965435518346615,
      "tokens_seen": 1757282304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047219658976930797,
      "loss": 1.3642,
      "theoretical_loss": 3.496532491834494,
      "tokens_seen": 1757347840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047217652958876633,
      "loss": 1.3074,
      "theoretical_loss": 3.4965214323622575,
      "tokens_seen": 1757413376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721564694082247,
      "loss": 1.4659,
      "theoretical_loss": 3.496510373417907,
      "tokens_seen": 1757478912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721364092276831,
      "loss": 1.3674,
      "theoretical_loss": 3.4964993150013974,
      "tokens_seen": 1757544448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047211634904714147,
      "loss": 1.4215,
      "theoretical_loss": 3.4964882571126834,
      "tokens_seen": 1757609984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047209628886659983,
      "loss": 1.4338,
      "theoretical_loss": 3.496477199751721,
      "tokens_seen": 1757675520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720762286860582,
      "loss": 1.304,
      "theoretical_loss": 3.496466142918465,
      "tokens_seen": 1757741056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047205616850551656,
      "loss": 1.4645,
      "theoretical_loss": 3.4964550866128703,
      "tokens_seen": 1757806592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047203610832497497,
      "loss": 1.4003,
      "theoretical_loss": 3.4964440308348923,
      "tokens_seen": 1757872128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047201604814443333,
      "loss": 1.4411,
      "theoretical_loss": 3.4964329755844856,
      "tokens_seen": 1757937664
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719959879638917,
      "loss": 1.3551,
      "theoretical_loss": 3.4964219208616067,
      "tokens_seen": 1758003200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047197592778335006,
      "loss": 1.3761,
      "theoretical_loss": 3.49641086666621,
      "tokens_seen": 1758068736
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 642262,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.330387830734253,
      "objective/train/theoretical_loss": 3.4963998129982494,
      "objective/train/tokens_used": 128593376,
      "theoretical_loss": 3.4963998129982494,
      "tokens_seen": 1758134272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719558676028085,
      "loss": 1.3445,
      "theoretical_loss": 3.4963998129982494,
      "tokens_seen": 1758134272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047193580742226684,
      "loss": 1.3591,
      "theoretical_loss": 3.496388759857682,
      "tokens_seen": 1758199808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719157472417252,
      "loss": 1.3132,
      "theoretical_loss": 3.4963777072444624,
      "tokens_seen": 1758265344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047189568706118356,
      "loss": 1.2909,
      "theoretical_loss": 3.4963666551585453,
      "tokens_seen": 1758330880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718756268806419,
      "loss": 1.5081,
      "theoretical_loss": 3.496355603599887,
      "tokens_seen": 1758396416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047185556670010034,
      "loss": 1.288,
      "theoretical_loss": 3.496344552568441,
      "tokens_seen": 1758461952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718355065195587,
      "loss": 1.3416,
      "theoretical_loss": 3.4963335020641635,
      "tokens_seen": 1758527488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047181544633901706,
      "loss": 1.4727,
      "theoretical_loss": 3.4963224520870098,
      "tokens_seen": 1758593024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717953861584754,
      "loss": 1.4002,
      "theoretical_loss": 3.496311402636935,
      "tokens_seen": 1758658560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047177532597793384,
      "loss": 1.4526,
      "theoretical_loss": 3.4963003537138944,
      "tokens_seen": 1758724096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717552657973922,
      "loss": 1.473,
      "theoretical_loss": 3.4962893053178425,
      "tokens_seen": 1758789632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047173520561685057,
      "loss": 1.4001,
      "theoretical_loss": 3.4962782574487354,
      "tokens_seen": 1758855168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047171514543630893,
      "loss": 1.5418,
      "theoretical_loss": 3.496267210106528,
      "tokens_seen": 1758920704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047169508525576735,
      "loss": 1.4355,
      "theoretical_loss": 3.496256163291176,
      "tokens_seen": 1758986240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716750250752257,
      "loss": 1.4155,
      "theoretical_loss": 3.496245117002634,
      "tokens_seen": 1759051776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047165496489468407,
      "loss": 1.3922,
      "theoretical_loss": 3.496234071240857,
      "tokens_seen": 1759117312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047163490471414243,
      "loss": 1.3249,
      "theoretical_loss": 3.4962230260058016,
      "tokens_seen": 1759182848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716148445336008,
      "loss": 1.3263,
      "theoretical_loss": 3.4962119812974217,
      "tokens_seen": 1759248384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715947843530592,
      "loss": 1.3491,
      "theoretical_loss": 3.496200937115673,
      "tokens_seen": 1759313920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715747241725176,
      "loss": 1.384,
      "theoretical_loss": 3.496189893460511,
      "tokens_seen": 1759379456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047155466399197594,
      "loss": 1.3865,
      "theoretical_loss": 3.4961788503318907,
      "tokens_seen": 1759444992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715346038114343,
      "loss": 1.3548,
      "theoretical_loss": 3.4961678077297673,
      "tokens_seen": 1759510528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715145436308927,
      "loss": 1.4088,
      "theoretical_loss": 3.4961567656540966,
      "tokens_seen": 1759576064
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714944834503511,
      "loss": 1.2849,
      "theoretical_loss": 3.4961457241048337,
      "tokens_seen": 1759641600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047147442326980944,
      "loss": 1.385,
      "theoretical_loss": 3.4961346830819338,
      "tokens_seen": 1759707136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714543630892678,
      "loss": 1.34,
      "theoretical_loss": 3.496123642585352,
      "tokens_seen": 1759772672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047143430290872616,
      "loss": 1.3061,
      "theoretical_loss": 3.4961126026150433,
      "tokens_seen": 1759838208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714142427281846,
      "loss": 1.4748,
      "theoretical_loss": 3.4961015631709644,
      "tokens_seen": 1759903744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047139418254764294,
      "loss": 1.4464,
      "theoretical_loss": 3.496090524253069,
      "tokens_seen": 1759969280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713741223671013,
      "loss": 1.5164,
      "theoretical_loss": 3.496079485861314,
      "tokens_seen": 1760034816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047135406218655967,
      "loss": 1.4403,
      "theoretical_loss": 3.496068447995653,
      "tokens_seen": 1760100352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713340020060181,
      "loss": 1.4814,
      "theoretical_loss": 3.496057410656043,
      "tokens_seen": 1760165888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047131394182547645,
      "loss": 1.363,
      "theoretical_loss": 3.4960463738424385,
      "tokens_seen": 1760231424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712938816449348,
      "loss": 1.2892,
      "theoretical_loss": 3.496035337554795,
      "tokens_seen": 1760296960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047127382146439317,
      "loss": 1.352,
      "theoretical_loss": 3.4960243017930672,
      "tokens_seen": 1760362496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047125376128385153,
      "loss": 1.4501,
      "theoretical_loss": 3.4960132665572115,
      "tokens_seen": 1760428032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047123370110330995,
      "loss": 1.4287,
      "theoretical_loss": 3.496002231847182,
      "tokens_seen": 1760493568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712136409227683,
      "loss": 1.3261,
      "theoretical_loss": 3.495991197662936,
      "tokens_seen": 1760559104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711935807422267,
      "loss": 1.398,
      "theoretical_loss": 3.4959801640044272,
      "tokens_seen": 1760624640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047117352056168504,
      "loss": 1.421,
      "theoretical_loss": 3.495969130871612,
      "tokens_seen": 1760690176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047115346038114345,
      "loss": 1.2906,
      "theoretical_loss": 3.4959580982644445,
      "tokens_seen": 1760755712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711334002006018,
      "loss": 1.3436,
      "theoretical_loss": 3.4959470661828815,
      "tokens_seen": 1760821248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711133400200602,
      "loss": 1.4988,
      "theoretical_loss": 3.495936034626878,
      "tokens_seen": 1760886784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047109327983951854,
      "loss": 1.4439,
      "theoretical_loss": 3.4959250035963887,
      "tokens_seen": 1760952320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710732196589769,
      "loss": 1.3507,
      "theoretical_loss": 3.49591397309137,
      "tokens_seen": 1761017856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710531594784353,
      "loss": 1.3673,
      "theoretical_loss": 3.495902943111777,
      "tokens_seen": 1761083392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710330992978937,
      "loss": 1.3384,
      "theoretical_loss": 3.4958919136575646,
      "tokens_seen": 1761148928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047101303911735204,
      "loss": 1.3313,
      "theoretical_loss": 3.4958808847286886,
      "tokens_seen": 1761214464
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709929789368104,
      "loss": 1.4544,
      "theoretical_loss": 3.495869856325105,
      "tokens_seen": 1761280000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709729187562688,
      "loss": 1.3167,
      "theoretical_loss": 3.4958588284467678,
      "tokens_seen": 1761345536
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 643454,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3653559684753418,
      "objective/train/theoretical_loss": 3.4958478010936336,
      "objective/train/tokens_used": 131870176,
      "theoretical_loss": 3.4958478010936336,
      "tokens_seen": 1761411072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709528585757272,
      "loss": 1.3219,
      "theoretical_loss": 3.4958478010936336,
      "tokens_seen": 1761411072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047093279839518555,
      "loss": 1.4581,
      "theoretical_loss": 3.4958367742656575,
      "tokens_seen": 1761476608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709127382146439,
      "loss": 1.4421,
      "theoretical_loss": 3.4958257479627957,
      "tokens_seen": 1761542144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047089267803410227,
      "loss": 1.3858,
      "theoretical_loss": 3.495814722185002,
      "tokens_seen": 1761607680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708726178535607,
      "loss": 1.4404,
      "theoretical_loss": 3.4958036969322337,
      "tokens_seen": 1761673216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047085255767301905,
      "loss": 1.3771,
      "theoretical_loss": 3.495792672204445,
      "tokens_seen": 1761738752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708324974924774,
      "loss": 1.4397,
      "theoretical_loss": 3.495781648001592,
      "tokens_seen": 1761804288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708124373119358,
      "loss": 1.3939,
      "theoretical_loss": 3.4957706243236295,
      "tokens_seen": 1761869824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707923771313942,
      "loss": 1.2862,
      "theoretical_loss": 3.495759601170514,
      "tokens_seen": 1761935360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047077231695085255,
      "loss": 1.501,
      "theoretical_loss": 3.4957485785422,
      "tokens_seen": 1762000896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707522567703109,
      "loss": 1.5213,
      "theoretical_loss": 3.4957375564386437,
      "tokens_seen": 1762066432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707321965897693,
      "loss": 1.3949,
      "theoretical_loss": 3.4957265348598,
      "tokens_seen": 1762131968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047071213640922775,
      "loss": 1.3351,
      "theoretical_loss": 3.495715513805625,
      "tokens_seen": 1762197504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706920762286861,
      "loss": 1.3716,
      "theoretical_loss": 3.495704493276074,
      "tokens_seen": 1762263040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047067201604814447,
      "loss": 1.4119,
      "theoretical_loss": 3.4956934732711025,
      "tokens_seen": 1762328576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047065195586760283,
      "loss": 1.4479,
      "theoretical_loss": 3.4956824537906654,
      "tokens_seen": 1762394112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706318956870612,
      "loss": 1.5315,
      "theoretical_loss": 3.4956714348347195,
      "tokens_seen": 1762459648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706118355065196,
      "loss": 1.345,
      "theoretical_loss": 3.4956604164032195,
      "tokens_seen": 1762525184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470591775325978,
      "loss": 1.4292,
      "theoretical_loss": 3.4956493984961208,
      "tokens_seen": 1762590720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047057171514543634,
      "loss": 1.4123,
      "theoretical_loss": 3.4956383811133795,
      "tokens_seen": 1762656256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705516549648947,
      "loss": 1.3416,
      "theoretical_loss": 3.4956273642549514,
      "tokens_seen": 1762721792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705315947843531,
      "loss": 1.2949,
      "theoretical_loss": 3.495616347920791,
      "tokens_seen": 1762787328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705115346038115,
      "loss": 1.6214,
      "theoretical_loss": 3.4956053321108547,
      "tokens_seen": 1762852864
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047049147442326984,
      "loss": 1.334,
      "theoretical_loss": 3.4955943168250974,
      "tokens_seen": 1762918400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704714142427282,
      "loss": 1.4307,
      "theoretical_loss": 3.4955833020634755,
      "tokens_seen": 1762983936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047045135406218656,
      "loss": 1.4839,
      "theoretical_loss": 3.4955722878259436,
      "tokens_seen": 1763049472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470431293881645,
      "loss": 1.4356,
      "theoretical_loss": 3.495561274112458,
      "tokens_seen": 1763115008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047041123370110334,
      "loss": 1.3556,
      "theoretical_loss": 3.4955502609229745,
      "tokens_seen": 1763180544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703911735205617,
      "loss": 1.3959,
      "theoretical_loss": 3.4955392482574474,
      "tokens_seen": 1763246080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047037111334002007,
      "loss": 1.3551,
      "theoretical_loss": 3.4955282361158337,
      "tokens_seen": 1763311616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703510531594785,
      "loss": 1.4848,
      "theoretical_loss": 3.495517224498089,
      "tokens_seen": 1763377152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047033099297893685,
      "loss": 1.488,
      "theoretical_loss": 3.495506213404168,
      "tokens_seen": 1763442688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703109327983952,
      "loss": 1.3503,
      "theoretical_loss": 3.495495202834027,
      "tokens_seen": 1763508224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047029087261785357,
      "loss": 1.5253,
      "theoretical_loss": 3.495484192787621,
      "tokens_seen": 1763573760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047027081243731193,
      "loss": 1.6191,
      "theoretical_loss": 3.495473183264906,
      "tokens_seen": 1763639296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047025075225677035,
      "loss": 1.4056,
      "theoretical_loss": 3.495462174265838,
      "tokens_seen": 1763704832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702306920762287,
      "loss": 1.3762,
      "theoretical_loss": 3.4954511657903717,
      "tokens_seen": 1763770368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702106318956871,
      "loss": 1.3131,
      "theoretical_loss": 3.4954401578384635,
      "tokens_seen": 1763835904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047019057171514544,
      "loss": 1.274,
      "theoretical_loss": 3.495429150410069,
      "tokens_seen": 1763901440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047017051153460385,
      "loss": 1.4417,
      "theoretical_loss": 3.4954181435051437,
      "tokens_seen": 1763966976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701504513540622,
      "loss": 1.4518,
      "theoretical_loss": 3.495407137123643,
      "tokens_seen": 1764032512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701303911735206,
      "loss": 1.3252,
      "theoretical_loss": 3.495396131265523,
      "tokens_seen": 1764098048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047011033099297894,
      "loss": 1.3624,
      "theoretical_loss": 3.495385125930739,
      "tokens_seen": 1764163584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700902708124373,
      "loss": 1.4579,
      "theoretical_loss": 3.495374121119247,
      "tokens_seen": 1764229120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700702106318957,
      "loss": 1.3647,
      "theoretical_loss": 3.4953631168310024,
      "tokens_seen": 1764294656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700501504513541,
      "loss": 1.4022,
      "theoretical_loss": 3.4953521130659615,
      "tokens_seen": 1764360192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047003009027081244,
      "loss": 1.3839,
      "theoretical_loss": 3.4953411098240785,
      "tokens_seen": 1764425728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700100300902708,
      "loss": 1.4308,
      "theoretical_loss": 3.4953301071053113,
      "tokens_seen": 1764491264
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699899699097292,
      "loss": 1.3421,
      "theoretical_loss": 3.4953191049096137,
      "tokens_seen": 1764556800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699699097291876,
      "loss": 1.356,
      "theoretical_loss": 3.495308103236942,
      "tokens_seen": 1764622336
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 644554,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3437608480453491,
      "objective/train/theoretical_loss": 3.4952971020872527,
      "objective/train/tokens_used": 135146976,
      "theoretical_loss": 3.4952971020872527,
      "tokens_seen": 1764687872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046994984954864595,
      "loss": 1.408,
      "theoretical_loss": 3.4952971020872527,
      "tokens_seen": 1764687872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699297893681043,
      "loss": 1.4481,
      "theoretical_loss": 3.4952861014605006,
      "tokens_seen": 1764753408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046990972918756267,
      "loss": 1.4172,
      "theoretical_loss": 3.4952751013566417,
      "tokens_seen": 1764818944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698896690070211,
      "loss": 1.4879,
      "theoretical_loss": 3.4952641017756316,
      "tokens_seen": 1764884480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046986960882647945,
      "loss": 1.3973,
      "theoretical_loss": 3.495253102717426,
      "tokens_seen": 1764950016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698495486459378,
      "loss": 1.4776,
      "theoretical_loss": 3.495242104181981,
      "tokens_seen": 1765015552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698294884653962,
      "loss": 1.4463,
      "theoretical_loss": 3.495231106169252,
      "tokens_seen": 1765081088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698094282848546,
      "loss": 1.3859,
      "theoretical_loss": 3.495220108679195,
      "tokens_seen": 1765146624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046978936810431295,
      "loss": 1.3913,
      "theoretical_loss": 3.4952091117117656,
      "tokens_seen": 1765212160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697693079237713,
      "loss": 1.3257,
      "theoretical_loss": 3.49519811526692,
      "tokens_seen": 1765277696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697492477432297,
      "loss": 1.3308,
      "theoretical_loss": 3.495187119344613,
      "tokens_seen": 1765343232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046972918756268804,
      "loss": 1.4673,
      "theoretical_loss": 3.495176123944801,
      "tokens_seen": 1765408768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046970912738214646,
      "loss": 1.4629,
      "theoretical_loss": 3.4951651290674395,
      "tokens_seen": 1765474304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696890672016048,
      "loss": 1.4301,
      "theoretical_loss": 3.495154134712485,
      "tokens_seen": 1765539840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696690070210632,
      "loss": 1.3324,
      "theoretical_loss": 3.4951431408798928,
      "tokens_seen": 1765605376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046964894684052154,
      "loss": 1.2896,
      "theoretical_loss": 3.4951321475696187,
      "tokens_seen": 1765670912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046962888665997996,
      "loss": 1.448,
      "theoretical_loss": 3.495121154781618,
      "tokens_seen": 1765736448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696088264794383,
      "loss": 1.3542,
      "theoretical_loss": 3.4951101625158474,
      "tokens_seen": 1765801984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695887662988967,
      "loss": 1.3436,
      "theoretical_loss": 3.495099170772262,
      "tokens_seen": 1765867520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046956870611835505,
      "loss": 1.3527,
      "theoretical_loss": 3.4950881795508177,
      "tokens_seen": 1765933056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046954864593781346,
      "loss": 1.5647,
      "theoretical_loss": 3.4950771888514716,
      "tokens_seen": 1765998592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695285857572718,
      "loss": 1.2566,
      "theoretical_loss": 3.495066198674177,
      "tokens_seen": 1766064128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695085255767302,
      "loss": 1.4403,
      "theoretical_loss": 3.495055209018892,
      "tokens_seen": 1766129664
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046948846539618855,
      "loss": 1.3159,
      "theoretical_loss": 3.495044219885572,
      "tokens_seen": 1766195200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694684052156469,
      "loss": 1.355,
      "theoretical_loss": 3.4950332312741716,
      "tokens_seen": 1766260736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694483450351053,
      "loss": 1.4411,
      "theoretical_loss": 3.495022243184648,
      "tokens_seen": 1766326272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694282848545637,
      "loss": 1.3657,
      "theoretical_loss": 3.495011255616957,
      "tokens_seen": 1766391808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046940822467402205,
      "loss": 1.3799,
      "theoretical_loss": 3.4950002685710535,
      "tokens_seen": 1766457344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693881644934804,
      "loss": 1.3737,
      "theoretical_loss": 3.4949892820468937,
      "tokens_seen": 1766522880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046936810431293883,
      "loss": 1.3805,
      "theoretical_loss": 3.494978296044434,
      "tokens_seen": 1766588416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693480441323972,
      "loss": 1.3766,
      "theoretical_loss": 3.49496731056363,
      "tokens_seen": 1766653952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046932798395185555,
      "loss": 1.6006,
      "theoretical_loss": 3.494956325604437,
      "tokens_seen": 1766719488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693079237713139,
      "loss": 1.3969,
      "theoretical_loss": 3.4949453411668117,
      "tokens_seen": 1766785024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692878635907723,
      "loss": 1.2856,
      "theoretical_loss": 3.4949343572507097,
      "tokens_seen": 1766850560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692678034102307,
      "loss": 1.3832,
      "theoretical_loss": 3.4949233738560874,
      "tokens_seen": 1766916096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046924774322968906,
      "loss": 1.3627,
      "theoretical_loss": 3.4949123909828996,
      "tokens_seen": 1766981632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692276830491474,
      "loss": 1.4127,
      "theoretical_loss": 3.4949014086311028,
      "tokens_seen": 1767047168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046920762286860584,
      "loss": 1.4154,
      "theoretical_loss": 3.4948904268006533,
      "tokens_seen": 1767112704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046918756268806425,
      "loss": 1.3805,
      "theoretical_loss": 3.494879445491506,
      "tokens_seen": 1767178240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691675025075226,
      "loss": 1.4497,
      "theoretical_loss": 3.494868464703618,
      "tokens_seen": 1767243776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469147442326981,
      "loss": 1.3187,
      "theoretical_loss": 3.4948574844369444,
      "tokens_seen": 1767309312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046912738214643934,
      "loss": 1.3534,
      "theoretical_loss": 3.4948465046914414,
      "tokens_seen": 1767374848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691073219658977,
      "loss": 1.3963,
      "theoretical_loss": 3.494835525467065,
      "tokens_seen": 1767440384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690872617853561,
      "loss": 1.3614,
      "theoretical_loss": 3.494824546763771,
      "tokens_seen": 1767505920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690672016048145,
      "loss": 1.3758,
      "theoretical_loss": 3.494813568581516,
      "tokens_seen": 1767571456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046904714142427284,
      "loss": 1.3334,
      "theoretical_loss": 3.494802590920255,
      "tokens_seen": 1767636992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690270812437312,
      "loss": 1.3513,
      "theoretical_loss": 3.494791613779944,
      "tokens_seen": 1767702528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690070210631896,
      "loss": 1.2967,
      "theoretical_loss": 3.49478063716054,
      "tokens_seen": 1767768064
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468986960882648,
      "loss": 1.6072,
      "theoretical_loss": 3.494769661061998,
      "tokens_seen": 1767833600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046896690070210635,
      "loss": 1.3412,
      "theoretical_loss": 3.494758685484274,
      "tokens_seen": 1767899136
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 645687,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4466071128845215,
      "objective/train/theoretical_loss": 3.4947477104273243,
      "objective/train/tokens_used": 138423776,
      "theoretical_loss": 3.4947477104273243,
      "tokens_seen": 1767964672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689468405215647,
      "loss": 1.3827,
      "theoretical_loss": 3.4947477104273243,
      "tokens_seen": 1767964672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046892678034102307,
      "loss": 1.4484,
      "theoretical_loss": 3.494736735891105,
      "tokens_seen": 1768030208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689067201604815,
      "loss": 1.3213,
      "theoretical_loss": 3.494725761875572,
      "tokens_seen": 1768095744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046888665997993985,
      "loss": 1.3091,
      "theoretical_loss": 3.494714788380681,
      "tokens_seen": 1768161280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688665997993982,
      "loss": 1.3501,
      "theoretical_loss": 3.494703815406388,
      "tokens_seen": 1768226816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688465396188566,
      "loss": 1.3446,
      "theoretical_loss": 3.4946928429526496,
      "tokens_seen": 1768292352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468826479438315,
      "loss": 1.3405,
      "theoretical_loss": 3.494681871019422,
      "tokens_seen": 1768357888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046880641925777335,
      "loss": 1.4153,
      "theoretical_loss": 3.4946708996066596,
      "tokens_seen": 1768423424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687863590772317,
      "loss": 1.4501,
      "theoretical_loss": 3.49465992871432,
      "tokens_seen": 1768488960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687662988966901,
      "loss": 1.4606,
      "theoretical_loss": 3.494648958342359,
      "tokens_seen": 1768554496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046874623871614844,
      "loss": 1.3804,
      "theoretical_loss": 3.494637988490732,
      "tokens_seen": 1768620032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046872617853560686,
      "loss": 1.474,
      "theoretical_loss": 3.4946270191593958,
      "tokens_seen": 1768685568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687061183550652,
      "loss": 1.4744,
      "theoretical_loss": 3.4946160503483057,
      "tokens_seen": 1768751104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686860581745236,
      "loss": 1.3864,
      "theoretical_loss": 3.494605082057418,
      "tokens_seen": 1768816640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046866599799398194,
      "loss": 1.3426,
      "theoretical_loss": 3.4945941142866888,
      "tokens_seen": 1768882176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046864593781344036,
      "loss": 1.3869,
      "theoretical_loss": 3.494583147036075,
      "tokens_seen": 1768947712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686258776328987,
      "loss": 1.3047,
      "theoretical_loss": 3.494572180305531,
      "tokens_seen": 1769013248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686058174523571,
      "loss": 1.3638,
      "theoretical_loss": 3.494561214095014,
      "tokens_seen": 1769078784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046858575727181545,
      "loss": 1.2741,
      "theoretical_loss": 3.4945502484044804,
      "tokens_seen": 1769144320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685656970912738,
      "loss": 1.3958,
      "theoretical_loss": 3.494539283233885,
      "tokens_seen": 1769209856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685456369107322,
      "loss": 1.38,
      "theoretical_loss": 3.4945283185831846,
      "tokens_seen": 1769275392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685255767301906,
      "loss": 1.4113,
      "theoretical_loss": 3.4945173544523356,
      "tokens_seen": 1769340928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046850551654964895,
      "loss": 1.3416,
      "theoretical_loss": 3.494506390841294,
      "tokens_seen": 1769406464
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684854563691073,
      "loss": 1.4602,
      "theoretical_loss": 3.4944954277500155,
      "tokens_seen": 1769472000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046846539618856573,
      "loss": 1.4125,
      "theoretical_loss": 3.4944844651784566,
      "tokens_seen": 1769537536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684453360080241,
      "loss": 1.426,
      "theoretical_loss": 3.4944735031265726,
      "tokens_seen": 1769603072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046842527582748245,
      "loss": 1.3728,
      "theoretical_loss": 3.494462541594321,
      "tokens_seen": 1769668608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684052156469408,
      "loss": 1.3692,
      "theoretical_loss": 3.4944515805816563,
      "tokens_seen": 1769734144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046838515546639923,
      "loss": 1.3873,
      "theoretical_loss": 3.494440620088536,
      "tokens_seen": 1769799680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683650952858576,
      "loss": 1.3983,
      "theoretical_loss": 3.494429660114916,
      "tokens_seen": 1769865216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046834503510531595,
      "loss": 1.2393,
      "theoretical_loss": 3.494418700660752,
      "tokens_seen": 1769930752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683249749247743,
      "loss": 1.403,
      "theoretical_loss": 3.4944077417260004,
      "tokens_seen": 1769996288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683049147442327,
      "loss": 1.3028,
      "theoretical_loss": 3.4943967833106173,
      "tokens_seen": 1770061824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682848545636911,
      "loss": 1.3154,
      "theoretical_loss": 3.4943858254145583,
      "tokens_seen": 1770127360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046826479438314946,
      "loss": 1.4234,
      "theoretical_loss": 3.494374868037781,
      "tokens_seen": 1770192896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682447342026078,
      "loss": 1.2604,
      "theoretical_loss": 3.49436391118024,
      "tokens_seen": 1770258432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682246740220662,
      "loss": 1.3015,
      "theoretical_loss": 3.4943529548418923,
      "tokens_seen": 1770323968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682046138415246,
      "loss": 1.4155,
      "theoretical_loss": 3.4943419990226943,
      "tokens_seen": 1770389504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046818455366098296,
      "loss": 1.3454,
      "theoretical_loss": 3.4943310437226014,
      "tokens_seen": 1770455040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681644934804413,
      "loss": 1.4116,
      "theoretical_loss": 3.49432008894157,
      "tokens_seen": 1770520576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681444332998997,
      "loss": 1.3459,
      "theoretical_loss": 3.4943091346795567,
      "tokens_seen": 1770586112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046812437311935805,
      "loss": 1.393,
      "theoretical_loss": 3.494298180936518,
      "tokens_seen": 1770651648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046810431293881646,
      "loss": 1.4032,
      "theoretical_loss": 3.4942872277124093,
      "tokens_seen": 1770717184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680842527582748,
      "loss": 1.4136,
      "theoretical_loss": 3.494276275007187,
      "tokens_seen": 1770782720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680641925777332,
      "loss": 1.3419,
      "theoretical_loss": 3.494265322820808,
      "tokens_seen": 1770848256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046804413239719155,
      "loss": 1.3964,
      "theoretical_loss": 3.494254371153227,
      "tokens_seen": 1770913792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046802407221664997,
      "loss": 1.3462,
      "theoretical_loss": 3.4942434200044015,
      "tokens_seen": 1770979328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046800401203610833,
      "loss": 1.3377,
      "theoretical_loss": 3.494232469374287,
      "tokens_seen": 1771044864
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679839518555667,
      "loss": 1.3578,
      "theoretical_loss": 3.494221519262841,
      "tokens_seen": 1771110400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046796389167502505,
      "loss": 1.3526,
      "theoretical_loss": 3.4942105696700185,
      "tokens_seen": 1771175936
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 646815,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4491500854492188,
      "objective/train/theoretical_loss": 3.4941996205957766,
      "objective/train/tokens_used": 141700576,
      "theoretical_loss": 3.4941996205957766,
      "tokens_seen": 1771241472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679438314944834,
      "loss": 1.4018,
      "theoretical_loss": 3.4941996205957766,
      "tokens_seen": 1771241472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046792377131394183,
      "loss": 1.3682,
      "theoretical_loss": 3.4941886720400706,
      "tokens_seen": 1771307008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679037111334002,
      "loss": 1.3883,
      "theoretical_loss": 3.4941777240028573,
      "tokens_seen": 1771372544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046788365095285856,
      "loss": 1.3136,
      "theoretical_loss": 3.494166776484093,
      "tokens_seen": 1771438080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678635907723169,
      "loss": 1.3225,
      "theoretical_loss": 3.494155829483734,
      "tokens_seen": 1771503616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046784353059177534,
      "loss": 1.4145,
      "theoretical_loss": 3.494144883001736,
      "tokens_seen": 1771569152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678234704112337,
      "loss": 1.3247,
      "theoretical_loss": 3.4941339370380566,
      "tokens_seen": 1771634688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046780341023069206,
      "loss": 1.307,
      "theoretical_loss": 3.4941229915926506,
      "tokens_seen": 1771700224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677833500501504,
      "loss": 1.4132,
      "theoretical_loss": 3.4941120466654745,
      "tokens_seen": 1771765760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677632898696088,
      "loss": 1.4054,
      "theoretical_loss": 3.4941011022564856,
      "tokens_seen": 1771831296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677432296890672,
      "loss": 1.3943,
      "theoretical_loss": 3.4940901583656396,
      "tokens_seen": 1771896832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046772316950852556,
      "loss": 1.4279,
      "theoretical_loss": 3.494079214992893,
      "tokens_seen": 1771962368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467703109327984,
      "loss": 1.2983,
      "theoretical_loss": 3.4940682721382017,
      "tokens_seen": 1772027904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046768304914744234,
      "loss": 1.2707,
      "theoretical_loss": 3.4940573298015223,
      "tokens_seen": 1772093440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046766298896690076,
      "loss": 1.3194,
      "theoretical_loss": 3.494046387982811,
      "tokens_seen": 1772158976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676429287863591,
      "loss": 1.4656,
      "theoretical_loss": 3.494035446682024,
      "tokens_seen": 1772224512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676228686058175,
      "loss": 1.493,
      "theoretical_loss": 3.494024505899118,
      "tokens_seen": 1772290048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046760280842527585,
      "loss": 1.3873,
      "theoretical_loss": 3.4940135656340496,
      "tokens_seen": 1772355584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675827482447342,
      "loss": 1.3803,
      "theoretical_loss": 3.4940026258867745,
      "tokens_seen": 1772421120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675626880641926,
      "loss": 1.4669,
      "theoretical_loss": 3.4939916866572487,
      "tokens_seen": 1772486656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467542627883651,
      "loss": 1.3403,
      "theoretical_loss": 3.4939807479454297,
      "tokens_seen": 1772552192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046752256770310935,
      "loss": 1.353,
      "theoretical_loss": 3.493969809751273,
      "tokens_seen": 1772617728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675025075225677,
      "loss": 1.3699,
      "theoretical_loss": 3.4939588720747357,
      "tokens_seen": 1772683264
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046748244734202613,
      "loss": 1.5135,
      "theoretical_loss": 3.493947934915773,
      "tokens_seen": 1772748800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674623871614845,
      "loss": 1.2624,
      "theoretical_loss": 3.493936998274343,
      "tokens_seen": 1772814336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046744232698094285,
      "loss": 1.3377,
      "theoretical_loss": 3.4939260621504005,
      "tokens_seen": 1772879872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674222668004012,
      "loss": 1.3574,
      "theoretical_loss": 3.493915126543902,
      "tokens_seen": 1772945408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046740220661985963,
      "loss": 1.3531,
      "theoretical_loss": 3.493904191454805,
      "tokens_seen": 1773010944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467382146439318,
      "loss": 1.3937,
      "theoretical_loss": 3.493893256883065,
      "tokens_seen": 1773076480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046736208625877636,
      "loss": 1.4041,
      "theoretical_loss": 3.4938823228286386,
      "tokens_seen": 1773142016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673420260782347,
      "loss": 1.3208,
      "theoretical_loss": 3.493871389291482,
      "tokens_seen": 1773207552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673219658976931,
      "loss": 1.3034,
      "theoretical_loss": 3.4938604562715527,
      "tokens_seen": 1773273088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673019057171515,
      "loss": 1.4102,
      "theoretical_loss": 3.4938495237688056,
      "tokens_seen": 1773338624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046728184553660986,
      "loss": 1.4702,
      "theoretical_loss": 3.4938385917831978,
      "tokens_seen": 1773404160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672617853560682,
      "loss": 1.3546,
      "theoretical_loss": 3.493827660314686,
      "tokens_seen": 1773469696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672417251755266,
      "loss": 1.4124,
      "theoretical_loss": 3.4938167293632265,
      "tokens_seen": 1773535232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467221664994985,
      "loss": 1.345,
      "theoretical_loss": 3.493805798928775,
      "tokens_seen": 1773600768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046720160481444336,
      "loss": 1.3367,
      "theoretical_loss": 3.493794869011289,
      "tokens_seen": 1773666304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004671815446339017,
      "loss": 1.4365,
      "theoretical_loss": 3.493783939610725,
      "tokens_seen": 1773731840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671614844533601,
      "loss": 1.518,
      "theoretical_loss": 3.493773010727038,
      "tokens_seen": 1773797376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046714142427281845,
      "loss": 1.5115,
      "theoretical_loss": 3.493762082360186,
      "tokens_seen": 1773862912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046712136409227686,
      "loss": 1.3766,
      "theoretical_loss": 3.4937511545101243,
      "tokens_seen": 1773928448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671013039117352,
      "loss": 1.532,
      "theoretical_loss": 3.4937402271768105,
      "tokens_seen": 1773993984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670812437311936,
      "loss": 1.3404,
      "theoretical_loss": 3.493729300360201,
      "tokens_seen": 1774059520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046706118355065195,
      "loss": 1.324,
      "theoretical_loss": 3.4937183740602507,
      "tokens_seen": 1774125056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046704112337011037,
      "loss": 1.4054,
      "theoretical_loss": 3.4937074482769175,
      "tokens_seen": 1774190592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046702106318956873,
      "loss": 1.3717,
      "theoretical_loss": 3.4936965230101578,
      "tokens_seen": 1774256128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670010030090271,
      "loss": 1.3153,
      "theoretical_loss": 3.493685598259928,
      "tokens_seen": 1774321664
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046698094282848545,
      "loss": 1.3947,
      "theoretical_loss": 3.4936746740261846,
      "tokens_seen": 1774387200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669608826479438,
      "loss": 1.4154,
      "theoretical_loss": 3.4936637503088837,
      "tokens_seen": 1774452736
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 647978,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3683605194091797,
      "objective/train/theoretical_loss": 3.493652827107982,
      "objective/train/tokens_used": 144977376,
      "theoretical_loss": 3.493652827107982,
      "tokens_seen": 1774518272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046694082246740223,
      "loss": 1.393,
      "theoretical_loss": 3.493652827107982,
      "tokens_seen": 1774518272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669207622868606,
      "loss": 1.3967,
      "theoretical_loss": 3.493641904423437,
      "tokens_seen": 1774583808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046690070210631896,
      "loss": 1.3406,
      "theoretical_loss": 3.493630982255203,
      "tokens_seen": 1774649344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668806419257773,
      "loss": 1.2683,
      "theoretical_loss": 3.493620060603239,
      "tokens_seen": 1774714880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046686058174523574,
      "loss": 1.2875,
      "theoretical_loss": 3.4936091394674995,
      "tokens_seen": 1774780416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668405215646941,
      "loss": 1.4032,
      "theoretical_loss": 3.4935982188479424,
      "tokens_seen": 1774845952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046682046138415246,
      "loss": 1.4913,
      "theoretical_loss": 3.493587298744524,
      "tokens_seen": 1774911488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668004012036108,
      "loss": 1.3395,
      "theoretical_loss": 3.4935763791572003,
      "tokens_seen": 1774977024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667803410230692,
      "loss": 1.3878,
      "theoretical_loss": 3.4935654600859283,
      "tokens_seen": 1775042560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667602808425276,
      "loss": 1.3406,
      "theoretical_loss": 3.4935545415306644,
      "tokens_seen": 1775108096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046674022066198596,
      "loss": 1.402,
      "theoretical_loss": 3.4935436234913655,
      "tokens_seen": 1775173632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667201604814443,
      "loss": 1.3797,
      "theoretical_loss": 3.4935327059679873,
      "tokens_seen": 1775239168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667001003009027,
      "loss": 1.2969,
      "theoretical_loss": 3.4935217889604875,
      "tokens_seen": 1775304704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666800401203611,
      "loss": 1.3547,
      "theoretical_loss": 3.493510872468822,
      "tokens_seen": 1775370240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046665997993981947,
      "loss": 1.3587,
      "theoretical_loss": 3.493499956492948,
      "tokens_seen": 1775435776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046663991975927783,
      "loss": 1.3586,
      "theoretical_loss": 3.493489041032821,
      "tokens_seen": 1775501312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666198595787362,
      "loss": 1.409,
      "theoretical_loss": 3.493478126088398,
      "tokens_seen": 1775566848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046659979939819455,
      "loss": 1.4863,
      "theoretical_loss": 3.4934672116596364,
      "tokens_seen": 1775632384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046657973921765297,
      "loss": 1.3135,
      "theoretical_loss": 3.4934562977464925,
      "tokens_seen": 1775697920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046655967903711133,
      "loss": 1.3717,
      "theoretical_loss": 3.493445384348922,
      "tokens_seen": 1775763456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665396188565697,
      "loss": 1.4024,
      "theoretical_loss": 3.493434471466882,
      "tokens_seen": 1775828992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046651955867602806,
      "loss": 1.3746,
      "theoretical_loss": 3.49342355910033,
      "tokens_seen": 1775894528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664994984954865,
      "loss": 1.5006,
      "theoretical_loss": 3.4934126472492215,
      "tokens_seen": 1775960064
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046647943831494484,
      "loss": 1.4287,
      "theoretical_loss": 3.4934017359135137,
      "tokens_seen": 1776025600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664593781344032,
      "loss": 1.3496,
      "theoretical_loss": 3.4933908250931633,
      "tokens_seen": 1776091136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046643931795386156,
      "loss": 1.3728,
      "theoretical_loss": 3.493379914788126,
      "tokens_seen": 1776156672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664192577733199,
      "loss": 1.4248,
      "theoretical_loss": 3.49336900499836,
      "tokens_seen": 1776222208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046639919759277834,
      "loss": 1.4654,
      "theoretical_loss": 3.4933580957238206,
      "tokens_seen": 1776287744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663791374122367,
      "loss": 1.2819,
      "theoretical_loss": 3.4933471869644652,
      "tokens_seen": 1776353280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046635907723169506,
      "loss": 1.3475,
      "theoretical_loss": 3.4933362787202507,
      "tokens_seen": 1776418816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663390170511534,
      "loss": 1.3553,
      "theoretical_loss": 3.4933253709911325,
      "tokens_seen": 1776484352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046631895687061184,
      "loss": 1.4399,
      "theoretical_loss": 3.493314463777069,
      "tokens_seen": 1776549888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662988966900702,
      "loss": 1.2961,
      "theoretical_loss": 3.493303557078015,
      "tokens_seen": 1776615424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046627883650952857,
      "loss": 1.4466,
      "theoretical_loss": 3.493292650893929,
      "tokens_seen": 1776680960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046625877632898693,
      "loss": 1.4076,
      "theoretical_loss": 3.4932817452247664,
      "tokens_seen": 1776746496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046623871614844535,
      "loss": 1.3671,
      "theoretical_loss": 3.493270840070484,
      "tokens_seen": 1776812032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662186559679037,
      "loss": 1.2753,
      "theoretical_loss": 3.4932599354310403,
      "tokens_seen": 1776877568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661985957873621,
      "loss": 1.3254,
      "theoretical_loss": 3.493249031306389,
      "tokens_seen": 1776943104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661785356068205,
      "loss": 1.336,
      "theoretical_loss": 3.4932381276964897,
      "tokens_seen": 1777008640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046615847542627885,
      "loss": 1.3294,
      "theoretical_loss": 3.4932272246012968,
      "tokens_seen": 1777074176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046613841524573727,
      "loss": 1.4117,
      "theoretical_loss": 3.4932163220207686,
      "tokens_seen": 1777139712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046611835506519563,
      "loss": 1.3507,
      "theoretical_loss": 3.493205419954861,
      "tokens_seen": 1777205248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466098294884654,
      "loss": 1.3653,
      "theoretical_loss": 3.4931945184035307,
      "tokens_seen": 1777270784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046607823470411235,
      "loss": 1.317,
      "theoretical_loss": 3.493183617366735,
      "tokens_seen": 1777336320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046605817452357077,
      "loss": 1.4495,
      "theoretical_loss": 3.4931727168444304,
      "tokens_seen": 1777401856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046603811434302913,
      "loss": 1.398,
      "theoretical_loss": 3.4931618168365737,
      "tokens_seen": 1777467392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004660180541624875,
      "loss": 1.387,
      "theoretical_loss": 3.4931509173431214,
      "tokens_seen": 1777532928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046599799398194586,
      "loss": 1.4403,
      "theoretical_loss": 3.4931400183640307,
      "tokens_seen": 1777598464
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004659779338014042,
      "loss": 1.4114,
      "theoretical_loss": 3.4931291198992582,
      "tokens_seen": 1777664000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046595787362086263,
      "loss": 1.4083,
      "theoretical_loss": 3.49311822194876,
      "tokens_seen": 1777729536
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 649180,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1798685789108276,
      "objective/train/theoretical_loss": 3.493107324512494,
      "objective/train/tokens_used": 148254176,
      "theoretical_loss": 3.493107324512494,
      "tokens_seen": 1777795072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465937813440321,
      "loss": 1.2022,
      "theoretical_loss": 3.493107324512494,
      "tokens_seen": 1777795072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046591775325977936,
      "loss": 1.3655,
      "theoretical_loss": 3.493096427590416,
      "tokens_seen": 1777860608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658976930792377,
      "loss": 1.3358,
      "theoretical_loss": 3.4930855311824835,
      "tokens_seen": 1777926144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046587763289869614,
      "loss": 1.3148,
      "theoretical_loss": 3.493074635288653,
      "tokens_seen": 1777991680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658575727181545,
      "loss": 1.407,
      "theoretical_loss": 3.493063739908881,
      "tokens_seen": 1778057216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046583751253761286,
      "loss": 1.3564,
      "theoretical_loss": 3.493052845043125,
      "tokens_seen": 1778122752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658174523570712,
      "loss": 1.3304,
      "theoretical_loss": 3.493041950691341,
      "tokens_seen": 1778188288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657973921765296,
      "loss": 1.3414,
      "theoretical_loss": 3.4930310568534866,
      "tokens_seen": 1778253824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465777331995988,
      "loss": 1.3462,
      "theoretical_loss": 3.4930201635295175,
      "tokens_seen": 1778319360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046575727181544636,
      "loss": 1.3212,
      "theoretical_loss": 3.493009270719392,
      "tokens_seen": 1778384896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657372116349047,
      "loss": 1.4643,
      "theoretical_loss": 3.4929983784230663,
      "tokens_seen": 1778450432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657171514543631,
      "loss": 1.2515,
      "theoretical_loss": 3.492987486640496,
      "tokens_seen": 1778515968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656970912738215,
      "loss": 1.3084,
      "theoretical_loss": 3.4929765953716405,
      "tokens_seen": 1778581504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046567703109327987,
      "loss": 1.3553,
      "theoretical_loss": 3.4929657046164544,
      "tokens_seen": 1778647040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046565697091273823,
      "loss": 1.3112,
      "theoretical_loss": 3.4929548143748956,
      "tokens_seen": 1778712576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656369107321966,
      "loss": 1.2687,
      "theoretical_loss": 3.4929439246469203,
      "tokens_seen": 1778778112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046561685055165495,
      "loss": 1.325,
      "theoretical_loss": 3.492933035432486,
      "tokens_seen": 1778843648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046559679037111337,
      "loss": 1.3414,
      "theoretical_loss": 3.492922146731549,
      "tokens_seen": 1778909184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046557673019057173,
      "loss": 1.3544,
      "theoretical_loss": 3.4929112585440674,
      "tokens_seen": 1778974720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655566700100301,
      "loss": 1.2733,
      "theoretical_loss": 3.492900370869996,
      "tokens_seen": 1779040256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046553660982948846,
      "loss": 1.3441,
      "theoretical_loss": 3.492889483709294,
      "tokens_seen": 1779105792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655165496489469,
      "loss": 1.3427,
      "theoretical_loss": 3.4928785970619165,
      "tokens_seen": 1779171328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046549648946840524,
      "loss": 1.3447,
      "theoretical_loss": 3.492867710927821,
      "tokens_seen": 1779236864
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654764292878636,
      "loss": 1.1811,
      "theoretical_loss": 3.4928568253069643,
      "tokens_seen": 1779302400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046545636910732196,
      "loss": 1.3992,
      "theoretical_loss": 3.4928459401993033,
      "tokens_seen": 1779367936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654363089267803,
      "loss": 1.3751,
      "theoretical_loss": 3.492835055604796,
      "tokens_seen": 1779433472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046541624874623874,
      "loss": 1.3041,
      "theoretical_loss": 3.492824171523397,
      "tokens_seen": 1779499008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653961885656971,
      "loss": 1.4208,
      "theoretical_loss": 3.492813287955065,
      "tokens_seen": 1779564544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046537612838515546,
      "loss": 1.2894,
      "theoretical_loss": 3.4928024048997566,
      "tokens_seen": 1779630080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653560682046138,
      "loss": 1.4283,
      "theoretical_loss": 3.492791522357429,
      "tokens_seen": 1779695616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046533600802407224,
      "loss": 1.254,
      "theoretical_loss": 3.4927806403280384,
      "tokens_seen": 1779761152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653159478435306,
      "loss": 1.2984,
      "theoretical_loss": 3.4927697588115416,
      "tokens_seen": 1779826688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046529588766298897,
      "loss": 1.3202,
      "theoretical_loss": 3.492758877807897,
      "tokens_seen": 1779892224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046527582748244733,
      "loss": 1.3428,
      "theoretical_loss": 3.4927479973170596,
      "tokens_seen": 1779957760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046525576730190575,
      "loss": 1.5392,
      "theoretical_loss": 3.4927371173389874,
      "tokens_seen": 1780023296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652357071213641,
      "loss": 1.4185,
      "theoretical_loss": 3.4927262378736375,
      "tokens_seen": 1780088832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046521564694082247,
      "loss": 1.3486,
      "theoretical_loss": 3.4927153589209667,
      "tokens_seen": 1780154368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046519558676028083,
      "loss": 1.4041,
      "theoretical_loss": 3.492704480480932,
      "tokens_seen": 1780219904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651755265797392,
      "loss": 1.3262,
      "theoretical_loss": 3.49269360255349,
      "tokens_seen": 1780285440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651554663991976,
      "loss": 1.3214,
      "theoretical_loss": 3.492682725138598,
      "tokens_seen": 1780350976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465135406218656,
      "loss": 1.3432,
      "theoretical_loss": 3.492671848236213,
      "tokens_seen": 1780416512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046511534603811434,
      "loss": 1.2627,
      "theoretical_loss": 3.4926609718462918,
      "tokens_seen": 1780482048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650952858575727,
      "loss": 1.3192,
      "theoretical_loss": 3.4926500959687914,
      "tokens_seen": 1780547584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650752256770311,
      "loss": 1.4248,
      "theoretical_loss": 3.492639220603669,
      "tokens_seen": 1780613120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650551654964895,
      "loss": 1.4194,
      "theoretical_loss": 3.4926283457508815,
      "tokens_seen": 1780678656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046503510531594784,
      "loss": 1.4209,
      "theoretical_loss": 3.4926174714103864,
      "tokens_seen": 1780744192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650150451354062,
      "loss": 1.4346,
      "theoretical_loss": 3.4926065975821396,
      "tokens_seen": 1780809728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046499498495486456,
      "loss": 1.242,
      "theoretical_loss": 3.4925957242660988,
      "tokens_seen": 1780875264
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464974924774323,
      "loss": 1.3773,
      "theoretical_loss": 3.4925848514622215,
      "tokens_seen": 1780940800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046495486459378134,
      "loss": 1.3099,
      "theoretical_loss": 3.4925739791704635,
      "tokens_seen": 1781006336
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 650314,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2257611751556396,
      "objective/train/theoretical_loss": 3.492563107390783,
      "objective/train/tokens_used": 151530976,
      "theoretical_loss": 3.492563107390783,
      "tokens_seen": 1781071872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649348044132397,
      "loss": 1.4112,
      "theoretical_loss": 3.492563107390783,
      "tokens_seen": 1781071872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046491474423269807,
      "loss": 1.4127,
      "theoretical_loss": 3.4925522361231365,
      "tokens_seen": 1781137408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648946840521565,
      "loss": 1.328,
      "theoretical_loss": 3.4925413653674813,
      "tokens_seen": 1781202944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046487462387161485,
      "loss": 1.3604,
      "theoretical_loss": 3.4925304951237734,
      "tokens_seen": 1781268480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648545636910732,
      "loss": 1.4498,
      "theoretical_loss": 3.4925196253919717,
      "tokens_seen": 1781334016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046483450351053157,
      "loss": 1.563,
      "theoretical_loss": 3.4925087561720325,
      "tokens_seen": 1781399552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046481444332998993,
      "loss": 1.3441,
      "theoretical_loss": 3.4924978874639114,
      "tokens_seen": 1781465088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046479438314944835,
      "loss": 1.3977,
      "theoretical_loss": 3.4924870192675677,
      "tokens_seen": 1781530624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647743229689067,
      "loss": 1.3104,
      "theoretical_loss": 3.4924761515829577,
      "tokens_seen": 1781596160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647542627883651,
      "loss": 1.3729,
      "theoretical_loss": 3.492465284410038,
      "tokens_seen": 1781661696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046473420260782344,
      "loss": 1.4054,
      "theoretical_loss": 3.4924544177487657,
      "tokens_seen": 1781727232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647141424272819,
      "loss": 1.2698,
      "theoretical_loss": 3.4924435515990986,
      "tokens_seen": 1781792768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046469408224674027,
      "loss": 1.3376,
      "theoretical_loss": 3.492432685960993,
      "tokens_seen": 1781858304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046467402206619863,
      "loss": 1.4646,
      "theoretical_loss": 3.492421820834406,
      "tokens_seen": 1781923840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464653961885657,
      "loss": 1.4831,
      "theoretical_loss": 3.492410956219296,
      "tokens_seen": 1781989376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046463390170511535,
      "loss": 1.4281,
      "theoretical_loss": 3.492400092115619,
      "tokens_seen": 1782054912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046461384152457377,
      "loss": 1.3922,
      "theoretical_loss": 3.492389228523332,
      "tokens_seen": 1782120448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046459378134403213,
      "loss": 1.3606,
      "theoretical_loss": 3.4923783654423923,
      "tokens_seen": 1782185984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645737211634905,
      "loss": 1.367,
      "theoretical_loss": 3.4923675028727574,
      "tokens_seen": 1782251520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046455366098294886,
      "loss": 1.3624,
      "theoretical_loss": 3.4923566408143842,
      "tokens_seen": 1782317056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645336008024073,
      "loss": 1.3743,
      "theoretical_loss": 3.49234577926723,
      "tokens_seen": 1782382592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046451354062186564,
      "loss": 1.3936,
      "theoretical_loss": 3.4923349182312515,
      "tokens_seen": 1782448128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464493480441324,
      "loss": 1.4711,
      "theoretical_loss": 3.4923240577064067,
      "tokens_seen": 1782513664
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046447342026078236,
      "loss": 1.3881,
      "theoretical_loss": 3.4923131976926514,
      "tokens_seen": 1782579200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644533600802407,
      "loss": 1.347,
      "theoretical_loss": 3.492302338189944,
      "tokens_seen": 1782644736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046443329989969914,
      "loss": 1.4536,
      "theoretical_loss": 3.492291479198241,
      "tokens_seen": 1782710272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644132397191575,
      "loss": 1.2682,
      "theoretical_loss": 3.4922806207174997,
      "tokens_seen": 1782775808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046439317953861586,
      "loss": 1.2763,
      "theoretical_loss": 3.4922697627476778,
      "tokens_seen": 1782841344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643731193580742,
      "loss": 1.3557,
      "theoretical_loss": 3.4922589052887316,
      "tokens_seen": 1782906880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046435305917753264,
      "loss": 1.2785,
      "theoretical_loss": 3.492248048340619,
      "tokens_seen": 1782972416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464332998996991,
      "loss": 1.4066,
      "theoretical_loss": 3.492237191903297,
      "tokens_seen": 1783037952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046431293881644937,
      "loss": 1.2322,
      "theoretical_loss": 3.4922263359767225,
      "tokens_seen": 1783103488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046429287863590773,
      "loss": 1.3669,
      "theoretical_loss": 3.492215480560853,
      "tokens_seen": 1783169024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046427281845536615,
      "loss": 1.3803,
      "theoretical_loss": 3.4922046256556456,
      "tokens_seen": 1783234560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642527582748245,
      "loss": 1.3788,
      "theoretical_loss": 3.4921937712610576,
      "tokens_seen": 1783300096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046423269809428287,
      "loss": 1.3734,
      "theoretical_loss": 3.492182917377046,
      "tokens_seen": 1783365632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046421263791374123,
      "loss": 1.3093,
      "theoretical_loss": 3.492172064003568,
      "tokens_seen": 1783431168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641925777331996,
      "loss": 1.4635,
      "theoretical_loss": 3.492161211140581,
      "tokens_seen": 1783496704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464172517552658,
      "loss": 1.2471,
      "theoretical_loss": 3.4921503587880425,
      "tokens_seen": 1783562240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641524573721164,
      "loss": 1.3886,
      "theoretical_loss": 3.4921395069459096,
      "tokens_seen": 1783627776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046413239719157474,
      "loss": 1.3439,
      "theoretical_loss": 3.492128655614139,
      "tokens_seen": 1783693312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641123370110331,
      "loss": 1.507,
      "theoretical_loss": 3.492117804792689,
      "tokens_seen": 1783758848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640922768304915,
      "loss": 1.3927,
      "theoretical_loss": 3.4921069544815153,
      "tokens_seen": 1783824384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640722166499499,
      "loss": 1.3649,
      "theoretical_loss": 3.4920961046805763,
      "tokens_seen": 1783889920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046405215646940824,
      "loss": 1.3722,
      "theoretical_loss": 3.4920852553898296,
      "tokens_seen": 1783955456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640320962888666,
      "loss": 1.4852,
      "theoretical_loss": 3.4920744066092313,
      "tokens_seen": 1784020992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046401203610832496,
      "loss": 1.4986,
      "theoretical_loss": 3.4920635583387396,
      "tokens_seen": 1784086528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639919759277834,
      "loss": 1.348,
      "theoretical_loss": 3.492052710578311,
      "tokens_seen": 1784152064
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046397191574724174,
      "loss": 1.3796,
      "theoretical_loss": 3.492041863327904,
      "tokens_seen": 1784217600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639518555667001,
      "loss": 1.3677,
      "theoretical_loss": 3.4920310165874744,
      "tokens_seen": 1784283136
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 651558,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.542694091796875,
      "objective/train/theoretical_loss": 3.49202017035698,
      "objective/train/tokens_used": 154807776,
      "theoretical_loss": 3.49202017035698,
      "tokens_seen": 1784348672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046393179538615847,
      "loss": 1.3991,
      "theoretical_loss": 3.49202017035698,
      "tokens_seen": 1784348672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639117352056169,
      "loss": 1.393,
      "theoretical_loss": 3.492009324636379,
      "tokens_seen": 1784414208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046389167502507525,
      "loss": 1.4695,
      "theoretical_loss": 3.491998479425628,
      "tokens_seen": 1784479744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638716148445336,
      "loss": 1.4816,
      "theoretical_loss": 3.491987634724684,
      "tokens_seen": 1784545280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046385155466399197,
      "loss": 1.4259,
      "theoretical_loss": 3.4919767905335046,
      "tokens_seen": 1784610816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046383149448345033,
      "loss": 1.3959,
      "theoretical_loss": 3.491965946852047,
      "tokens_seen": 1784676352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046381143430290875,
      "loss": 1.4587,
      "theoretical_loss": 3.491955103680269,
      "tokens_seen": 1784741888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637913741223671,
      "loss": 1.3978,
      "theoretical_loss": 3.4919442610181277,
      "tokens_seen": 1784807424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637713139418255,
      "loss": 1.5788,
      "theoretical_loss": 3.49193341886558,
      "tokens_seen": 1784872960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046375125376128384,
      "loss": 1.3937,
      "theoretical_loss": 3.491922577222584,
      "tokens_seen": 1784938496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046373119358074225,
      "loss": 1.3733,
      "theoretical_loss": 3.4919117360890963,
      "tokens_seen": 1785004032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637111334002006,
      "loss": 1.4229,
      "theoretical_loss": 3.4919008954650748,
      "tokens_seen": 1785069568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463691073219659,
      "loss": 1.5605,
      "theoretical_loss": 3.491890055350477,
      "tokens_seen": 1785135104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046367101303911734,
      "loss": 1.4093,
      "theoretical_loss": 3.491879215745259,
      "tokens_seen": 1785200640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636509528585757,
      "loss": 1.4728,
      "theoretical_loss": 3.4918683766493794,
      "tokens_seen": 1785266176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636308926780341,
      "loss": 1.3199,
      "theoretical_loss": 3.491857538062795,
      "tokens_seen": 1785331712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636108324974925,
      "loss": 1.3595,
      "theoretical_loss": 3.491846699985464,
      "tokens_seen": 1785397248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046359077231695084,
      "loss": 1.4362,
      "theoretical_loss": 3.491835862417343,
      "tokens_seen": 1785462784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635707121364092,
      "loss": 1.4483,
      "theoretical_loss": 3.491825025358389,
      "tokens_seen": 1785528320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635506519558676,
      "loss": 1.3289,
      "theoretical_loss": 3.4918141888085605,
      "tokens_seen": 1785593856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463530591775326,
      "loss": 1.5213,
      "theoretical_loss": 3.4918033527678145,
      "tokens_seen": 1785659392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046351053159478434,
      "loss": 1.2615,
      "theoretical_loss": 3.491792517236108,
      "tokens_seen": 1785724928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634904714142427,
      "loss": 1.3449,
      "theoretical_loss": 3.4917816822133982,
      "tokens_seen": 1785790464
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046347041123370107,
      "loss": 1.3547,
      "theoretical_loss": 3.4917708476996436,
      "tokens_seen": 1785856000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634503510531595,
      "loss": 1.3417,
      "theoretical_loss": 3.491760013694801,
      "tokens_seen": 1785921536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046343029087261785,
      "loss": 1.4067,
      "theoretical_loss": 3.491749180198828,
      "tokens_seen": 1785987072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634102306920762,
      "loss": 1.347,
      "theoretical_loss": 3.4917383472116814,
      "tokens_seen": 1786052608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046339017051153457,
      "loss": 1.3094,
      "theoretical_loss": 3.491727514733319,
      "tokens_seen": 1786118144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463370110330993,
      "loss": 1.2779,
      "theoretical_loss": 3.4917166827636987,
      "tokens_seen": 1786183680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046335005015045135,
      "loss": 1.409,
      "theoretical_loss": 3.491705851302777,
      "tokens_seen": 1786249216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633299899699097,
      "loss": 1.3644,
      "theoretical_loss": 3.4916950203505124,
      "tokens_seen": 1786314752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633099297893681,
      "loss": 1.3594,
      "theoretical_loss": 3.491684189906862,
      "tokens_seen": 1786380288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046328986960882644,
      "loss": 1.3757,
      "theoretical_loss": 3.4916733599717826,
      "tokens_seen": 1786445824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046326980942828485,
      "loss": 1.3782,
      "theoretical_loss": 3.4916625305452325,
      "tokens_seen": 1786511360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632497492477432,
      "loss": 1.2859,
      "theoretical_loss": 3.491651701627169,
      "tokens_seen": 1786576896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632296890672016,
      "loss": 1.5382,
      "theoretical_loss": 3.4916408732175492,
      "tokens_seen": 1786642432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046320962888666,
      "loss": 1.3755,
      "theoretical_loss": 3.491630045316331,
      "tokens_seen": 1786707968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631895687061184,
      "loss": 1.4508,
      "theoretical_loss": 3.4916192179234713,
      "tokens_seen": 1786773504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631695085255768,
      "loss": 1.5339,
      "theoretical_loss": 3.491608391038928,
      "tokens_seen": 1786839040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046314944834503514,
      "loss": 1.31,
      "theoretical_loss": 3.4915975646626594,
      "tokens_seen": 1786904576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631293881644935,
      "loss": 1.3931,
      "theoretical_loss": 3.491586738794621,
      "tokens_seen": 1786970112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631093279839519,
      "loss": 1.3781,
      "theoretical_loss": 3.491575913434772,
      "tokens_seen": 1787035648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630892678034103,
      "loss": 1.2669,
      "theoretical_loss": 3.4915650885830694,
      "tokens_seen": 1787101184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046306920762286864,
      "loss": 1.3142,
      "theoretical_loss": 3.4915542642394706,
      "tokens_seen": 1787166720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463049147442327,
      "loss": 1.4238,
      "theoretical_loss": 3.4915434404039334,
      "tokens_seen": 1787232256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046302908726178536,
      "loss": 1.2684,
      "theoretical_loss": 3.491532617076415,
      "tokens_seen": 1787297792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630090270812438,
      "loss": 1.4976,
      "theoretical_loss": 3.491521794256873,
      "tokens_seen": 1787363328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046298896690070214,
      "loss": 1.3864,
      "theoretical_loss": 3.4915109719452646,
      "tokens_seen": 1787428864
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629689067201605,
      "loss": 1.3209,
      "theoretical_loss": 3.4915001501415484,
      "tokens_seen": 1787494400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046294884653961887,
      "loss": 1.4777,
      "theoretical_loss": 3.491489328845681,
      "tokens_seen": 1787559936
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 652772,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3689403533935547,
      "objective/train/theoretical_loss": 3.4914785080576207,
      "objective/train/tokens_used": 158084576,
      "theoretical_loss": 3.4914785080576207,
      "tokens_seen": 1787625472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629287863590773,
      "loss": 1.4206,
      "theoretical_loss": 3.4914785080576207,
      "tokens_seen": 1787625472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046290872617853565,
      "loss": 1.3574,
      "theoretical_loss": 3.491467687777324,
      "tokens_seen": 1787691008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462888665997994,
      "loss": 1.3886,
      "theoretical_loss": 3.4914568680047493,
      "tokens_seen": 1787756544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046286860581745237,
      "loss": 1.4557,
      "theoretical_loss": 3.4914460487398538,
      "tokens_seen": 1787822080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046284854563691073,
      "loss": 1.3635,
      "theoretical_loss": 3.491435229982595,
      "tokens_seen": 1787887616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046282848545636915,
      "loss": 1.4842,
      "theoretical_loss": 3.4914244117329307,
      "tokens_seen": 1787953152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628084252758275,
      "loss": 1.392,
      "theoretical_loss": 3.491413593990819,
      "tokens_seen": 1788018688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627883650952859,
      "loss": 1.3914,
      "theoretical_loss": 3.4914027767562166,
      "tokens_seen": 1788084224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046276830491474424,
      "loss": 1.3906,
      "theoretical_loss": 3.4913919600290813,
      "tokens_seen": 1788149760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046274824473420265,
      "loss": 1.2698,
      "theoretical_loss": 3.491381143809371,
      "tokens_seen": 1788215296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462728184553661,
      "loss": 1.3993,
      "theoretical_loss": 3.4913703280970427,
      "tokens_seen": 1788280832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627081243731194,
      "loss": 1.4036,
      "theoretical_loss": 3.4913595128920547,
      "tokens_seen": 1788346368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046268806419257774,
      "loss": 1.2978,
      "theoretical_loss": 3.4913486981943644,
      "tokens_seen": 1788411904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626680040120361,
      "loss": 1.2322,
      "theoretical_loss": 3.4913378840039293,
      "tokens_seen": 1788477440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626479438314945,
      "loss": 1.3615,
      "theoretical_loss": 3.491327070320707,
      "tokens_seen": 1788542976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626278836509529,
      "loss": 1.3352,
      "theoretical_loss": 3.491316257144655,
      "tokens_seen": 1788608512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046260782347041124,
      "loss": 1.4185,
      "theoretical_loss": 3.4913054444757314,
      "tokens_seen": 1788674048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625877632898696,
      "loss": 1.3946,
      "theoretical_loss": 3.4912946323138936,
      "tokens_seen": 1788739584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462567703109328,
      "loss": 1.2568,
      "theoretical_loss": 3.491283820659099,
      "tokens_seen": 1788805120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625476429287864,
      "loss": 1.4322,
      "theoretical_loss": 3.491273009511306,
      "tokens_seen": 1788870656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046252758274824475,
      "loss": 1.4246,
      "theoretical_loss": 3.491262198870471,
      "tokens_seen": 1788936192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625075225677031,
      "loss": 1.351,
      "theoretical_loss": 3.491251388736553,
      "tokens_seen": 1789001728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046248746238716147,
      "loss": 1.357,
      "theoretical_loss": 3.4912405791095082,
      "tokens_seen": 1789067264
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624674022066199,
      "loss": 1.3887,
      "theoretical_loss": 3.491229769989296,
      "tokens_seen": 1789132800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046244734202607825,
      "loss": 1.3465,
      "theoretical_loss": 3.4912189613758726,
      "tokens_seen": 1789198336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624272818455366,
      "loss": 1.2618,
      "theoretical_loss": 3.4912081532691963,
      "tokens_seen": 1789263872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462407221664995,
      "loss": 1.3755,
      "theoretical_loss": 3.491197345669225,
      "tokens_seen": 1789329408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623871614844534,
      "loss": 1.3202,
      "theoretical_loss": 3.4911865385759158,
      "tokens_seen": 1789394944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046236710130391175,
      "loss": 1.3203,
      "theoretical_loss": 3.4911757319892267,
      "tokens_seen": 1789460480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623470411233701,
      "loss": 1.4261,
      "theoretical_loss": 3.4911649259091155,
      "tokens_seen": 1789526016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623269809428285,
      "loss": 1.2647,
      "theoretical_loss": 3.4911541203355396,
      "tokens_seen": 1789591552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046230692076228684,
      "loss": 1.2963,
      "theoretical_loss": 3.4911433152684572,
      "tokens_seen": 1789657088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046228686058174525,
      "loss": 1.3134,
      "theoretical_loss": 3.4911325107078257,
      "tokens_seen": 1789722624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622668004012036,
      "loss": 1.4385,
      "theoretical_loss": 3.491121706653603,
      "tokens_seen": 1789788160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462246740220662,
      "loss": 1.3566,
      "theoretical_loss": 3.4911109031057457,
      "tokens_seen": 1789853696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046222668004012034,
      "loss": 1.2555,
      "theoretical_loss": 3.4911001000642132,
      "tokens_seen": 1789919232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046220661985957876,
      "loss": 1.4008,
      "theoretical_loss": 3.491089297528963,
      "tokens_seen": 1789984768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621865596790371,
      "loss": 1.2429,
      "theoretical_loss": 3.4910784954999516,
      "tokens_seen": 1790050304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621664994984955,
      "loss": 1.319,
      "theoretical_loss": 3.491067693977138,
      "tokens_seen": 1790115840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046214643931795384,
      "loss": 1.3774,
      "theoretical_loss": 3.4910568929604793,
      "tokens_seen": 1790181376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621263791374122,
      "loss": 1.4883,
      "theoretical_loss": 3.491046092449933,
      "tokens_seen": 1790246912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004621063189568706,
      "loss": 1.5182,
      "theoretical_loss": 3.4910352924454577,
      "tokens_seen": 1790312448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000462086258776329,
      "loss": 1.4293,
      "theoretical_loss": 3.4910244929470107,
      "tokens_seen": 1790377984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046206619859578735,
      "loss": 1.3163,
      "theoretical_loss": 3.4910136939545495,
      "tokens_seen": 1790443520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620461384152457,
      "loss": 1.2555,
      "theoretical_loss": 3.4910028954680326,
      "tokens_seen": 1790509056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620260782347041,
      "loss": 1.3977,
      "theoretical_loss": 3.490992097487417,
      "tokens_seen": 1790574592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620060180541625,
      "loss": 1.4721,
      "theoretical_loss": 3.490981300012661,
      "tokens_seen": 1790640128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046198595787362085,
      "loss": 1.2129,
      "theoretical_loss": 3.490970503043722,
      "tokens_seen": 1790705664
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619658976930792,
      "loss": 1.4094,
      "theoretical_loss": 3.4909597065805587,
      "tokens_seen": 1790771200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046194583751253763,
      "loss": 1.4175,
      "theoretical_loss": 3.4909489106231275,
      "tokens_seen": 1790836736
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 653832,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5036251544952393,
      "objective/train/theoretical_loss": 3.490938115171387,
      "objective/train/tokens_used": 161361376,
      "theoretical_loss": 3.490938115171387,
      "tokens_seen": 1790902272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461925777331996,
      "loss": 1.482,
      "theoretical_loss": 3.490938115171387,
      "tokens_seen": 1790902272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046190571715145435,
      "loss": 1.3238,
      "theoretical_loss": 3.4909273202252953,
      "tokens_seen": 1790967808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618856569709127,
      "loss": 1.3115,
      "theoretical_loss": 3.49091652578481,
      "tokens_seen": 1791033344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618655967903711,
      "loss": 1.4552,
      "theoretical_loss": 3.490905731849888,
      "tokens_seen": 1791098880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618455366098295,
      "loss": 1.3243,
      "theoretical_loss": 3.4908949384204884,
      "tokens_seen": 1791164416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046182547642928786,
      "loss": 1.2971,
      "theoretical_loss": 3.490884145496568,
      "tokens_seen": 1791229952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618054162487462,
      "loss": 1.421,
      "theoretical_loss": 3.4908733530780856,
      "tokens_seen": 1791295488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617853560682046,
      "loss": 1.2402,
      "theoretical_loss": 3.490862561164999,
      "tokens_seen": 1791361024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461765295887663,
      "loss": 1.3192,
      "theoretical_loss": 3.490851769757265,
      "tokens_seen": 1791426560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046174523570712136,
      "loss": 1.2511,
      "theoretical_loss": 3.490840978854842,
      "tokens_seen": 1791492096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617251755265797,
      "loss": 1.2764,
      "theoretical_loss": 3.4908301884576884,
      "tokens_seen": 1791557632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046170511534603814,
      "loss": 1.4033,
      "theoretical_loss": 3.4908193985657614,
      "tokens_seen": 1791623168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616850551654965,
      "loss": 1.3282,
      "theoretical_loss": 3.490808609179019,
      "tokens_seen": 1791688704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616649949849549,
      "loss": 1.4277,
      "theoretical_loss": 3.490797820297419,
      "tokens_seen": 1791754240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616449348044133,
      "loss": 1.364,
      "theoretical_loss": 3.4907870319209198,
      "tokens_seen": 1791819776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046162487462387164,
      "loss": 1.3934,
      "theoretical_loss": 3.4907762440494787,
      "tokens_seen": 1791885312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046160481444333,
      "loss": 1.3087,
      "theoretical_loss": 3.4907654566830537,
      "tokens_seen": 1791950848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615847542627884,
      "loss": 1.414,
      "theoretical_loss": 3.4907546698216025,
      "tokens_seen": 1792016384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615646940822468,
      "loss": 1.4455,
      "theoretical_loss": 3.4907438834650844,
      "tokens_seen": 1792081920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046154463390170515,
      "loss": 1.2248,
      "theoretical_loss": 3.4907330976134547,
      "tokens_seen": 1792147456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615245737211635,
      "loss": 1.3332,
      "theoretical_loss": 3.4907223122666737,
      "tokens_seen": 1792212992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046150451354062187,
      "loss": 1.5327,
      "theoretical_loss": 3.490711527424698,
      "tokens_seen": 1792278528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614844533600803,
      "loss": 1.3328,
      "theoretical_loss": 3.490700743087486,
      "tokens_seen": 1792344064
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046146439317953865,
      "loss": 1.3008,
      "theoretical_loss": 3.4906899592549956,
      "tokens_seen": 1792409600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461444332998997,
      "loss": 1.4446,
      "theoretical_loss": 3.4906791759271845,
      "tokens_seen": 1792475136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614242728184554,
      "loss": 1.3194,
      "theoretical_loss": 3.490668393104011,
      "tokens_seen": 1792540672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614042126379138,
      "loss": 1.2333,
      "theoretical_loss": 3.4906576107854326,
      "tokens_seen": 1792606208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046138415245737215,
      "loss": 1.3816,
      "theoretical_loss": 3.490646828971407,
      "tokens_seen": 1792671744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613640922768305,
      "loss": 1.3335,
      "theoretical_loss": 3.4906360476618934,
      "tokens_seen": 1792737280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613440320962889,
      "loss": 1.2906,
      "theoretical_loss": 3.4906252668568483,
      "tokens_seen": 1792802816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046132397191574724,
      "loss": 1.3803,
      "theoretical_loss": 3.4906144865562307,
      "tokens_seen": 1792868352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046130391173520566,
      "loss": 1.4188,
      "theoretical_loss": 3.490603706759998,
      "tokens_seen": 1792933888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461283851554664,
      "loss": 1.2697,
      "theoretical_loss": 3.490592927468108,
      "tokens_seen": 1792999424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612637913741224,
      "loss": 1.3661,
      "theoretical_loss": 3.4905821486805197,
      "tokens_seen": 1793064960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046124373119358074,
      "loss": 1.3723,
      "theoretical_loss": 3.49057137039719,
      "tokens_seen": 1793130496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046122367101303916,
      "loss": 1.2416,
      "theoretical_loss": 3.490560592618077,
      "tokens_seen": 1793196032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612036108324975,
      "loss": 1.3244,
      "theoretical_loss": 3.4905498153431394,
      "tokens_seen": 1793261568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611835506519559,
      "loss": 1.4122,
      "theoretical_loss": 3.4905390385723347,
      "tokens_seen": 1793327104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046116349047141425,
      "loss": 1.298,
      "theoretical_loss": 3.49052826230562,
      "tokens_seen": 1793392640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611434302908726,
      "loss": 1.3589,
      "theoretical_loss": 3.490517486542956,
      "tokens_seen": 1793458176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461123370110331,
      "loss": 1.3116,
      "theoretical_loss": 3.490506711284297,
      "tokens_seen": 1793523712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611033099297894,
      "loss": 1.5182,
      "theoretical_loss": 3.490495936529604,
      "tokens_seen": 1793589248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046108324974924775,
      "loss": 1.3556,
      "theoretical_loss": 3.490485162278834,
      "tokens_seen": 1793654784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610631895687061,
      "loss": 1.3344,
      "theoretical_loss": 3.490474388531945,
      "tokens_seen": 1793720320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610431293881645,
      "loss": 1.4541,
      "theoretical_loss": 3.4904636152888946,
      "tokens_seen": 1793785856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610230692076229,
      "loss": 1.3576,
      "theoretical_loss": 3.4904528425496415,
      "tokens_seen": 1793851392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046100300902708125,
      "loss": 1.32,
      "theoretical_loss": 3.490442070314143,
      "tokens_seen": 1793916928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609829488465396,
      "loss": 1.3711,
      "theoretical_loss": 3.490431298582358,
      "tokens_seen": 1793982464
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046096288866599803,
      "loss": 1.2296,
      "theoretical_loss": 3.490420527354244,
      "tokens_seen": 1794048000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609428284854564,
      "loss": 1.413,
      "theoretical_loss": 3.49040975662976,
      "tokens_seen": 1794113536
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 655001,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.234043002128601,
      "objective/train/theoretical_loss": 3.4903989864088625,
      "objective/train/tokens_used": 164638176,
      "theoretical_loss": 3.4903989864088625,
      "tokens_seen": 1794179072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046092276830491475,
      "loss": 1.3584,
      "theoretical_loss": 3.4903989864088625,
      "tokens_seen": 1794179072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609027081243731,
      "loss": 1.3314,
      "theoretical_loss": 3.4903882166915103,
      "tokens_seen": 1794244608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608826479438315,
      "loss": 1.2828,
      "theoretical_loss": 3.4903774474776617,
      "tokens_seen": 1794310144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608625877632899,
      "loss": 1.1977,
      "theoretical_loss": 3.4903666787672747,
      "tokens_seen": 1794375680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046084252758274826,
      "loss": 1.3539,
      "theoretical_loss": 3.4903559105603073,
      "tokens_seen": 1794441216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608224674022066,
      "loss": 1.33,
      "theoretical_loss": 3.490345142856717,
      "tokens_seen": 1794506752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460802407221665,
      "loss": 1.3558,
      "theoretical_loss": 3.4903343756564627,
      "tokens_seen": 1794572288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607823470411234,
      "loss": 1.4131,
      "theoretical_loss": 3.4903236089595024,
      "tokens_seen": 1794637824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046076228686058176,
      "loss": 1.4542,
      "theoretical_loss": 3.4903128427657935,
      "tokens_seen": 1794703360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607422266800401,
      "loss": 1.3535,
      "theoretical_loss": 3.4903020770752953,
      "tokens_seen": 1794768896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607221664994985,
      "loss": 1.3594,
      "theoretical_loss": 3.4902913118879644,
      "tokens_seen": 1794834432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046070210631895685,
      "loss": 1.3839,
      "theoretical_loss": 3.4902805472037604,
      "tokens_seen": 1794899968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046068204613841526,
      "loss": 1.3381,
      "theoretical_loss": 3.4902697830226406,
      "tokens_seen": 1794965504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606619859578736,
      "loss": 1.3986,
      "theoretical_loss": 3.490259019344563,
      "tokens_seen": 1795031040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460641925777332,
      "loss": 1.2497,
      "theoretical_loss": 3.490248256169486,
      "tokens_seen": 1795096576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046062186559679035,
      "loss": 1.3083,
      "theoretical_loss": 3.490237493497368,
      "tokens_seen": 1795162112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046060180541624877,
      "loss": 1.4295,
      "theoretical_loss": 3.4902267313281667,
      "tokens_seen": 1795227648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046058174523570713,
      "loss": 1.2517,
      "theoretical_loss": 3.4902159696618402,
      "tokens_seen": 1795293184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605616850551655,
      "loss": 1.2911,
      "theoretical_loss": 3.490205208498347,
      "tokens_seen": 1795358720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046054162487462385,
      "loss": 1.2823,
      "theoretical_loss": 3.490194447837645,
      "tokens_seen": 1795424256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605215646940822,
      "loss": 1.3681,
      "theoretical_loss": 3.490183687679693,
      "tokens_seen": 1795489792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046050150451354063,
      "loss": 1.2534,
      "theoretical_loss": 3.490172928024448,
      "tokens_seen": 1795555328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460481444332999,
      "loss": 1.2252,
      "theoretical_loss": 3.490162168871869,
      "tokens_seen": 1795620864
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046046138415245736,
      "loss": 1.2917,
      "theoretical_loss": 3.4901514102219138,
      "tokens_seen": 1795686400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604413239719157,
      "loss": 1.3955,
      "theoretical_loss": 3.490140652074541,
      "tokens_seen": 1795751936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046042126379137414,
      "loss": 1.537,
      "theoretical_loss": 3.4901298944297086,
      "tokens_seen": 1795817472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604012036108325,
      "loss": 1.3599,
      "theoretical_loss": 3.4901191372873743,
      "tokens_seen": 1795883008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046038114343029086,
      "loss": 1.3454,
      "theoretical_loss": 3.490108380647497,
      "tokens_seen": 1795948544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603610832497492,
      "loss": 1.4816,
      "theoretical_loss": 3.4900976245100344,
      "tokens_seen": 1796014080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603410230692076,
      "loss": 1.3968,
      "theoretical_loss": 3.4900868688749442,
      "tokens_seen": 1796079616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460320962888666,
      "loss": 1.3785,
      "theoretical_loss": 3.4900761137421865,
      "tokens_seen": 1796145152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046030090270812436,
      "loss": 1.3047,
      "theoretical_loss": 3.4900653591117177,
      "tokens_seen": 1796210688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602808425275827,
      "loss": 1.2954,
      "theoretical_loss": 3.490054604983497,
      "tokens_seen": 1796276224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602607823470411,
      "loss": 1.2261,
      "theoretical_loss": 3.4900438513574814,
      "tokens_seen": 1796341760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602407221664995,
      "loss": 1.3646,
      "theoretical_loss": 3.490033098233631,
      "tokens_seen": 1796407296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046022066198595787,
      "loss": 1.4895,
      "theoretical_loss": 3.4900223456119024,
      "tokens_seen": 1796472832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602006018054163,
      "loss": 1.2432,
      "theoretical_loss": 3.4900115934922544,
      "tokens_seen": 1796538368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046018054162487465,
      "loss": 1.4947,
      "theoretical_loss": 3.4900008418746458,
      "tokens_seen": 1796603904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460160481444333,
      "loss": 1.365,
      "theoretical_loss": 3.4899900907590338,
      "tokens_seen": 1796669440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601404212637914,
      "loss": 1.3777,
      "theoretical_loss": 3.4899793401453776,
      "tokens_seen": 1796734976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601203610832498,
      "loss": 1.3186,
      "theoretical_loss": 3.4899685900336346,
      "tokens_seen": 1796800512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046010030090270815,
      "loss": 1.2987,
      "theoretical_loss": 3.489957840423764,
      "tokens_seen": 1796866048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600802407221665,
      "loss": 1.3973,
      "theoretical_loss": 3.489947091315723,
      "tokens_seen": 1796931584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046006018054162493,
      "loss": 1.1846,
      "theoretical_loss": 3.4899363427094707,
      "tokens_seen": 1796997120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600401203610833,
      "loss": 1.497,
      "theoretical_loss": 3.489925594604965,
      "tokens_seen": 1797062656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046002006018054165,
      "loss": 1.4801,
      "theoretical_loss": 3.4899148470021646,
      "tokens_seen": 1797128192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046,
      "loss": 1.4469,
      "theoretical_loss": 3.489904099901027,
      "tokens_seen": 1797193728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045997993981945843,
      "loss": 1.3574,
      "theoretical_loss": 3.4898933533015115,
      "tokens_seen": 1797259264
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599598796389168,
      "loss": 1.4101,
      "theoretical_loss": 3.489882607203576,
      "tokens_seen": 1797324800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045993981945837515,
      "loss": 1.3496,
      "theoretical_loss": 3.489871861607178,
      "tokens_seen": 1797390336
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 656301,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.390718936920166,
      "objective/train/theoretical_loss": 3.4898611165122766,
      "objective/train/tokens_used": 167914976,
      "theoretical_loss": 3.4898611165122766,
      "tokens_seen": 1797455872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599197592778335,
      "loss": 1.4125,
      "theoretical_loss": 3.4898611165122766,
      "tokens_seen": 1797455872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598996990972919,
      "loss": 1.4021,
      "theoretical_loss": 3.4898503719188305,
      "tokens_seen": 1797521408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598796389167503,
      "loss": 1.4134,
      "theoretical_loss": 3.489839627826797,
      "tokens_seen": 1797586944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045985957873620866,
      "loss": 1.3264,
      "theoretical_loss": 3.489828884236135,
      "tokens_seen": 1797652480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459839518555667,
      "loss": 1.2744,
      "theoretical_loss": 3.489818141146803,
      "tokens_seen": 1797718016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598194583751254,
      "loss": 1.4155,
      "theoretical_loss": 3.489807398558759,
      "tokens_seen": 1797783552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597993981945838,
      "loss": 1.3576,
      "theoretical_loss": 3.4897966564719614,
      "tokens_seen": 1797849088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045977933801404216,
      "loss": 1.5046,
      "theoretical_loss": 3.489785914886369,
      "tokens_seen": 1797914624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597592778335005,
      "loss": 1.3326,
      "theoretical_loss": 3.489775173801939,
      "tokens_seen": 1797980160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597392176529589,
      "loss": 1.2927,
      "theoretical_loss": 3.489764433218631,
      "tokens_seen": 1798045696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045971915747241725,
      "loss": 1.4271,
      "theoretical_loss": 3.489753693136402,
      "tokens_seen": 1798111232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045969909729187566,
      "loss": 1.372,
      "theoretical_loss": 3.4897429535552122,
      "tokens_seen": 1798176768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459679037111334,
      "loss": 1.2677,
      "theoretical_loss": 3.4897322144750182,
      "tokens_seen": 1798242304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596589769307924,
      "loss": 1.3884,
      "theoretical_loss": 3.4897214758957795,
      "tokens_seen": 1798307840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045963891675025075,
      "loss": 1.4863,
      "theoretical_loss": 3.489710737817454,
      "tokens_seen": 1798373376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045961885656970917,
      "loss": 1.3901,
      "theoretical_loss": 3.48970000024,
      "tokens_seen": 1798438912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045959879638916753,
      "loss": 1.3629,
      "theoretical_loss": 3.4896892631633762,
      "tokens_seen": 1798504448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595787362086259,
      "loss": 1.3527,
      "theoretical_loss": 3.489678526587541,
      "tokens_seen": 1798569984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045955867602808425,
      "loss": 1.4832,
      "theoretical_loss": 3.4896677905124527,
      "tokens_seen": 1798635520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595386158475426,
      "loss": 1.3014,
      "theoretical_loss": 3.4896570549380694,
      "tokens_seen": 1798701056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045951855566700103,
      "loss": 1.3594,
      "theoretical_loss": 3.4896463198643497,
      "tokens_seen": 1798766592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594984954864594,
      "loss": 1.4096,
      "theoretical_loss": 3.4896355852912526,
      "tokens_seen": 1798832128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045947843530591776,
      "loss": 1.3733,
      "theoretical_loss": 3.4896248512187356,
      "tokens_seen": 1798897664
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594583751253761,
      "loss": 1.3307,
      "theoretical_loss": 3.4896141176467577,
      "tokens_seen": 1798963200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045943831494483454,
      "loss": 1.3302,
      "theoretical_loss": 3.4896033845752763,
      "tokens_seen": 1799028736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594182547642929,
      "loss": 1.3078,
      "theoretical_loss": 3.4895926520042515,
      "tokens_seen": 1799094272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045939819458375126,
      "loss": 1.2989,
      "theoretical_loss": 3.4895819199336406,
      "tokens_seen": 1799159808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593781344032096,
      "loss": 1.2548,
      "theoretical_loss": 3.489571188363403,
      "tokens_seen": 1799225344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459358074222668,
      "loss": 1.3335,
      "theoretical_loss": 3.4895604572934955,
      "tokens_seen": 1799290880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593380140421264,
      "loss": 1.3241,
      "theoretical_loss": 3.489549726723878,
      "tokens_seen": 1799356416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045931795386158476,
      "loss": 1.3731,
      "theoretical_loss": 3.4895389966545087,
      "tokens_seen": 1799421952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592978936810431,
      "loss": 1.2981,
      "theoretical_loss": 3.489528267085346,
      "tokens_seen": 1799487488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592778335005015,
      "loss": 1.4009,
      "theoretical_loss": 3.4895175380163472,
      "tokens_seen": 1799553024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592577733199599,
      "loss": 1.277,
      "theoretical_loss": 3.4895068094474726,
      "tokens_seen": 1799618560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045923771313941827,
      "loss": 1.2825,
      "theoretical_loss": 3.4894960813786797,
      "tokens_seen": 1799684096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045921765295887663,
      "loss": 1.3936,
      "theoretical_loss": 3.4894853538099273,
      "tokens_seen": 1799749632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459197592778335,
      "loss": 1.337,
      "theoretical_loss": 3.489474626741174,
      "tokens_seen": 1799815168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045917753259779335,
      "loss": 1.4326,
      "theoretical_loss": 3.4894639001723773,
      "tokens_seen": 1799880704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045915747241725177,
      "loss": 1.3389,
      "theoretical_loss": 3.4894531741034966,
      "tokens_seen": 1799946240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045913741223671013,
      "loss": 1.3869,
      "theoretical_loss": 3.489442448534491,
      "tokens_seen": 1800011776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004591173520561685,
      "loss": 1.3062,
      "theoretical_loss": 3.489431723465317,
      "tokens_seen": 1800077312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045909729187562686,
      "loss": 1.2805,
      "theoretical_loss": 3.489420998895935,
      "tokens_seen": 1800142848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004590772316950853,
      "loss": 1.334,
      "theoretical_loss": 3.489410274826303,
      "tokens_seen": 1800208384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045905717151454364,
      "loss": 1.4646,
      "theoretical_loss": 3.489399551256379,
      "tokens_seen": 1800273920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459037111334002,
      "loss": 1.4175,
      "theoretical_loss": 3.4893888281861223,
      "tokens_seen": 1800339456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045901705115346036,
      "loss": 1.5074,
      "theoretical_loss": 3.489378105615491,
      "tokens_seen": 1800404992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589969909729187,
      "loss": 1.3638,
      "theoretical_loss": 3.489367383544443,
      "tokens_seen": 1800470528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045897693079237714,
      "loss": 1.2953,
      "theoretical_loss": 3.489356661972938,
      "tokens_seen": 1800536064
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589568706118355,
      "loss": 1.4398,
      "theoretical_loss": 3.489345940900934,
      "tokens_seen": 1800601600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045893681043129386,
      "loss": 1.4388,
      "theoretical_loss": 3.48933522032839,
      "tokens_seen": 1800667136
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 657521,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3838400840759277,
      "objective/train/theoretical_loss": 3.4893245002552638,
      "objective/train/tokens_used": 171191776,
      "theoretical_loss": 3.4893245002552638,
      "tokens_seen": 1800732672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589167502507522,
      "loss": 1.4321,
      "theoretical_loss": 3.4893245002552638,
      "tokens_seen": 1800732672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045889669007021064,
      "loss": 1.2989,
      "theoretical_loss": 3.489313780681514,
      "tokens_seen": 1800798208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458876629889669,
      "loss": 1.3539,
      "theoretical_loss": 3.4893030616070995,
      "tokens_seen": 1800863744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045885656970912737,
      "loss": 1.3443,
      "theoretical_loss": 3.4892923430319795,
      "tokens_seen": 1800929280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045883650952858573,
      "loss": 1.3797,
      "theoretical_loss": 3.4892816249561114,
      "tokens_seen": 1800994816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045881644934804415,
      "loss": 1.3086,
      "theoretical_loss": 3.4892709073794546,
      "tokens_seen": 1801060352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587963891675025,
      "loss": 1.3856,
      "theoretical_loss": 3.489260190301967,
      "tokens_seen": 1801125888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045877632898696087,
      "loss": 1.4438,
      "theoretical_loss": 3.489249473723608,
      "tokens_seen": 1801191424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045875626880641923,
      "loss": 1.3116,
      "theoretical_loss": 3.4892387576443356,
      "tokens_seen": 1801256960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587362086258776,
      "loss": 1.2708,
      "theoretical_loss": 3.4892280420641084,
      "tokens_seen": 1801322496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458716148445336,
      "loss": 1.3552,
      "theoretical_loss": 3.4892173269828852,
      "tokens_seen": 1801388032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586960882647944,
      "loss": 1.4093,
      "theoretical_loss": 3.489206612400625,
      "tokens_seen": 1801453568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586760280842528,
      "loss": 1.3794,
      "theoretical_loss": 3.489195898317286,
      "tokens_seen": 1801519104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045865596790371115,
      "loss": 1.4323,
      "theoretical_loss": 3.4891851847328264,
      "tokens_seen": 1801584640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045863590772316957,
      "loss": 1.3559,
      "theoretical_loss": 3.489174471647205,
      "tokens_seen": 1801650176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045861584754262793,
      "loss": 1.3762,
      "theoretical_loss": 3.4891637590603812,
      "tokens_seen": 1801715712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585957873620863,
      "loss": 1.332,
      "theoretical_loss": 3.489153046972313,
      "tokens_seen": 1801781248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045857572718154465,
      "loss": 1.4762,
      "theoretical_loss": 3.4891423353829594,
      "tokens_seen": 1801846784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458555667001003,
      "loss": 1.3516,
      "theoretical_loss": 3.4891316242922787,
      "tokens_seen": 1801912320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045853560682046143,
      "loss": 1.4266,
      "theoretical_loss": 3.489120913700229,
      "tokens_seen": 1801977856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585155466399198,
      "loss": 1.3399,
      "theoretical_loss": 3.4891102036067703,
      "tokens_seen": 1802043392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045849548645937816,
      "loss": 1.3181,
      "theoretical_loss": 3.4890994940118603,
      "tokens_seen": 1802108928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584754262788365,
      "loss": 1.3428,
      "theoretical_loss": 3.4890887849154577,
      "tokens_seen": 1802174464
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045845536609829494,
      "loss": 1.2375,
      "theoretical_loss": 3.4890780763175218,
      "tokens_seen": 1802240000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584353059177533,
      "loss": 1.3274,
      "theoretical_loss": 3.4890673682180107,
      "tokens_seen": 1802305536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045841524573721166,
      "loss": 1.2824,
      "theoretical_loss": 3.489056660616883,
      "tokens_seen": 1802371072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045839518555667,
      "loss": 1.2304,
      "theoretical_loss": 3.489045953514098,
      "tokens_seen": 1802436608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583751253761284,
      "loss": 1.4738,
      "theoretical_loss": 3.489035246909614,
      "tokens_seen": 1802502144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583550651955868,
      "loss": 1.4322,
      "theoretical_loss": 3.4890245408033893,
      "tokens_seen": 1802567680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045833500501504516,
      "loss": 1.3753,
      "theoretical_loss": 3.489013835195383,
      "tokens_seen": 1802633216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583149448345035,
      "loss": 1.3823,
      "theoretical_loss": 3.489003130085554,
      "tokens_seen": 1802698752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582948846539619,
      "loss": 1.3556,
      "theoretical_loss": 3.488992425473861,
      "tokens_seen": 1802764288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582748244734203,
      "loss": 1.4368,
      "theoretical_loss": 3.4889817213602625,
      "tokens_seen": 1802829824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045825476429287867,
      "loss": 1.3289,
      "theoretical_loss": 3.488971017744717,
      "tokens_seen": 1802895360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045823470411233703,
      "loss": 1.2775,
      "theoretical_loss": 3.4889603146271835,
      "tokens_seen": 1802960896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582146439317954,
      "loss": 1.437,
      "theoretical_loss": 3.488949612007621,
      "tokens_seen": 1803026432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045819458375125375,
      "loss": 1.4029,
      "theoretical_loss": 3.488938909885988,
      "tokens_seen": 1803091968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045817452357071217,
      "loss": 1.3429,
      "theoretical_loss": 3.4889282082622426,
      "tokens_seen": 1803157504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045815446339017053,
      "loss": 1.4094,
      "theoretical_loss": 3.4889175071363443,
      "tokens_seen": 1803223040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581344032096289,
      "loss": 1.4227,
      "theoretical_loss": 3.4889068065082522,
      "tokens_seen": 1803288576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045811434302908726,
      "loss": 1.3736,
      "theoretical_loss": 3.488896106377924,
      "tokens_seen": 1803354112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580942828485457,
      "loss": 1.2447,
      "theoretical_loss": 3.488885406745319,
      "tokens_seen": 1803419648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045807422266800404,
      "loss": 1.4373,
      "theoretical_loss": 3.4888747076103956,
      "tokens_seen": 1803485184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580541624874624,
      "loss": 1.2957,
      "theoretical_loss": 3.4888640089731133,
      "tokens_seen": 1803550720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045803410230692076,
      "loss": 1.3422,
      "theoretical_loss": 3.4888533108334303,
      "tokens_seen": 1803616256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580140421263791,
      "loss": 1.3404,
      "theoretical_loss": 3.4888426131913057,
      "tokens_seen": 1803681792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045799398194583754,
      "loss": 1.2702,
      "theoretical_loss": 3.4888319160466983,
      "tokens_seen": 1803747328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579739217652959,
      "loss": 1.3288,
      "theoretical_loss": 3.488821219399566,
      "tokens_seen": 1803812864
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045795386158475426,
      "loss": 1.4499,
      "theoretical_loss": 3.488810523249869,
      "tokens_seen": 1803878400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579338014042126,
      "loss": 1.4317,
      "theoretical_loss": 3.488799827597565,
      "tokens_seen": 1803943936
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 658552,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3652687072753906,
      "objective/train/theoretical_loss": 3.4887891324426135,
      "objective/train/tokens_used": 174468576,
      "theoretical_loss": 3.4887891324426135,
      "tokens_seen": 1804009472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045791374122367104,
      "loss": 1.4031,
      "theoretical_loss": 3.4887891324426135,
      "tokens_seen": 1804009472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578936810431294,
      "loss": 1.2971,
      "theoretical_loss": 3.4887784377849727,
      "tokens_seen": 1804075008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045787362086258777,
      "loss": 1.3763,
      "theoretical_loss": 3.488767743624602,
      "tokens_seen": 1804140544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045785356068204613,
      "loss": 1.2835,
      "theoretical_loss": 3.4887570499614595,
      "tokens_seen": 1804206080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045783350050150455,
      "loss": 1.4371,
      "theoretical_loss": 3.488746356795505,
      "tokens_seen": 1804271616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578134403209629,
      "loss": 1.2313,
      "theoretical_loss": 3.488735664126697,
      "tokens_seen": 1804337152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045779338014042127,
      "loss": 1.3781,
      "theoretical_loss": 3.4887249719549933,
      "tokens_seen": 1804402688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045777331995987963,
      "loss": 1.2699,
      "theoretical_loss": 3.488714280280354,
      "tokens_seen": 1804468224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457753259779338,
      "loss": 1.4195,
      "theoretical_loss": 3.4887035891027374,
      "tokens_seen": 1804533760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577331995987964,
      "loss": 1.3151,
      "theoretical_loss": 3.4886928984221024,
      "tokens_seen": 1804599296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577131394182548,
      "loss": 1.3726,
      "theoretical_loss": 3.4886822082384077,
      "tokens_seen": 1804664832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045769307923771314,
      "loss": 1.4224,
      "theoretical_loss": 3.4886715185516124,
      "tokens_seen": 1804730368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576730190571715,
      "loss": 1.3777,
      "theoretical_loss": 3.4886608293616757,
      "tokens_seen": 1804795904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576529588766299,
      "loss": 1.2689,
      "theoretical_loss": 3.4886501406685557,
      "tokens_seen": 1804861440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576328986960883,
      "loss": 1.2477,
      "theoretical_loss": 3.4886394524722117,
      "tokens_seen": 1804926976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045761283851554664,
      "loss": 1.392,
      "theoretical_loss": 3.488628764772603,
      "tokens_seen": 1804992512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457592778335005,
      "loss": 1.2839,
      "theoretical_loss": 3.4886180775696873,
      "tokens_seen": 1805058048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045757271815446336,
      "loss": 1.3361,
      "theoretical_loss": 3.4886073908634243,
      "tokens_seen": 1805123584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575526579739218,
      "loss": 1.3321,
      "theoretical_loss": 3.488596704653773,
      "tokens_seen": 1805189120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045753259779338014,
      "loss": 1.3662,
      "theoretical_loss": 3.488586018940692,
      "tokens_seen": 1805254656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575125376128385,
      "loss": 1.3051,
      "theoretical_loss": 3.4885753337241403,
      "tokens_seen": 1805320192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045749247743229687,
      "loss": 1.3048,
      "theoretical_loss": 3.4885646490040765,
      "tokens_seen": 1805385728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574724172517553,
      "loss": 1.5203,
      "theoretical_loss": 3.48855396478046,
      "tokens_seen": 1805451264
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045745235707121364,
      "loss": 1.2229,
      "theoretical_loss": 3.4885432810532495,
      "tokens_seen": 1805516800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457432296890672,
      "loss": 1.4225,
      "theoretical_loss": 3.4885325978224038,
      "tokens_seen": 1805582336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045741223671013037,
      "loss": 1.3852,
      "theoretical_loss": 3.488521915087882,
      "tokens_seen": 1805647872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045739217652958873,
      "loss": 1.3862,
      "theoretical_loss": 3.4885112328496426,
      "tokens_seen": 1805713408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045737211634904715,
      "loss": 1.3808,
      "theoretical_loss": 3.4885005511076455,
      "tokens_seen": 1805778944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004573520561685055,
      "loss": 1.4036,
      "theoretical_loss": 3.4884898698618487,
      "tokens_seen": 1805844480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045733199598796387,
      "loss": 1.2988,
      "theoretical_loss": 3.4884791891122116,
      "tokens_seen": 1805910016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045731193580742223,
      "loss": 1.5018,
      "theoretical_loss": 3.488468508858693,
      "tokens_seen": 1805975552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045729187562688065,
      "loss": 1.2976,
      "theoretical_loss": 3.488457829101252,
      "tokens_seen": 1806041088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457271815446339,
      "loss": 1.3444,
      "theoretical_loss": 3.488447149839847,
      "tokens_seen": 1806106624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572517552657974,
      "loss": 1.4244,
      "theoretical_loss": 3.4884364710744378,
      "tokens_seen": 1806172160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045723169508525574,
      "loss": 1.3512,
      "theoretical_loss": 3.488425792804983,
      "tokens_seen": 1806237696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572116349047141,
      "loss": 1.2197,
      "theoretical_loss": 3.488415115031441,
      "tokens_seen": 1806303232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045719157472417257,
      "loss": 1.3635,
      "theoretical_loss": 3.488404437753772,
      "tokens_seen": 1806368768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045717151454363093,
      "loss": 1.3916,
      "theoretical_loss": 3.4883937609719338,
      "tokens_seen": 1806434304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571514543630893,
      "loss": 1.5014,
      "theoretical_loss": 3.4883830846858856,
      "tokens_seen": 1806499840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045713139418254766,
      "loss": 1.3277,
      "theoretical_loss": 3.4883724088955876,
      "tokens_seen": 1806565376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571113340020061,
      "loss": 1.4299,
      "theoretical_loss": 3.488361733600997,
      "tokens_seen": 1806630912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045709127382146444,
      "loss": 1.4107,
      "theoretical_loss": 3.488351058802074,
      "tokens_seen": 1806696448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004570712136409228,
      "loss": 1.4041,
      "theoretical_loss": 3.4883403844987777,
      "tokens_seen": 1806761984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045705115346038116,
      "loss": 1.3683,
      "theoretical_loss": 3.4883297106910662,
      "tokens_seen": 1806827520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004570310932798395,
      "loss": 1.3187,
      "theoretical_loss": 3.4883190373788997,
      "tokens_seen": 1806893056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045701103309929794,
      "loss": 1.3595,
      "theoretical_loss": 3.4883083645622355,
      "tokens_seen": 1806958592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004569909729187563,
      "loss": 1.3336,
      "theoretical_loss": 3.4882976922410345,
      "tokens_seen": 1807024128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045697091273821466,
      "loss": 1.4033,
      "theoretical_loss": 3.4882870204152545,
      "tokens_seen": 1807089664
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456950852557673,
      "loss": 1.2869,
      "theoretical_loss": 3.488276349084855,
      "tokens_seen": 1807155200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045693079237713144,
      "loss": 1.3787,
      "theoretical_loss": 3.488265678249795,
      "tokens_seen": 1807220736
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 659683,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4606060981750488,
      "objective/train/theoretical_loss": 3.4882550079100336,
      "objective/train/tokens_used": 177745376,
      "theoretical_loss": 3.4882550079100336,
      "tokens_seen": 1807286272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004569107321965898,
      "loss": 1.3264,
      "theoretical_loss": 3.4882550079100336,
      "tokens_seen": 1807286272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045689067201604817,
      "loss": 1.3305,
      "theoretical_loss": 3.4882443380655297,
      "tokens_seen": 1807351808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045687061183550653,
      "loss": 1.4957,
      "theoretical_loss": 3.4882336687162425,
      "tokens_seen": 1807417344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045685055165496495,
      "loss": 1.4547,
      "theoretical_loss": 3.4882229998621304,
      "tokens_seen": 1807482880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004568304914744233,
      "loss": 1.3334,
      "theoretical_loss": 3.4882123315031537,
      "tokens_seen": 1807548416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045681043129388167,
      "loss": 1.4867,
      "theoretical_loss": 3.4882016636392708,
      "tokens_seen": 1807613952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045679037111334003,
      "loss": 1.317,
      "theoretical_loss": 3.4881909962704407,
      "tokens_seen": 1807679488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567703109327984,
      "loss": 1.3876,
      "theoretical_loss": 3.4881803293966227,
      "tokens_seen": 1807745024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567502507522568,
      "loss": 1.3145,
      "theoretical_loss": 3.488169663017775,
      "tokens_seen": 1807810560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567301905717152,
      "loss": 1.3685,
      "theoretical_loss": 3.4881589971338585,
      "tokens_seen": 1807876096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045671013039117354,
      "loss": 1.2868,
      "theoretical_loss": 3.4881483317448305,
      "tokens_seen": 1807941632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566900702106319,
      "loss": 1.3468,
      "theoretical_loss": 3.488137666850651,
      "tokens_seen": 1808007168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566700100300903,
      "loss": 1.351,
      "theoretical_loss": 3.488127002451279,
      "tokens_seen": 1808072704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566499498495487,
      "loss": 1.391,
      "theoretical_loss": 3.488116338546674,
      "tokens_seen": 1808138240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045662988966900704,
      "loss": 1.4638,
      "theoretical_loss": 3.488105675136794,
      "tokens_seen": 1808203776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566098294884654,
      "loss": 1.4516,
      "theoretical_loss": 3.4880950122215992,
      "tokens_seen": 1808269312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045658976930792376,
      "loss": 1.281,
      "theoretical_loss": 3.4880843498010483,
      "tokens_seen": 1808334848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565697091273822,
      "loss": 1.3846,
      "theoretical_loss": 3.4880736878751,
      "tokens_seen": 1808400384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045654964894684054,
      "loss": 1.314,
      "theoretical_loss": 3.4880630264437142,
      "tokens_seen": 1808465920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565295887662989,
      "loss": 1.3936,
      "theoretical_loss": 3.48805236550685,
      "tokens_seen": 1808531456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045650952858575727,
      "loss": 1.4567,
      "theoretical_loss": 3.488041705064466,
      "tokens_seen": 1808596992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564894684052157,
      "loss": 1.3893,
      "theoretical_loss": 3.4880310451165215,
      "tokens_seen": 1808662528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045646940822467405,
      "loss": 1.3522,
      "theoretical_loss": 3.488020385662976,
      "tokens_seen": 1808728064
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564493480441324,
      "loss": 1.3825,
      "theoretical_loss": 3.4880097267037877,
      "tokens_seen": 1808793600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045642928786359077,
      "loss": 1.3196,
      "theoretical_loss": 3.4879990682389175,
      "tokens_seen": 1808859136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045640922768304913,
      "loss": 1.2819,
      "theoretical_loss": 3.487988410268323,
      "tokens_seen": 1808924672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045638916750250755,
      "loss": 1.4451,
      "theoretical_loss": 3.4879777527919638,
      "tokens_seen": 1808990208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004563691073219659,
      "loss": 1.3774,
      "theoretical_loss": 3.4879670958097995,
      "tokens_seen": 1809055744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045634904714142427,
      "loss": 1.3554,
      "theoretical_loss": 3.4879564393217892,
      "tokens_seen": 1809121280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045632898696088263,
      "loss": 1.4159,
      "theoretical_loss": 3.4879457833278913,
      "tokens_seen": 1809186816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045630892678034105,
      "loss": 1.3398,
      "theoretical_loss": 3.4879351278280657,
      "tokens_seen": 1809252352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562888665997994,
      "loss": 1.3942,
      "theoretical_loss": 3.4879244728222716,
      "tokens_seen": 1809317888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562688064192578,
      "loss": 1.4348,
      "theoretical_loss": 3.487913818310468,
      "tokens_seen": 1809383424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045624874623871614,
      "loss": 1.3525,
      "theoretical_loss": 3.4879031642926144,
      "tokens_seen": 1809448960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562286860581745,
      "loss": 1.319,
      "theoretical_loss": 3.4878925107686696,
      "tokens_seen": 1809514496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562086258776329,
      "loss": 1.4513,
      "theoretical_loss": 3.487881857738593,
      "tokens_seen": 1809580032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561885656970913,
      "loss": 1.3961,
      "theoretical_loss": 3.4878712052023433,
      "tokens_seen": 1809645568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045616850551654964,
      "loss": 1.327,
      "theoretical_loss": 3.487860553159881,
      "tokens_seen": 1809711104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456148445336008,
      "loss": 1.3921,
      "theoretical_loss": 3.487849901611164,
      "tokens_seen": 1809776640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561283851554664,
      "loss": 1.3918,
      "theoretical_loss": 3.487839250556153,
      "tokens_seen": 1809842176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561083249749248,
      "loss": 1.2764,
      "theoretical_loss": 3.4878285999948053,
      "tokens_seen": 1809907712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045608826479438314,
      "loss": 1.3769,
      "theoretical_loss": 3.4878179499270816,
      "tokens_seen": 1809973248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004560682046138415,
      "loss": 1.303,
      "theoretical_loss": 3.487807300352941,
      "tokens_seen": 1810038784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045604814443329987,
      "loss": 1.3821,
      "theoretical_loss": 3.487796651272342,
      "tokens_seen": 1810104320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004560280842527583,
      "loss": 1.3369,
      "theoretical_loss": 3.4877860026852443,
      "tokens_seen": 1810169856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045600802407221665,
      "loss": 1.3372,
      "theoretical_loss": 3.4877753545916077,
      "tokens_seen": 1810235392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455987963891675,
      "loss": 1.4035,
      "theoretical_loss": 3.487764706991391,
      "tokens_seen": 1810300928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045596790371113337,
      "loss": 1.3171,
      "theoretical_loss": 3.487754059884553,
      "tokens_seen": 1810366464
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004559478435305918,
      "loss": 1.3921,
      "theoretical_loss": 3.487743413271054,
      "tokens_seen": 1810432000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045592778335005015,
      "loss": 1.362,
      "theoretical_loss": 3.487732767150852,
      "tokens_seen": 1810497536
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 660859,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3827738761901855,
      "objective/train/theoretical_loss": 3.487722121523907,
      "objective/train/tokens_used": 181022176,
      "theoretical_loss": 3.487722121523907,
      "tokens_seen": 1810563072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004559077231695085,
      "loss": 1.3172,
      "theoretical_loss": 3.487722121523907,
      "tokens_seen": 1810563072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558876629889669,
      "loss": 1.3648,
      "theoretical_loss": 3.487711476390179,
      "tokens_seen": 1810628608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045586760280842524,
      "loss": 1.4288,
      "theoretical_loss": 3.4877008317496268,
      "tokens_seen": 1810694144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045584754262788365,
      "loss": 1.4121,
      "theoretical_loss": 3.4876901876022086,
      "tokens_seen": 1810759680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455827482447342,
      "loss": 1.3978,
      "theoretical_loss": 3.487679543947885,
      "tokens_seen": 1810825216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558074222668004,
      "loss": 1.4279,
      "theoretical_loss": 3.4876689007866153,
      "tokens_seen": 1810890752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045578736208625874,
      "loss": 1.2723,
      "theoretical_loss": 3.4876582581183575,
      "tokens_seen": 1810956288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045576730190571716,
      "loss": 1.4098,
      "theoretical_loss": 3.4876476159430725,
      "tokens_seen": 1811021824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557472417251755,
      "loss": 1.3927,
      "theoretical_loss": 3.4876369742607194,
      "tokens_seen": 1811087360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557271815446339,
      "loss": 1.3357,
      "theoretical_loss": 3.4876263330712565,
      "tokens_seen": 1811152896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045570712136409224,
      "loss": 1.4501,
      "theoretical_loss": 3.4876156923746437,
      "tokens_seen": 1811218432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556870611835507,
      "loss": 1.3736,
      "theoretical_loss": 3.487605052170841,
      "tokens_seen": 1811283968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556670010030091,
      "loss": 1.4069,
      "theoretical_loss": 3.4875944124598064,
      "tokens_seen": 1811349504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045564694082246744,
      "loss": 1.3743,
      "theoretical_loss": 3.487583773241501,
      "tokens_seen": 1811415040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556268806419258,
      "loss": 1.4264,
      "theoretical_loss": 3.4875731345158822,
      "tokens_seen": 1811480576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045560682046138416,
      "loss": 1.4066,
      "theoretical_loss": 3.4875624962829104,
      "tokens_seen": 1811546112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555867602808426,
      "loss": 1.3675,
      "theoretical_loss": 3.4875518585425453,
      "tokens_seen": 1811611648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045556670010030094,
      "loss": 1.3768,
      "theoretical_loss": 3.487541221294746,
      "tokens_seen": 1811677184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555466399197593,
      "loss": 1.3791,
      "theoretical_loss": 3.4875305845394715,
      "tokens_seen": 1811742720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045552657973921767,
      "loss": 1.2643,
      "theoretical_loss": 3.487519948276681,
      "tokens_seen": 1811808256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555065195586761,
      "loss": 1.443,
      "theoretical_loss": 3.487509312506335,
      "tokens_seen": 1811873792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045548645937813445,
      "loss": 1.3575,
      "theoretical_loss": 3.4874986772283916,
      "tokens_seen": 1811939328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554663991975928,
      "loss": 1.4148,
      "theoretical_loss": 3.487488042442811,
      "tokens_seen": 1812004864
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045544633901705117,
      "loss": 1.3839,
      "theoretical_loss": 3.4874774081495525,
      "tokens_seen": 1812070400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045542627883650953,
      "loss": 1.3475,
      "theoretical_loss": 3.487466774348575,
      "tokens_seen": 1812135936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045540621865596795,
      "loss": 1.3223,
      "theoretical_loss": 3.4874561410398393,
      "tokens_seen": 1812201472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553861584754263,
      "loss": 1.4257,
      "theoretical_loss": 3.4874455082233027,
      "tokens_seen": 1812267008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553660982948847,
      "loss": 1.4016,
      "theoretical_loss": 3.487434875898926,
      "tokens_seen": 1812332544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045534603811434304,
      "loss": 1.2334,
      "theoretical_loss": 3.4874242440666685,
      "tokens_seen": 1812398080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045532597793380145,
      "loss": 1.3541,
      "theoretical_loss": 3.487413612726489,
      "tokens_seen": 1812463616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553059177532598,
      "loss": 1.3471,
      "theoretical_loss": 3.4874029818783487,
      "tokens_seen": 1812529152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552858575727182,
      "loss": 1.3846,
      "theoretical_loss": 3.4873923515222045,
      "tokens_seen": 1812594688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045526579739217654,
      "loss": 1.4484,
      "theoretical_loss": 3.487381721658018,
      "tokens_seen": 1812660224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552457372116349,
      "loss": 1.3441,
      "theoretical_loss": 3.487371092285747,
      "tokens_seen": 1812725760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552256770310933,
      "loss": 1.4135,
      "theoretical_loss": 3.487360463405352,
      "tokens_seen": 1812791296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552056168505517,
      "loss": 1.4303,
      "theoretical_loss": 3.4873498350167917,
      "tokens_seen": 1812856832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045518555667001004,
      "loss": 1.2703,
      "theoretical_loss": 3.4873392071200264,
      "tokens_seen": 1812922368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551654964894684,
      "loss": 1.346,
      "theoretical_loss": 3.4873285797150153,
      "tokens_seen": 1812987904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551454363089268,
      "loss": 1.2833,
      "theoretical_loss": 3.4873179528017175,
      "tokens_seen": 1813053440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551253761283852,
      "loss": 1.276,
      "theoretical_loss": 3.487307326380093,
      "tokens_seen": 1813118976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045510531594784354,
      "loss": 1.3792,
      "theoretical_loss": 3.487296700450101,
      "tokens_seen": 1813184512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550852557673019,
      "loss": 1.409,
      "theoretical_loss": 3.487286075011701,
      "tokens_seen": 1813250048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045506519558676027,
      "loss": 1.3866,
      "theoretical_loss": 3.4872754500648524,
      "tokens_seen": 1813315584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550451354062187,
      "loss": 1.3271,
      "theoretical_loss": 3.4872648256095147,
      "tokens_seen": 1813381120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045502507522567705,
      "loss": 1.351,
      "theoretical_loss": 3.487254201645648,
      "tokens_seen": 1813446656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550050150451354,
      "loss": 1.3199,
      "theoretical_loss": 3.4872435781732105,
      "tokens_seen": 1813512192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045498495486459377,
      "loss": 1.3552,
      "theoretical_loss": 3.487232955192163,
      "tokens_seen": 1813577728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549648946840522,
      "loss": 1.3602,
      "theoretical_loss": 3.487222332702464,
      "tokens_seen": 1813643264
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045494483450351055,
      "loss": 1.4444,
      "theoretical_loss": 3.487211710704074,
      "tokens_seen": 1813708800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549247743229689,
      "loss": 1.2832,
      "theoretical_loss": 3.487201089196952,
      "tokens_seen": 1813774336
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 661946,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.294579029083252,
      "objective/train/theoretical_loss": 3.4871904681810575,
      "objective/train/tokens_used": 184298976,
      "theoretical_loss": 3.4871904681810575,
      "tokens_seen": 1813839872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549047141424273,
      "loss": 1.2835,
      "theoretical_loss": 3.4871904681810575,
      "tokens_seen": 1813839872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045488465396188564,
      "loss": 1.4588,
      "theoretical_loss": 3.48717984765635,
      "tokens_seen": 1813905408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045486459378134405,
      "loss": 1.4644,
      "theoretical_loss": 3.4871692276227897,
      "tokens_seen": 1813970944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548445336008024,
      "loss": 1.4076,
      "theoretical_loss": 3.487158608080335,
      "tokens_seen": 1814036480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548244734202608,
      "loss": 1.3952,
      "theoretical_loss": 3.487147989028946,
      "tokens_seen": 1814102016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045480441323971914,
      "loss": 1.3131,
      "theoretical_loss": 3.4871373704685826,
      "tokens_seen": 1814167552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045478435305917756,
      "loss": 1.4066,
      "theoretical_loss": 3.487126752399204,
      "tokens_seen": 1814233088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547642928786359,
      "loss": 1.3782,
      "theoretical_loss": 3.48711613482077,
      "tokens_seen": 1814298624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547442326980943,
      "loss": 1.2663,
      "theoretical_loss": 3.48710551773324,
      "tokens_seen": 1814364160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045472417251755264,
      "loss": 1.4287,
      "theoretical_loss": 3.4870949011365733,
      "tokens_seen": 1814429696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454704112337011,
      "loss": 1.431,
      "theoretical_loss": 3.48708428503073,
      "tokens_seen": 1814495232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546840521564694,
      "loss": 1.4615,
      "theoretical_loss": 3.487073669415669,
      "tokens_seen": 1814560768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546639919759278,
      "loss": 1.447,
      "theoretical_loss": 3.487063054291351,
      "tokens_seen": 1814626304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045464393179538615,
      "loss": 1.2898,
      "theoretical_loss": 3.4870524396577345,
      "tokens_seen": 1814691840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546238716148445,
      "loss": 1.4077,
      "theoretical_loss": 3.4870418255147793,
      "tokens_seen": 1814757376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546038114343029,
      "loss": 1.5181,
      "theoretical_loss": 3.487031211862446,
      "tokens_seen": 1814822912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545837512537613,
      "loss": 1.3867,
      "theoretical_loss": 3.4870205987006924,
      "tokens_seen": 1814888448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045456369107321965,
      "loss": 1.3475,
      "theoretical_loss": 3.48700998602948,
      "tokens_seen": 1814953984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454543630892678,
      "loss": 1.3641,
      "theoretical_loss": 3.486999373848767,
      "tokens_seen": 1815019520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045452357071213643,
      "loss": 1.4215,
      "theoretical_loss": 3.4869887621585134,
      "tokens_seen": 1815085056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545035105315948,
      "loss": 1.3982,
      "theoretical_loss": 3.486978150958679,
      "tokens_seen": 1815150592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045448345035105315,
      "loss": 1.4663,
      "theoretical_loss": 3.486967540249224,
      "tokens_seen": 1815216128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544633901705115,
      "loss": 1.3237,
      "theoretical_loss": 3.486956930030107,
      "tokens_seen": 1815281664
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544433299899699,
      "loss": 1.3386,
      "theoretical_loss": 3.4869463203012883,
      "tokens_seen": 1815347200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544232698094283,
      "loss": 1.389,
      "theoretical_loss": 3.486935711062727,
      "tokens_seen": 1815412736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045440320962888666,
      "loss": 1.3784,
      "theoretical_loss": 3.4869251023143835,
      "tokens_seen": 1815478272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454383149448345,
      "loss": 1.4512,
      "theoretical_loss": 3.486914494056217,
      "tokens_seen": 1815543808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543630892678034,
      "loss": 1.3793,
      "theoretical_loss": 3.4869038862881867,
      "tokens_seen": 1815609344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543430290872618,
      "loss": 1.3004,
      "theoretical_loss": 3.4868932790102534,
      "tokens_seen": 1815674880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045432296890672016,
      "loss": 1.3104,
      "theoretical_loss": 3.4868826722223756,
      "tokens_seen": 1815740416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543029087261785,
      "loss": 1.4026,
      "theoretical_loss": 3.486872065924514,
      "tokens_seen": 1815805952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004542828485456369,
      "loss": 1.3596,
      "theoretical_loss": 3.486861460116627,
      "tokens_seen": 1815871488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045426278836509525,
      "loss": 1.3373,
      "theoretical_loss": 3.4868508547986754,
      "tokens_seen": 1815937024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045424272818455366,
      "loss": 1.3375,
      "theoretical_loss": 3.486840249970619,
      "tokens_seen": 1816002560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454222668004012,
      "loss": 1.2957,
      "theoretical_loss": 3.4868296456324166,
      "tokens_seen": 1816068096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045420260782347044,
      "loss": 1.2682,
      "theoretical_loss": 3.4868190417840284,
      "tokens_seen": 1816133632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541825476429288,
      "loss": 1.3165,
      "theoretical_loss": 3.486808438425414,
      "tokens_seen": 1816199168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541624874623872,
      "loss": 1.4053,
      "theoretical_loss": 3.4867978355565334,
      "tokens_seen": 1816264704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541424272818456,
      "loss": 1.3571,
      "theoretical_loss": 3.4867872331773455,
      "tokens_seen": 1816330240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045412236710130395,
      "loss": 1.4024,
      "theoretical_loss": 3.4867766312878112,
      "tokens_seen": 1816395776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541023069207623,
      "loss": 1.4612,
      "theoretical_loss": 3.486766029887889,
      "tokens_seen": 1816461312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045408224674022067,
      "loss": 1.2546,
      "theoretical_loss": 3.4867554289775398,
      "tokens_seen": 1816526848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540621865596791,
      "loss": 1.3132,
      "theoretical_loss": 3.4867448285567226,
      "tokens_seen": 1816592384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045404212637913745,
      "loss": 1.4106,
      "theoretical_loss": 3.486734228625397,
      "tokens_seen": 1816657920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540220661985958,
      "loss": 1.3342,
      "theoretical_loss": 3.486723629183523,
      "tokens_seen": 1816723456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045400200601805417,
      "loss": 1.4007,
      "theoretical_loss": 3.4867130302310607,
      "tokens_seen": 1816788992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539819458375126,
      "loss": 1.3379,
      "theoretical_loss": 3.486702431767969,
      "tokens_seen": 1816854528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045396188565697095,
      "loss": 1.3287,
      "theoretical_loss": 3.4866918337942088,
      "tokens_seen": 1816920064
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539418254764293,
      "loss": 1.3228,
      "theoretical_loss": 3.486681236309739,
      "tokens_seen": 1816985600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539217652958877,
      "loss": 1.4388,
      "theoretical_loss": 3.4866706393145197,
      "tokens_seen": 1817051136
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 663068,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.466980218887329,
      "objective/train/theoretical_loss": 3.48666004280851,
      "objective/train/tokens_used": 187575776,
      "theoretical_loss": 3.48666004280851,
      "tokens_seen": 1817116672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045390170511534604,
      "loss": 1.477,
      "theoretical_loss": 3.48666004280851,
      "tokens_seen": 1817116672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045388164493480445,
      "loss": 1.3778,
      "theoretical_loss": 3.4866494467916707,
      "tokens_seen": 1817182208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538615847542628,
      "loss": 1.334,
      "theoretical_loss": 3.486638851263961,
      "tokens_seen": 1817247744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538415245737212,
      "loss": 1.3123,
      "theoretical_loss": 3.486628256225341,
      "tokens_seen": 1817313280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045382146439317954,
      "loss": 1.4716,
      "theoretical_loss": 3.48661766167577,
      "tokens_seen": 1817378816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045380140421263796,
      "loss": 1.2645,
      "theoretical_loss": 3.486607067615208,
      "tokens_seen": 1817444352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537813440320963,
      "loss": 1.427,
      "theoretical_loss": 3.486596474043615,
      "tokens_seen": 1817509888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537612838515547,
      "loss": 1.2429,
      "theoretical_loss": 3.4865858809609507,
      "tokens_seen": 1817575424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045374122367101304,
      "loss": 1.3738,
      "theoretical_loss": 3.486575288367175,
      "tokens_seen": 1817640960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537211634904714,
      "loss": 1.272,
      "theoretical_loss": 3.4865646962622474,
      "tokens_seen": 1817706496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537011033099298,
      "loss": 1.391,
      "theoretical_loss": 3.486554104646128,
      "tokens_seen": 1817772032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536810431293882,
      "loss": 1.3263,
      "theoretical_loss": 3.486543513518776,
      "tokens_seen": 1817837568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045366098294884655,
      "loss": 1.2763,
      "theoretical_loss": 3.4865329228801523,
      "tokens_seen": 1817903104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536409227683049,
      "loss": 1.2687,
      "theoretical_loss": 3.486522332730216,
      "tokens_seen": 1817968640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536208625877633,
      "loss": 1.4156,
      "theoretical_loss": 3.4865117430689274,
      "tokens_seen": 1818034176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536008024072217,
      "loss": 1.2561,
      "theoretical_loss": 3.4865011538962456,
      "tokens_seen": 1818099712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045358074222668005,
      "loss": 1.3223,
      "theoretical_loss": 3.486490565212131,
      "tokens_seen": 1818165248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535606820461384,
      "loss": 1.4152,
      "theoretical_loss": 3.486479977016544,
      "tokens_seen": 1818230784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045354062186559683,
      "loss": 1.4437,
      "theoretical_loss": 3.486469389309443,
      "tokens_seen": 1818296320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535205616850552,
      "loss": 1.5072,
      "theoretical_loss": 3.486458802090789,
      "tokens_seen": 1818361856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045350050150451355,
      "loss": 1.2958,
      "theoretical_loss": 3.4864482153605416,
      "tokens_seen": 1818427392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534804413239719,
      "loss": 1.3456,
      "theoretical_loss": 3.4864376291186607,
      "tokens_seen": 1818492928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534603811434303,
      "loss": 1.3086,
      "theoretical_loss": 3.4864270433651052,
      "tokens_seen": 1818558464
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534403209628887,
      "loss": 1.552,
      "theoretical_loss": 3.486416458099837,
      "tokens_seen": 1818624000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045342026078234706,
      "loss": 1.2934,
      "theoretical_loss": 3.486405873322814,
      "tokens_seen": 1818689536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534002006018054,
      "loss": 1.3957,
      "theoretical_loss": 3.4863952890339975,
      "tokens_seen": 1818755072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533801404212638,
      "loss": 1.3605,
      "theoretical_loss": 3.486384705233346,
      "tokens_seen": 1818820608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533600802407222,
      "loss": 1.388,
      "theoretical_loss": 3.4863741219208206,
      "tokens_seen": 1818886144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045334002006018056,
      "loss": 1.4278,
      "theoretical_loss": 3.486363539096381,
      "tokens_seen": 1818951680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533199598796389,
      "loss": 1.4162,
      "theoretical_loss": 3.486352956759987,
      "tokens_seen": 1819017216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532998996990973,
      "loss": 1.4359,
      "theoretical_loss": 3.4863423749115983,
      "tokens_seen": 1819082752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045327983951855565,
      "loss": 1.374,
      "theoretical_loss": 3.486331793551175,
      "tokens_seen": 1819148288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045325977933801406,
      "loss": 1.3729,
      "theoretical_loss": 3.4863212126786762,
      "tokens_seen": 1819213824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532397191574724,
      "loss": 1.3984,
      "theoretical_loss": 3.4863106322940633,
      "tokens_seen": 1819279360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532196589769308,
      "loss": 1.2672,
      "theoretical_loss": 3.4863000523972953,
      "tokens_seen": 1819344896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045319959879638915,
      "loss": 1.4197,
      "theoretical_loss": 3.4862894729883322,
      "tokens_seen": 1819410432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045317953861584757,
      "loss": 1.3061,
      "theoretical_loss": 3.4862788940671345,
      "tokens_seen": 1819475968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045315947843530593,
      "loss": 1.5047,
      "theoretical_loss": 3.4862683156336614,
      "tokens_seen": 1819541504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531394182547643,
      "loss": 1.4117,
      "theoretical_loss": 3.486257737687873,
      "tokens_seen": 1819607040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045311935807422265,
      "loss": 1.3772,
      "theoretical_loss": 3.4862471602297296,
      "tokens_seen": 1819672576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000453099297893681,
      "loss": 1.3383,
      "theoretical_loss": 3.486236583259191,
      "tokens_seen": 1819738112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045307923771313943,
      "loss": 1.2922,
      "theoretical_loss": 3.4862260067762176,
      "tokens_seen": 1819803648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530591775325978,
      "loss": 1.489,
      "theoretical_loss": 3.486215430780769,
      "tokens_seen": 1819869184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045303911735205616,
      "loss": 1.2933,
      "theoretical_loss": 3.486204855272804,
      "tokens_seen": 1819934720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530190571715145,
      "loss": 1.3882,
      "theoretical_loss": 3.4861942802522843,
      "tokens_seen": 1820000256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045299899699097294,
      "loss": 1.3231,
      "theoretical_loss": 3.486183705719169,
      "tokens_seen": 1820065792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529789368104313,
      "loss": 1.378,
      "theoretical_loss": 3.486173131673419,
      "tokens_seen": 1820131328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045295887662988966,
      "loss": 1.3546,
      "theoretical_loss": 3.4861625581149926,
      "tokens_seen": 1820196864
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452938816449348,
      "loss": 1.4891,
      "theoretical_loss": 3.4861519850438514,
      "tokens_seen": 1820262400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529187562688064,
      "loss": 1.3358,
      "theoretical_loss": 3.486141412459955,
      "tokens_seen": 1820327936
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 664244,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.473057746887207,
      "objective/train/theoretical_loss": 3.486130840363263,
      "objective/train/tokens_used": 190852576,
      "theoretical_loss": 3.486130840363263,
      "tokens_seen": 1820393472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528986960882648,
      "loss": 1.4186,
      "theoretical_loss": 3.486130840363263,
      "tokens_seen": 1820393472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045287863590772316,
      "loss": 1.5056,
      "theoretical_loss": 3.4861202687537354,
      "tokens_seen": 1820459008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528585757271815,
      "loss": 1.3785,
      "theoretical_loss": 3.486109697631333,
      "tokens_seen": 1820524544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528385155466399,
      "loss": 1.4482,
      "theoretical_loss": 3.486099126996015,
      "tokens_seen": 1820590080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528184553660983,
      "loss": 1.3856,
      "theoretical_loss": 3.4860885568477413,
      "tokens_seen": 1820655616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045279839518555667,
      "loss": 1.2543,
      "theoretical_loss": 3.486077987186473,
      "tokens_seen": 1820721152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045277833500501503,
      "loss": 1.2726,
      "theoretical_loss": 3.486067418012169,
      "tokens_seen": 1820786688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004527582748244734,
      "loss": 1.4595,
      "theoretical_loss": 3.4860568493247897,
      "tokens_seen": 1820852224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045273821464393175,
      "loss": 1.4927,
      "theoretical_loss": 3.4860462811242954,
      "tokens_seen": 1820917760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045271815446339017,
      "loss": 1.3728,
      "theoretical_loss": 3.4860357134106463,
      "tokens_seen": 1820983296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526980942828486,
      "loss": 1.4341,
      "theoretical_loss": 3.4860251461838017,
      "tokens_seen": 1821048832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045267803410230695,
      "loss": 1.3573,
      "theoretical_loss": 3.4860145794437223,
      "tokens_seen": 1821114368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526579739217653,
      "loss": 1.3299,
      "theoretical_loss": 3.486004013190368,
      "tokens_seen": 1821179904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526379137412237,
      "loss": 1.4155,
      "theoretical_loss": 3.4859934474236987,
      "tokens_seen": 1821245440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526178535606821,
      "loss": 1.4428,
      "theoretical_loss": 3.4859828821436745,
      "tokens_seen": 1821310976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045259779338014045,
      "loss": 1.4274,
      "theoretical_loss": 3.4859723173502557,
      "tokens_seen": 1821376512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525777331995988,
      "loss": 1.3713,
      "theoretical_loss": 3.4859617530434024,
      "tokens_seen": 1821442048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045255767301905723,
      "loss": 1.2971,
      "theoretical_loss": 3.485951189223074,
      "tokens_seen": 1821507584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525376128385156,
      "loss": 1.4422,
      "theoretical_loss": 3.4859406258892314,
      "tokens_seen": 1821573120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045251755265797395,
      "loss": 1.3715,
      "theoretical_loss": 3.4859300630418346,
      "tokens_seen": 1821638656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524974924774323,
      "loss": 1.3374,
      "theoretical_loss": 3.4859195006808434,
      "tokens_seen": 1821704192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524774322968907,
      "loss": 1.415,
      "theoretical_loss": 3.485908938806218,
      "tokens_seen": 1821769728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524573721163491,
      "loss": 1.2894,
      "theoretical_loss": 3.485898377417919,
      "tokens_seen": 1821835264
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045243731193580746,
      "loss": 1.3543,
      "theoretical_loss": 3.4858878165159055,
      "tokens_seen": 1821900800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524172517552658,
      "loss": 1.4854,
      "theoretical_loss": 3.485877256100138,
      "tokens_seen": 1821966336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523971915747242,
      "loss": 1.3344,
      "theoretical_loss": 3.4858666961705773,
      "tokens_seen": 1822031872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523771313941826,
      "loss": 1.3174,
      "theoretical_loss": 3.4858561367271825,
      "tokens_seen": 1822097408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045235707121364096,
      "loss": 1.3643,
      "theoretical_loss": 3.485845577769915,
      "tokens_seen": 1822162944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523370110330993,
      "loss": 1.5148,
      "theoretical_loss": 3.4858350192987335,
      "tokens_seen": 1822228480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523169508525577,
      "loss": 1.3437,
      "theoretical_loss": 3.485824461313599,
      "tokens_seen": 1822294016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045229689067201605,
      "loss": 1.4285,
      "theoretical_loss": 3.4858139038144715,
      "tokens_seen": 1822359552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045227683049147446,
      "loss": 1.4553,
      "theoretical_loss": 3.485803346801311,
      "tokens_seen": 1822425088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004522567703109328,
      "loss": 1.4443,
      "theoretical_loss": 3.485792790274078,
      "tokens_seen": 1822490624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004522367101303912,
      "loss": 1.4322,
      "theoretical_loss": 3.4857822342327323,
      "tokens_seen": 1822556160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045221664994984955,
      "loss": 1.4931,
      "theoretical_loss": 3.4857716786772346,
      "tokens_seen": 1822621696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045219658976930797,
      "loss": 1.443,
      "theoretical_loss": 3.4857611236075443,
      "tokens_seen": 1822687232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045217652958876633,
      "loss": 1.454,
      "theoretical_loss": 3.4857505690236223,
      "tokens_seen": 1822752768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521564694082247,
      "loss": 1.2928,
      "theoretical_loss": 3.485740014925428,
      "tokens_seen": 1822818304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045213640922768305,
      "loss": 1.3843,
      "theoretical_loss": 3.485729461312922,
      "tokens_seen": 1822883840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521163490471414,
      "loss": 1.3998,
      "theoretical_loss": 3.4857189081860644,
      "tokens_seen": 1822949376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045209628886659983,
      "loss": 1.396,
      "theoretical_loss": 3.485708355544816,
      "tokens_seen": 1823014912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520762286860582,
      "loss": 1.4106,
      "theoretical_loss": 3.4856978033891366,
      "tokens_seen": 1823080448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045205616850551656,
      "loss": 1.2949,
      "theoretical_loss": 3.4856872517189856,
      "tokens_seen": 1823145984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520361083249749,
      "loss": 1.3768,
      "theoretical_loss": 3.4856767005343245,
      "tokens_seen": 1823211520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045201604814443334,
      "loss": 1.3806,
      "theoretical_loss": 3.4856661498351125,
      "tokens_seen": 1823277056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519959879638917,
      "loss": 1.2996,
      "theoretical_loss": 3.4856555996213103,
      "tokens_seen": 1823342592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045197592778335006,
      "loss": 1.3818,
      "theoretical_loss": 3.485645049892878,
      "tokens_seen": 1823408128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519558676028084,
      "loss": 1.2801,
      "theoretical_loss": 3.485634500649777,
      "tokens_seen": 1823473664
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519358074222668,
      "loss": 1.4582,
      "theoretical_loss": 3.4856239518919647,
      "tokens_seen": 1823539200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519157472417252,
      "loss": 1.3983,
      "theoretical_loss": 3.4856134036194044,
      "tokens_seen": 1823604736
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 664954,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.428945541381836,
      "objective/train/theoretical_loss": 3.485602855832054,
      "objective/train/tokens_used": 194129376,
      "theoretical_loss": 3.485602855832054,
      "tokens_seen": 1823670272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045189568706118356,
      "loss": 1.4432,
      "theoretical_loss": 3.485602855832054,
      "tokens_seen": 1823670272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518756268806419,
      "loss": 1.3504,
      "theoretical_loss": 3.4855923085298754,
      "tokens_seen": 1823735808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518555667001003,
      "loss": 1.4151,
      "theoretical_loss": 3.4855817617128277,
      "tokens_seen": 1823801344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518355065195587,
      "loss": 1.4631,
      "theoretical_loss": 3.4855712153808724,
      "tokens_seen": 1823866880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045181544633901707,
      "loss": 1.364,
      "theoretical_loss": 3.4855606695339683,
      "tokens_seen": 1823932416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045179538615847543,
      "loss": 1.341,
      "theoretical_loss": 3.485550124172077,
      "tokens_seen": 1823997952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004517753259779338,
      "loss": 1.4471,
      "theoretical_loss": 3.4855395792951573,
      "tokens_seen": 1824063488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045175526579739215,
      "loss": 1.3904,
      "theoretical_loss": 3.4855290349031707,
      "tokens_seen": 1824129024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045173520561685057,
      "loss": 1.4009,
      "theoretical_loss": 3.4855184909960775,
      "tokens_seen": 1824194560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045171514543630893,
      "loss": 1.3067,
      "theoretical_loss": 3.4855079475738373,
      "tokens_seen": 1824260096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004516950852557673,
      "loss": 1.3469,
      "theoretical_loss": 3.4854974046364102,
      "tokens_seen": 1824325632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045167502507522566,
      "loss": 1.3843,
      "theoretical_loss": 3.485486862183757,
      "tokens_seen": 1824391168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045165496489468407,
      "loss": 1.41,
      "theoretical_loss": 3.4854763202158385,
      "tokens_seen": 1824456704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045163490471414244,
      "loss": 1.3258,
      "theoretical_loss": 3.485465778732614,
      "tokens_seen": 1824522240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004516148445336008,
      "loss": 1.3487,
      "theoretical_loss": 3.485455237734045,
      "tokens_seen": 1824587776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045159478435305916,
      "loss": 1.3223,
      "theoretical_loss": 3.4854446972200903,
      "tokens_seen": 1824653312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515747241725175,
      "loss": 1.3633,
      "theoretical_loss": 3.485434157190711,
      "tokens_seen": 1824718848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045155466399197594,
      "loss": 1.2066,
      "theoretical_loss": 3.4854236176458677,
      "tokens_seen": 1824784384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515346038114343,
      "loss": 1.3287,
      "theoretical_loss": 3.4854130785855206,
      "tokens_seen": 1824849920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045151454363089266,
      "loss": 1.4036,
      "theoretical_loss": 3.4854025400096296,
      "tokens_seen": 1824915456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451494483450351,
      "loss": 1.3777,
      "theoretical_loss": 3.485392001918155,
      "tokens_seen": 1824980992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045147442326980944,
      "loss": 1.3453,
      "theoretical_loss": 3.485381464311058,
      "tokens_seen": 1825046528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004514543630892678,
      "loss": 1.289,
      "theoretical_loss": 3.485370927188298,
      "tokens_seen": 1825112064
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045143430290872617,
      "loss": 1.4624,
      "theoretical_loss": 3.4853603905498356,
      "tokens_seen": 1825177600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045141424272818453,
      "loss": 1.3451,
      "theoretical_loss": 3.485349854395632,
      "tokens_seen": 1825243136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045139418254764294,
      "loss": 1.4064,
      "theoretical_loss": 3.4853393187256456,
      "tokens_seen": 1825308672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513741223671013,
      "loss": 1.2947,
      "theoretical_loss": 3.4853287835398388,
      "tokens_seen": 1825374208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045135406218655967,
      "loss": 1.396,
      "theoretical_loss": 3.485318248838171,
      "tokens_seen": 1825439744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045133400200601803,
      "loss": 1.3771,
      "theoretical_loss": 3.485307714620603,
      "tokens_seen": 1825505280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513139418254764,
      "loss": 1.4019,
      "theoretical_loss": 3.4852971808870947,
      "tokens_seen": 1825570816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512938816449348,
      "loss": 1.3241,
      "theoretical_loss": 3.4852866476376065,
      "tokens_seen": 1825636352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045127382146439317,
      "loss": 1.4185,
      "theoretical_loss": 3.485276114872099,
      "tokens_seen": 1825701888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045125376128385153,
      "loss": 1.3577,
      "theoretical_loss": 3.4852655825905328,
      "tokens_seen": 1825767424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512337011033099,
      "loss": 1.2659,
      "theoretical_loss": 3.485255050792868,
      "tokens_seen": 1825832960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512136409227683,
      "loss": 1.3476,
      "theoretical_loss": 3.485244519479065,
      "tokens_seen": 1825898496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045119358074222673,
      "loss": 1.3852,
      "theoretical_loss": 3.4852339886490844,
      "tokens_seen": 1825964032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511735205616851,
      "loss": 1.3941,
      "theoretical_loss": 3.4852234583028863,
      "tokens_seen": 1826029568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045115346038114345,
      "loss": 1.346,
      "theoretical_loss": 3.485212928440431,
      "tokens_seen": 1826095104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511334002006018,
      "loss": 1.5541,
      "theoretical_loss": 3.48520239906168,
      "tokens_seen": 1826160640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045111334002006023,
      "loss": 1.4031,
      "theoretical_loss": 3.4851918701665925,
      "tokens_seen": 1826226176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510932798395186,
      "loss": 1.3364,
      "theoretical_loss": 3.485181341755129,
      "tokens_seen": 1826291712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045107321965897696,
      "loss": 1.3323,
      "theoretical_loss": 3.4851708138272506,
      "tokens_seen": 1826357248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510531594784353,
      "loss": 1.3312,
      "theoretical_loss": 3.4851602863829174,
      "tokens_seen": 1826422784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045103309929789374,
      "loss": 1.2774,
      "theoretical_loss": 3.48514975942209,
      "tokens_seen": 1826488320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510130391173521,
      "loss": 1.3689,
      "theoretical_loss": 3.4851392329447286,
      "tokens_seen": 1826553856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045099297893681046,
      "loss": 1.4138,
      "theoretical_loss": 3.4851287069507935,
      "tokens_seen": 1826619392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509729187562688,
      "loss": 1.3661,
      "theoretical_loss": 3.485118181440246,
      "tokens_seen": 1826684928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509528585757272,
      "loss": 1.3387,
      "theoretical_loss": 3.4851076564130454,
      "tokens_seen": 1826750464
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509327983951856,
      "loss": 1.3327,
      "theoretical_loss": 3.485097131869153,
      "tokens_seen": 1826816000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045091273821464396,
      "loss": 1.2156,
      "theoretical_loss": 3.4850866078085287,
      "tokens_seen": 1826881536
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 666029,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2888948917388916,
      "objective/train/theoretical_loss": 3.485076084231134,
      "objective/train/tokens_used": 197406176,
      "theoretical_loss": 3.485076084231134,
      "tokens_seen": 1826947072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508926780341023,
      "loss": 1.3517,
      "theoretical_loss": 3.485076084231134,
      "tokens_seen": 1826947072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508726178535607,
      "loss": 1.4546,
      "theoretical_loss": 3.485065561136928,
      "tokens_seen": 1827012608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508525576730191,
      "loss": 1.2362,
      "theoretical_loss": 3.4850550385258723,
      "tokens_seen": 1827078144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045083249749247747,
      "loss": 1.4626,
      "theoretical_loss": 3.4850445163979265,
      "tokens_seen": 1827143680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045081243731193583,
      "loss": 1.4563,
      "theoretical_loss": 3.485033994753052,
      "tokens_seen": 1827209216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507923771313942,
      "loss": 1.3422,
      "theoretical_loss": 3.4850234735912085,
      "tokens_seen": 1827274752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045077231695085255,
      "loss": 1.3542,
      "theoretical_loss": 3.4850129529123564,
      "tokens_seen": 1827340288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045075225677031097,
      "loss": 1.4255,
      "theoretical_loss": 3.4850024327164575,
      "tokens_seen": 1827405824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045073219658976933,
      "loss": 1.3217,
      "theoretical_loss": 3.484991913003471,
      "tokens_seen": 1827471360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507121364092277,
      "loss": 1.4181,
      "theoretical_loss": 3.484981393773358,
      "tokens_seen": 1827536896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045069207622868606,
      "loss": 1.3229,
      "theoretical_loss": 3.484970875026079,
      "tokens_seen": 1827602432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506720160481445,
      "loss": 1.4248,
      "theoretical_loss": 3.4849603567615937,
      "tokens_seen": 1827667968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045065195586760284,
      "loss": 1.368,
      "theoretical_loss": 3.4849498389798637,
      "tokens_seen": 1827733504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506318956870612,
      "loss": 1.4122,
      "theoretical_loss": 3.4849393216808497,
      "tokens_seen": 1827799040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045061183550651956,
      "loss": 1.307,
      "theoretical_loss": 3.484928804864511,
      "tokens_seen": 1827864576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505917753259779,
      "loss": 1.2867,
      "theoretical_loss": 3.4849182885308094,
      "tokens_seen": 1827930112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045057171514543634,
      "loss": 1.4847,
      "theoretical_loss": 3.484907772679705,
      "tokens_seen": 1827995648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505516549648947,
      "loss": 1.2788,
      "theoretical_loss": 3.4848972573111574,
      "tokens_seen": 1828061184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045053159478435306,
      "loss": 1.3949,
      "theoretical_loss": 3.484886742425129,
      "tokens_seen": 1828126720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505115346038114,
      "loss": 1.3631,
      "theoretical_loss": 3.4848762280215784,
      "tokens_seen": 1828192256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045049147442326984,
      "loss": 1.4437,
      "theoretical_loss": 3.4848657141004677,
      "tokens_seen": 1828257792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504714142427282,
      "loss": 1.4086,
      "theoretical_loss": 3.484855200661757,
      "tokens_seen": 1828323328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045045135406218657,
      "loss": 1.3689,
      "theoretical_loss": 3.4848446877054062,
      "tokens_seen": 1828388864
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045043129388164493,
      "loss": 1.3715,
      "theoretical_loss": 3.484834175231377,
      "tokens_seen": 1828454400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045041123370110334,
      "loss": 1.3624,
      "theoretical_loss": 3.4848236632396294,
      "tokens_seen": 1828519936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503911735205617,
      "loss": 1.4685,
      "theoretical_loss": 3.484813151730124,
      "tokens_seen": 1828585472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045037111334002007,
      "loss": 1.3944,
      "theoretical_loss": 3.484802640702821,
      "tokens_seen": 1828651008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045035105315947843,
      "loss": 1.3633,
      "theoretical_loss": 3.484792130157682,
      "tokens_seen": 1828716544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503309929789368,
      "loss": 1.3572,
      "theoretical_loss": 3.484781620094667,
      "tokens_seen": 1828782080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503109327983952,
      "loss": 1.294,
      "theoretical_loss": 3.4847711105137367,
      "tokens_seen": 1828847616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045029087261785357,
      "loss": 1.3522,
      "theoretical_loss": 3.4847606014148518,
      "tokens_seen": 1828913152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045027081243731193,
      "loss": 1.3408,
      "theoretical_loss": 3.484750092797972,
      "tokens_seen": 1828978688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502507522567703,
      "loss": 1.2818,
      "theoretical_loss": 3.4847395846630596,
      "tokens_seen": 1829044224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502306920762287,
      "loss": 1.2774,
      "theoretical_loss": 3.4847290770100736,
      "tokens_seen": 1829109760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502106318956871,
      "loss": 1.2751,
      "theoretical_loss": 3.484718569838976,
      "tokens_seen": 1829175296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045019057171514544,
      "loss": 1.2768,
      "theoretical_loss": 3.484708063149726,
      "tokens_seen": 1829240832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501705115346038,
      "loss": 1.3734,
      "theoretical_loss": 3.484697556942286,
      "tokens_seen": 1829306368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045015045135406216,
      "loss": 1.3498,
      "theoretical_loss": 3.484687051216615,
      "tokens_seen": 1829371904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501303911735206,
      "loss": 1.332,
      "theoretical_loss": 3.4846765459726745,
      "tokens_seen": 1829437440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045011033099297894,
      "loss": 1.3605,
      "theoretical_loss": 3.484666041210425,
      "tokens_seen": 1829502976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500902708124373,
      "loss": 1.34,
      "theoretical_loss": 3.484655536929827,
      "tokens_seen": 1829568512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045007021063189567,
      "loss": 1.2935,
      "theoretical_loss": 3.4846450331308416,
      "tokens_seen": 1829634048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500501504513541,
      "loss": 1.398,
      "theoretical_loss": 3.484634529813429,
      "tokens_seen": 1829699584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045003009027081244,
      "loss": 1.3657,
      "theoretical_loss": 3.48462402697755,
      "tokens_seen": 1829765120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500100300902708,
      "loss": 1.3742,
      "theoretical_loss": 3.4846135246231658,
      "tokens_seen": 1829830656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044998996990972917,
      "loss": 1.361,
      "theoretical_loss": 3.4846030227502363,
      "tokens_seen": 1829896192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044996990972918753,
      "loss": 1.4564,
      "theoretical_loss": 3.484592521358722,
      "tokens_seen": 1829961728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044994984954864595,
      "loss": 1.3517,
      "theoretical_loss": 3.4845820204485847,
      "tokens_seen": 1830027264
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004499297893681043,
      "loss": 1.3877,
      "theoretical_loss": 3.4845715200197844,
      "tokens_seen": 1830092800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044990972918756267,
      "loss": 1.3119,
      "theoretical_loss": 3.484561020072282,
      "tokens_seen": 1830158336
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 667256,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3890573978424072,
      "objective/train/theoretical_loss": 3.484550520606038,
      "objective/train/tokens_used": 200682976,
      "theoretical_loss": 3.484550520606038,
      "tokens_seen": 1830223872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044988966900702103,
      "loss": 1.3316,
      "theoretical_loss": 3.484550520606038,
      "tokens_seen": 1830223872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044986960882647945,
      "loss": 1.4165,
      "theoretical_loss": 3.484540021621013,
      "tokens_seen": 1830289408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498495486459378,
      "loss": 1.3301,
      "theoretical_loss": 3.484529523117168,
      "tokens_seen": 1830354944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498294884653962,
      "loss": 1.3621,
      "theoretical_loss": 3.4845190250944635,
      "tokens_seen": 1830420480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044980942828485454,
      "loss": 1.23,
      "theoretical_loss": 3.484508527552861,
      "tokens_seen": 1830486016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497893681043129,
      "loss": 1.2959,
      "theoretical_loss": 3.48449803049232,
      "tokens_seen": 1830551552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497693079237713,
      "loss": 1.2651,
      "theoretical_loss": 3.4844875339128016,
      "tokens_seen": 1830617088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497492477432297,
      "loss": 1.3512,
      "theoretical_loss": 3.4844770378142673,
      "tokens_seen": 1830682624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044972918756268804,
      "loss": 1.4223,
      "theoretical_loss": 3.484466542196677,
      "tokens_seen": 1830748160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497091273821464,
      "loss": 1.346,
      "theoretical_loss": 3.484456047059992,
      "tokens_seen": 1830813696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496890672016049,
      "loss": 1.2978,
      "theoretical_loss": 3.484445552404173,
      "tokens_seen": 1830879232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044966900702106324,
      "loss": 1.3466,
      "theoretical_loss": 3.48443505822918,
      "tokens_seen": 1830944768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496489468405216,
      "loss": 1.3783,
      "theoretical_loss": 3.484424564534975,
      "tokens_seen": 1831010304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044962888665997996,
      "loss": 1.4081,
      "theoretical_loss": 3.4844140713215177,
      "tokens_seen": 1831075840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496088264794383,
      "loss": 1.4309,
      "theoretical_loss": 3.4844035785887693,
      "tokens_seen": 1831141376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044958876629889674,
      "loss": 1.3216,
      "theoretical_loss": 3.4843930863366905,
      "tokens_seen": 1831206912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495687061183551,
      "loss": 1.3568,
      "theoretical_loss": 3.4843825945652425,
      "tokens_seen": 1831272448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044954864593781346,
      "loss": 1.3302,
      "theoretical_loss": 3.484372103274385,
      "tokens_seen": 1831337984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495285857572718,
      "loss": 1.4184,
      "theoretical_loss": 3.4843616124640797,
      "tokens_seen": 1831403520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044950852557673024,
      "loss": 1.2208,
      "theoretical_loss": 3.4843511221342873,
      "tokens_seen": 1831469056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494884653961886,
      "loss": 1.3212,
      "theoretical_loss": 3.484340632284969,
      "tokens_seen": 1831534592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044946840521564697,
      "loss": 1.3291,
      "theoretical_loss": 3.4843301429160842,
      "tokens_seen": 1831600128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044944834503510533,
      "loss": 1.3375,
      "theoretical_loss": 3.484319654027595,
      "tokens_seen": 1831665664
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044942828485456375,
      "loss": 1.3828,
      "theoretical_loss": 3.484309165619462,
      "tokens_seen": 1831731200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494082246740221,
      "loss": 1.3023,
      "theoretical_loss": 3.484298677691645,
      "tokens_seen": 1831796736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044938816449348047,
      "loss": 1.3586,
      "theoretical_loss": 3.484288190244106,
      "tokens_seen": 1831862272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044936810431293883,
      "loss": 1.3609,
      "theoretical_loss": 3.484277703276806,
      "tokens_seen": 1831927808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493480441323972,
      "loss": 1.3817,
      "theoretical_loss": 3.484267216789705,
      "tokens_seen": 1831993344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493279839518556,
      "loss": 1.388,
      "theoretical_loss": 3.4842567307827643,
      "tokens_seen": 1832058880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044930792377131397,
      "loss": 1.2866,
      "theoretical_loss": 3.484246245255944,
      "tokens_seen": 1832124416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044928786359077234,
      "loss": 1.4305,
      "theoretical_loss": 3.484235760209206,
      "tokens_seen": 1832189952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492678034102307,
      "loss": 1.336,
      "theoretical_loss": 3.4842252756425105,
      "tokens_seen": 1832255488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492477432296891,
      "loss": 1.5091,
      "theoretical_loss": 3.4842147915558184,
      "tokens_seen": 1832321024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492276830491475,
      "loss": 1.289,
      "theoretical_loss": 3.4842043079490908,
      "tokens_seen": 1832386560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044920762286860584,
      "loss": 1.4503,
      "theoretical_loss": 3.4841938248222886,
      "tokens_seen": 1832452096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491875626880642,
      "loss": 1.3954,
      "theoretical_loss": 3.4841833421753723,
      "tokens_seen": 1832517632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044916750250752256,
      "loss": 1.4462,
      "theoretical_loss": 3.484172860008303,
      "tokens_seen": 1832583168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449147442326981,
      "loss": 1.4106,
      "theoretical_loss": 3.4841623783210416,
      "tokens_seen": 1832648704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044912738214643934,
      "loss": 1.3435,
      "theoretical_loss": 3.4841518971135486,
      "tokens_seen": 1832714240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491073219658977,
      "loss": 1.1785,
      "theoretical_loss": 3.4841414163857856,
      "tokens_seen": 1832779776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044908726178535607,
      "loss": 1.4378,
      "theoretical_loss": 3.484130936137713,
      "tokens_seen": 1832845312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490672016048145,
      "loss": 1.4084,
      "theoretical_loss": 3.4841204563692916,
      "tokens_seen": 1832910848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044904714142427284,
      "loss": 1.3024,
      "theoretical_loss": 3.484109977080483,
      "tokens_seen": 1832976384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490270812437312,
      "loss": 1.2725,
      "theoretical_loss": 3.484099498271247,
      "tokens_seen": 1833041920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044900702106318957,
      "loss": 1.3649,
      "theoretical_loss": 3.4840890199415453,
      "tokens_seen": 1833107456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044898696088264793,
      "loss": 1.4148,
      "theoretical_loss": 3.4840785420913387,
      "tokens_seen": 1833172992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044896690070210635,
      "loss": 1.3436,
      "theoretical_loss": 3.484068064720588,
      "tokens_seen": 1833238528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004489468405215647,
      "loss": 1.4146,
      "theoretical_loss": 3.4840575878292546,
      "tokens_seen": 1833304064
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044892678034102307,
      "loss": 1.4709,
      "theoretical_loss": 3.484047111417298,
      "tokens_seen": 1833369600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044890672016048143,
      "loss": 1.4155,
      "theoretical_loss": 3.484036635484681,
      "tokens_seen": 1833435136
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 668449,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3851547241210938,
      "objective/train/theoretical_loss": 3.4840261600313633,
      "objective/train/tokens_used": 203959776,
      "theoretical_loss": 3.4840261600313633,
      "tokens_seen": 1833500672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044888665997993985,
      "loss": 1.3764,
      "theoretical_loss": 3.4840261600313633,
      "tokens_seen": 1833500672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488665997993982,
      "loss": 1.3619,
      "theoretical_loss": 3.4840156850573063,
      "tokens_seen": 1833566208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488465396188566,
      "loss": 1.4809,
      "theoretical_loss": 3.4840052105624704,
      "tokens_seen": 1833631744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044882647943831494,
      "loss": 1.3097,
      "theoretical_loss": 3.4839947365468174,
      "tokens_seen": 1833697280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488064192577733,
      "loss": 1.4207,
      "theoretical_loss": 3.483984263010308,
      "tokens_seen": 1833762816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487863590772317,
      "loss": 1.3273,
      "theoretical_loss": 3.483973789952903,
      "tokens_seen": 1833828352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487662988966901,
      "loss": 1.3316,
      "theoretical_loss": 3.4839633173745628,
      "tokens_seen": 1833893888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044874623871614844,
      "loss": 1.3361,
      "theoretical_loss": 3.4839528452752493,
      "tokens_seen": 1833959424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487261785356068,
      "loss": 1.3571,
      "theoretical_loss": 3.483942373654923,
      "tokens_seen": 1834024960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487061183550652,
      "loss": 1.3918,
      "theoretical_loss": 3.483931902513545,
      "tokens_seen": 1834090496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486860581745236,
      "loss": 1.3251,
      "theoretical_loss": 3.4839214318510763,
      "tokens_seen": 1834156032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044866599799398194,
      "loss": 1.4931,
      "theoretical_loss": 3.483910961667478,
      "tokens_seen": 1834221568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486459378134403,
      "loss": 1.3209,
      "theoretical_loss": 3.483900491962711,
      "tokens_seen": 1834287104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044862587763289867,
      "loss": 1.3772,
      "theoretical_loss": 3.4838900227367358,
      "tokens_seen": 1834352640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486058174523571,
      "loss": 1.3448,
      "theoretical_loss": 3.4838795539895138,
      "tokens_seen": 1834418176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044858575727181545,
      "loss": 1.4248,
      "theoretical_loss": 3.483869085721006,
      "tokens_seen": 1834483712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485656970912738,
      "loss": 1.451,
      "theoretical_loss": 3.483858617931174,
      "tokens_seen": 1834549248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044854563691073217,
      "loss": 1.4277,
      "theoretical_loss": 3.483848150619978,
      "tokens_seen": 1834614784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485255767301906,
      "loss": 1.3668,
      "theoretical_loss": 3.483837683787379,
      "tokens_seen": 1834680320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044850551654964895,
      "loss": 1.2602,
      "theoretical_loss": 3.4838272174333387,
      "tokens_seen": 1834745856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484854563691073,
      "loss": 1.2785,
      "theoretical_loss": 3.4838167515578173,
      "tokens_seen": 1834811392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484653961885657,
      "loss": 1.2475,
      "theoretical_loss": 3.4838062861607764,
      "tokens_seen": 1834876928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044844533600802404,
      "loss": 1.2824,
      "theoretical_loss": 3.483795821242177,
      "tokens_seen": 1834942464
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044842527582748245,
      "loss": 1.2873,
      "theoretical_loss": 3.48378535680198,
      "tokens_seen": 1835008000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484052156469408,
      "loss": 1.3887,
      "theoretical_loss": 3.4837748928401466,
      "tokens_seen": 1835073536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483851554663992,
      "loss": 1.3698,
      "theoretical_loss": 3.4837644293566377,
      "tokens_seen": 1835139072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044836509528585754,
      "loss": 1.3546,
      "theoretical_loss": 3.483753966351414,
      "tokens_seen": 1835204608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044834503510531596,
      "loss": 1.3905,
      "theoretical_loss": 3.4837435038244378,
      "tokens_seen": 1835270144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483249749247743,
      "loss": 1.2918,
      "theoretical_loss": 3.4837330417756682,
      "tokens_seen": 1835335680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483049147442327,
      "loss": 1.2334,
      "theoretical_loss": 3.483722580205068,
      "tokens_seen": 1835401216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044828485456369104,
      "loss": 1.3341,
      "theoretical_loss": 3.4837121191125977,
      "tokens_seen": 1835466752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482647943831494,
      "loss": 1.3696,
      "theoretical_loss": 3.483701658498218,
      "tokens_seen": 1835532288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482447342026078,
      "loss": 1.4185,
      "theoretical_loss": 3.483691198361891,
      "tokens_seen": 1835597824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482246740220662,
      "loss": 1.371,
      "theoretical_loss": 3.483680738703576,
      "tokens_seen": 1835663360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044820461384152455,
      "loss": 1.2848,
      "theoretical_loss": 3.4836702795232357,
      "tokens_seen": 1835728896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044818455366098296,
      "loss": 1.3714,
      "theoretical_loss": 3.4836598208208307,
      "tokens_seen": 1835794432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481644934804414,
      "loss": 1.3666,
      "theoretical_loss": 3.483649362596322,
      "tokens_seen": 1835859968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044814443329989974,
      "loss": 1.4868,
      "theoretical_loss": 3.4836389048496708,
      "tokens_seen": 1835925504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481243731193581,
      "loss": 1.3598,
      "theoretical_loss": 3.4836284475808386,
      "tokens_seen": 1835991040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044810431293881647,
      "loss": 1.4479,
      "theoretical_loss": 3.4836179907897855,
      "tokens_seen": 1836056576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004480842527582749,
      "loss": 1.3072,
      "theoretical_loss": 3.4836075344764734,
      "tokens_seen": 1836122112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044806419257773324,
      "loss": 1.3202,
      "theoretical_loss": 3.483597078640863,
      "tokens_seen": 1836187648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004480441323971916,
      "loss": 1.2778,
      "theoretical_loss": 3.4835866232829167,
      "tokens_seen": 1836253184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044802407221664997,
      "loss": 1.3599,
      "theoretical_loss": 3.4835761684025934,
      "tokens_seen": 1836318720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044800401203610833,
      "loss": 1.3083,
      "theoretical_loss": 3.483565713999856,
      "tokens_seen": 1836384256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044798395185556675,
      "loss": 1.4001,
      "theoretical_loss": 3.483555260074665,
      "tokens_seen": 1836449792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004479638916750251,
      "loss": 1.2928,
      "theoretical_loss": 3.4835448066269814,
      "tokens_seen": 1836515328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044794383149448347,
      "loss": 1.3265,
      "theoretical_loss": 3.483534353656767,
      "tokens_seen": 1836580864
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044792377131394183,
      "loss": 1.3513,
      "theoretical_loss": 3.4835239011639825,
      "tokens_seen": 1836646400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044790371113340025,
      "loss": 1.3269,
      "theoretical_loss": 3.483513449148589,
      "tokens_seen": 1836711936
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 669661,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3857011795043945,
      "objective/train/theoretical_loss": 3.4835029976105476,
      "objective/train/tokens_used": 207236576,
      "theoretical_loss": 3.4835029976105476,
      "tokens_seen": 1836777472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478836509528586,
      "loss": 1.4506,
      "theoretical_loss": 3.4835029976105476,
      "tokens_seen": 1836777472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447863590772317,
      "loss": 1.3854,
      "theoretical_loss": 3.4834925465498197,
      "tokens_seen": 1836843008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044784353059177534,
      "loss": 1.2658,
      "theoretical_loss": 3.483482095966366,
      "tokens_seen": 1836908544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478234704112337,
      "loss": 1.3701,
      "theoretical_loss": 3.4834716458601482,
      "tokens_seen": 1836974080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478034102306921,
      "loss": 1.405,
      "theoretical_loss": 3.483461196231128,
      "tokens_seen": 1837039616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477833500501505,
      "loss": 1.4214,
      "theoretical_loss": 3.4834507470792655,
      "tokens_seen": 1837105152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044776328986960884,
      "loss": 1.4264,
      "theoretical_loss": 3.4834402984045223,
      "tokens_seen": 1837170688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477432296890672,
      "loss": 1.4037,
      "theoretical_loss": 3.4834298502068597,
      "tokens_seen": 1837236224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477231695085256,
      "loss": 1.3278,
      "theoretical_loss": 3.4834194024862386,
      "tokens_seen": 1837301760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447703109327984,
      "loss": 1.2432,
      "theoretical_loss": 3.483408955242621,
      "tokens_seen": 1837367296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044768304914744234,
      "loss": 1.3156,
      "theoretical_loss": 3.483398508475967,
      "tokens_seen": 1837432832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476629889669007,
      "loss": 1.3817,
      "theoretical_loss": 3.4833880621862385,
      "tokens_seen": 1837498368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044764292878635907,
      "loss": 1.4581,
      "theoretical_loss": 3.4833776163733967,
      "tokens_seen": 1837563904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476228686058175,
      "loss": 1.4544,
      "theoretical_loss": 3.4833671710374023,
      "tokens_seen": 1837629440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044760280842527585,
      "loss": 1.4868,
      "theoretical_loss": 3.4833567261782177,
      "tokens_seen": 1837694976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475827482447342,
      "loss": 1.2919,
      "theoretical_loss": 3.4833462817958027,
      "tokens_seen": 1837760512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044756268806419257,
      "loss": 1.4497,
      "theoretical_loss": 3.4833358378901194,
      "tokens_seen": 1837826048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447542627883651,
      "loss": 1.395,
      "theoretical_loss": 3.4833253944611284,
      "tokens_seen": 1837891584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044752256770310935,
      "loss": 1.392,
      "theoretical_loss": 3.483314951508792,
      "tokens_seen": 1837957120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475025075225677,
      "loss": 1.4253,
      "theoretical_loss": 3.4833045090330703,
      "tokens_seen": 1838022656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474824473420261,
      "loss": 1.3788,
      "theoretical_loss": 3.483294067033926,
      "tokens_seen": 1838088192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044746238716148444,
      "loss": 1.369,
      "theoretical_loss": 3.4832836255113184,
      "tokens_seen": 1838153728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044744232698094285,
      "loss": 1.2772,
      "theoretical_loss": 3.48327318446521,
      "tokens_seen": 1838219264
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474222668004012,
      "loss": 1.33,
      "theoretical_loss": 3.483262743895562,
      "tokens_seen": 1838284800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474022066198596,
      "loss": 1.3684,
      "theoretical_loss": 3.4832523038023355,
      "tokens_seen": 1838350336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044738214643931794,
      "loss": 1.3565,
      "theoretical_loss": 3.483241864185492,
      "tokens_seen": 1838415872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044736208625877636,
      "loss": 1.2278,
      "theoretical_loss": 3.483231425044993,
      "tokens_seen": 1838481408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473420260782347,
      "loss": 1.3463,
      "theoretical_loss": 3.4832209863807986,
      "tokens_seen": 1838546944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473219658976931,
      "loss": 1.3481,
      "theoretical_loss": 3.4832105481928712,
      "tokens_seen": 1838612480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044730190571715144,
      "loss": 1.3891,
      "theoretical_loss": 3.483200110481172,
      "tokens_seen": 1838678016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472818455366098,
      "loss": 1.3344,
      "theoretical_loss": 3.4831896732456613,
      "tokens_seen": 1838743552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472617853560682,
      "loss": 1.3094,
      "theoretical_loss": 3.483179236486302,
      "tokens_seen": 1838809088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472417251755266,
      "loss": 1.3006,
      "theoretical_loss": 3.483168800203054,
      "tokens_seen": 1838874624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044722166499498495,
      "loss": 1.2804,
      "theoretical_loss": 3.4831583643958797,
      "tokens_seen": 1838940160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472016048144433,
      "loss": 1.4075,
      "theoretical_loss": 3.4831479290647396,
      "tokens_seen": 1839005696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471815446339017,
      "loss": 1.359,
      "theoretical_loss": 3.483137494209595,
      "tokens_seen": 1839071232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471614844533601,
      "loss": 1.4215,
      "theoretical_loss": 3.4831270598304087,
      "tokens_seen": 1839136768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044714142427281845,
      "loss": 1.4103,
      "theoretical_loss": 3.48311662592714,
      "tokens_seen": 1839202304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471213640922768,
      "loss": 1.3851,
      "theoretical_loss": 3.4831061924997515,
      "tokens_seen": 1839267840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044710130391173523,
      "loss": 1.3227,
      "theoretical_loss": 3.4830957595482035,
      "tokens_seen": 1839333376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470812437311936,
      "loss": 1.3078,
      "theoretical_loss": 3.483085327072459,
      "tokens_seen": 1839398912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044706118355065195,
      "loss": 1.426,
      "theoretical_loss": 3.4830748950724777,
      "tokens_seen": 1839464448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470411233701103,
      "loss": 1.3378,
      "theoretical_loss": 3.4830644635482217,
      "tokens_seen": 1839529984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470210631895687,
      "loss": 1.3558,
      "theoretical_loss": 3.483054032499653,
      "tokens_seen": 1839595520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470010030090271,
      "loss": 1.3384,
      "theoretical_loss": 3.483043601926731,
      "tokens_seen": 1839661056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044698094282848546,
      "loss": 1.3258,
      "theoretical_loss": 3.483033171829419,
      "tokens_seen": 1839726592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469608826479438,
      "loss": 1.3786,
      "theoretical_loss": 3.4830227422076776,
      "tokens_seen": 1839792128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469408224674022,
      "loss": 1.272,
      "theoretical_loss": 3.4830123130614687,
      "tokens_seen": 1839857664
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469207622868606,
      "loss": 1.4065,
      "theoretical_loss": 3.4830018843907524,
      "tokens_seen": 1839923200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044690070210631896,
      "loss": 1.4026,
      "theoretical_loss": 3.4829914561954913,
      "tokens_seen": 1839988736
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 670853,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3770475387573242,
      "objective/train/theoretical_loss": 3.4829810284756464,
      "objective/train/tokens_used": 210513376,
      "theoretical_loss": 3.4829810284756464,
      "tokens_seen": 1840054272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468806419257773,
      "loss": 1.2874,
      "theoretical_loss": 3.4829810284756464,
      "tokens_seen": 1840054272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468605817452357,
      "loss": 1.2706,
      "theoretical_loss": 3.4829706012311794,
      "tokens_seen": 1840119808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044684052156469405,
      "loss": 1.4379,
      "theoretical_loss": 3.4829601744620513,
      "tokens_seen": 1840185344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044682046138415246,
      "loss": 1.4284,
      "theoretical_loss": 3.4829497481682234,
      "tokens_seen": 1840250880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468004012036108,
      "loss": 1.3468,
      "theoretical_loss": 3.482939322349657,
      "tokens_seen": 1840316416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467803410230692,
      "loss": 1.3455,
      "theoretical_loss": 3.482928897006315,
      "tokens_seen": 1840381952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044676028084252755,
      "loss": 1.2919,
      "theoretical_loss": 3.4829184721381568,
      "tokens_seen": 1840447488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044674022066198597,
      "loss": 1.3297,
      "theoretical_loss": 3.482908047745145,
      "tokens_seen": 1840513024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044672016048144433,
      "loss": 1.3592,
      "theoretical_loss": 3.48289762382724,
      "tokens_seen": 1840578560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467001003009027,
      "loss": 1.3635,
      "theoretical_loss": 3.4828872003844045,
      "tokens_seen": 1840644096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466800401203611,
      "loss": 1.3362,
      "theoretical_loss": 3.482876777416599,
      "tokens_seen": 1840709632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044665997993981947,
      "loss": 1.2639,
      "theoretical_loss": 3.482866354923786,
      "tokens_seen": 1840775168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466399197592779,
      "loss": 1.3608,
      "theoretical_loss": 3.4828559329059257,
      "tokens_seen": 1840840704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044661985957873625,
      "loss": 1.3963,
      "theoretical_loss": 3.48284551136298,
      "tokens_seen": 1840906240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465997993981946,
      "loss": 1.4256,
      "theoretical_loss": 3.482835090294911,
      "tokens_seen": 1840971776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044657973921765297,
      "loss": 1.307,
      "theoretical_loss": 3.482824669701679,
      "tokens_seen": 1841037312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465596790371114,
      "loss": 1.337,
      "theoretical_loss": 3.4828142495832464,
      "tokens_seen": 1841102848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044653961885656975,
      "loss": 1.3583,
      "theoretical_loss": 3.4828038299395745,
      "tokens_seen": 1841168384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465195586760281,
      "loss": 1.3492,
      "theoretical_loss": 3.482793410770624,
      "tokens_seen": 1841233920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464994984954865,
      "loss": 1.4215,
      "theoretical_loss": 3.4827829920763573,
      "tokens_seen": 1841299456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044647943831494484,
      "loss": 1.321,
      "theoretical_loss": 3.482772573856736,
      "tokens_seen": 1841364992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044645937813440325,
      "loss": 1.3915,
      "theoretical_loss": 3.4827621561117206,
      "tokens_seen": 1841430528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464393179538616,
      "loss": 1.4487,
      "theoretical_loss": 3.4827517388412734,
      "tokens_seen": 1841496064
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044641925777332,
      "loss": 1.3752,
      "theoretical_loss": 3.4827413220453556,
      "tokens_seen": 1841561600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044639919759277834,
      "loss": 1.4058,
      "theoretical_loss": 3.482730905723929,
      "tokens_seen": 1841627136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044637913741223676,
      "loss": 1.2613,
      "theoretical_loss": 3.482720489876954,
      "tokens_seen": 1841692672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463590772316951,
      "loss": 1.3246,
      "theoretical_loss": 3.4827100745043937,
      "tokens_seen": 1841758208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463390170511535,
      "loss": 1.444,
      "theoretical_loss": 3.4826996596062085,
      "tokens_seen": 1841823744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044631895687061184,
      "loss": 1.2866,
      "theoretical_loss": 3.4826892451823603,
      "tokens_seen": 1841889280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462988966900702,
      "loss": 1.2348,
      "theoretical_loss": 3.4826788312328105,
      "tokens_seen": 1841954816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462788365095286,
      "loss": 1.3608,
      "theoretical_loss": 3.4826684177575213,
      "tokens_seen": 1842020352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446258776328987,
      "loss": 1.3317,
      "theoretical_loss": 3.482658004756453,
      "tokens_seen": 1842085888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044623871614844535,
      "loss": 1.4414,
      "theoretical_loss": 3.4826475922295677,
      "tokens_seen": 1842151424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462186559679037,
      "loss": 1.3356,
      "theoretical_loss": 3.4826371801768277,
      "tokens_seen": 1842216960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461985957873621,
      "loss": 1.4634,
      "theoretical_loss": 3.4826267685981933,
      "tokens_seen": 1842282496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461785356068205,
      "loss": 1.2762,
      "theoretical_loss": 3.482616357493627,
      "tokens_seen": 1842348032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044615847542627885,
      "loss": 1.4386,
      "theoretical_loss": 3.4826059468630897,
      "tokens_seen": 1842413568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461384152457372,
      "loss": 1.4646,
      "theoretical_loss": 3.482595536706543,
      "tokens_seen": 1842479104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044611835506519563,
      "loss": 1.3202,
      "theoretical_loss": 3.482585127023949,
      "tokens_seen": 1842544640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446098294884654,
      "loss": 1.3861,
      "theoretical_loss": 3.482574717815269,
      "tokens_seen": 1842610176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044607823470411235,
      "loss": 1.4658,
      "theoretical_loss": 3.4825643090804643,
      "tokens_seen": 1842675712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460581745235707,
      "loss": 1.3338,
      "theoretical_loss": 3.4825539008194966,
      "tokens_seen": 1842741248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460381143430291,
      "loss": 1.3567,
      "theoretical_loss": 3.4825434930323276,
      "tokens_seen": 1842806784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460180541624875,
      "loss": 1.4042,
      "theoretical_loss": 3.482533085718919,
      "tokens_seen": 1842872320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044599799398194586,
      "loss": 1.337,
      "theoretical_loss": 3.4825226788792323,
      "tokens_seen": 1842937856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004459779338014042,
      "loss": 1.3534,
      "theoretical_loss": 3.4825122725132287,
      "tokens_seen": 1843003392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004459578736208626,
      "loss": 1.5066,
      "theoretical_loss": 3.4825018666208702,
      "tokens_seen": 1843068928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000445937813440321,
      "loss": 1.383,
      "theoretical_loss": 3.4824914612021187,
      "tokens_seen": 1843134464
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044591775325977936,
      "loss": 1.2521,
      "theoretical_loss": 3.482481056256935,
      "tokens_seen": 1843200000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458976930792377,
      "loss": 1.3552,
      "theoretical_loss": 3.4824706517852815,
      "tokens_seen": 1843265536
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 672182,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2414642572402954,
      "objective/train/theoretical_loss": 3.4824602477871194,
      "objective/train/tokens_used": 213790176,
      "theoretical_loss": 3.4824602477871194,
      "tokens_seen": 1843331072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458776328986961,
      "loss": 1.3945,
      "theoretical_loss": 3.4824602477871194,
      "tokens_seen": 1843331072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044585757271815445,
      "loss": 1.4466,
      "theoretical_loss": 3.48244984426241,
      "tokens_seen": 1843396608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044583751253761286,
      "loss": 1.2829,
      "theoretical_loss": 3.482439441211116,
      "tokens_seen": 1843462144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458174523570712,
      "loss": 1.2853,
      "theoretical_loss": 3.4824290386331977,
      "tokens_seen": 1843527680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457973921765296,
      "loss": 1.3401,
      "theoretical_loss": 3.4824186365286174,
      "tokens_seen": 1843593216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044577733199598795,
      "loss": 1.3557,
      "theoretical_loss": 3.4824082348973366,
      "tokens_seen": 1843658752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044575727181544637,
      "loss": 1.4321,
      "theoretical_loss": 3.4823978337393173,
      "tokens_seen": 1843724288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044573721163490473,
      "loss": 1.2747,
      "theoretical_loss": 3.4823874330545213,
      "tokens_seen": 1843789824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457171514543631,
      "loss": 1.4164,
      "theoretical_loss": 3.4823770328429093,
      "tokens_seen": 1843855360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044569709127382145,
      "loss": 1.3941,
      "theoretical_loss": 3.4823666331044434,
      "tokens_seen": 1843920896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456770310932798,
      "loss": 1.4176,
      "theoretical_loss": 3.4823562338390857,
      "tokens_seen": 1843986432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044565697091273823,
      "loss": 1.3957,
      "theoretical_loss": 3.4823458350467975,
      "tokens_seen": 1844051968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456369107321966,
      "loss": 1.4266,
      "theoretical_loss": 3.4823354367275403,
      "tokens_seen": 1844117504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044561685055165496,
      "loss": 1.3711,
      "theoretical_loss": 3.482325038881276,
      "tokens_seen": 1844183040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455967903711133,
      "loss": 1.4531,
      "theoretical_loss": 3.482314641507966,
      "tokens_seen": 1844248576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044557673019057173,
      "loss": 1.3947,
      "theoretical_loss": 3.4823042446075725,
      "tokens_seen": 1844314112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455566700100301,
      "loss": 1.2882,
      "theoretical_loss": 3.4822938481800567,
      "tokens_seen": 1844379648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044553660982948846,
      "loss": 1.3367,
      "theoretical_loss": 3.4822834522253805,
      "tokens_seen": 1844445184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455165496489468,
      "loss": 1.3525,
      "theoretical_loss": 3.482273056743506,
      "tokens_seen": 1844510720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454964894684052,
      "loss": 1.3884,
      "theoretical_loss": 3.482262661734394,
      "tokens_seen": 1844576256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454764292878636,
      "loss": 1.2907,
      "theoretical_loss": 3.4822522671980067,
      "tokens_seen": 1844641792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044545636910732196,
      "loss": 1.4096,
      "theoretical_loss": 3.4822418731343063,
      "tokens_seen": 1844707328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454363089267803,
      "loss": 1.4257,
      "theoretical_loss": 3.4822314795432536,
      "tokens_seen": 1844772864
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454162487462387,
      "loss": 1.3026,
      "theoretical_loss": 3.4822210864248113,
      "tokens_seen": 1844838400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453961885656971,
      "loss": 1.2793,
      "theoretical_loss": 3.4822106937789394,
      "tokens_seen": 1844903936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044537612838515547,
      "loss": 1.4845,
      "theoretical_loss": 3.4822003016056016,
      "tokens_seen": 1844969472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044535606820461383,
      "loss": 1.2822,
      "theoretical_loss": 3.4821899099047586,
      "tokens_seen": 1845035008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453360080240722,
      "loss": 1.5107,
      "theoretical_loss": 3.4821795186763724,
      "tokens_seen": 1845100544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044531594784353055,
      "loss": 1.2597,
      "theoretical_loss": 3.482169127920405,
      "tokens_seen": 1845166080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044529588766298897,
      "loss": 1.4273,
      "theoretical_loss": 3.4821587376368175,
      "tokens_seen": 1845231616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044527582748244733,
      "loss": 1.3833,
      "theoretical_loss": 3.482148347825572,
      "tokens_seen": 1845297152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004452557673019057,
      "loss": 1.3197,
      "theoretical_loss": 3.4821379584866303,
      "tokens_seen": 1845362688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044523570712136406,
      "loss": 1.3255,
      "theoretical_loss": 3.482127569619954,
      "tokens_seen": 1845428224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044521564694082247,
      "loss": 1.2742,
      "theoretical_loss": 3.482117181225505,
      "tokens_seen": 1845493760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044519558676028083,
      "loss": 1.3858,
      "theoretical_loss": 3.482106793303245,
      "tokens_seen": 1845559296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044517552657973925,
      "loss": 1.4442,
      "theoretical_loss": 3.4820964058531354,
      "tokens_seen": 1845624832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451554663991976,
      "loss": 1.4046,
      "theoretical_loss": 3.482086018875139,
      "tokens_seen": 1845690368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044513540621865603,
      "loss": 1.3367,
      "theoretical_loss": 3.4820756323692166,
      "tokens_seen": 1845755904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451153460381144,
      "loss": 1.2899,
      "theoretical_loss": 3.4820652463353303,
      "tokens_seen": 1845821440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044509528585757275,
      "loss": 1.4217,
      "theoretical_loss": 3.4820548607734425,
      "tokens_seen": 1845886976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450752256770311,
      "loss": 1.4319,
      "theoretical_loss": 3.4820444756835136,
      "tokens_seen": 1845952512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450551654964895,
      "loss": 1.5386,
      "theoretical_loss": 3.4820340910655068,
      "tokens_seen": 1846018048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450351053159479,
      "loss": 1.3271,
      "theoretical_loss": 3.482023706919383,
      "tokens_seen": 1846083584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044501504513540626,
      "loss": 1.366,
      "theoretical_loss": 3.482013323245104,
      "tokens_seen": 1846149120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449949849548646,
      "loss": 1.4011,
      "theoretical_loss": 3.4820029400426327,
      "tokens_seen": 1846214656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444974924774323,
      "loss": 1.277,
      "theoretical_loss": 3.481992557311929,
      "tokens_seen": 1846280192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449548645937814,
      "loss": 1.3212,
      "theoretical_loss": 3.481982175052957,
      "tokens_seen": 1846345728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044493480441323976,
      "loss": 1.5374,
      "theoretical_loss": 3.481971793265677,
      "tokens_seen": 1846411264
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449147442326981,
      "loss": 1.3373,
      "theoretical_loss": 3.481961411950051,
      "tokens_seen": 1846476800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448946840521565,
      "loss": 1.3793,
      "theoretical_loss": 3.4819510311060418,
      "tokens_seen": 1846542336
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 673286,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5119043588638306,
      "objective/train/theoretical_loss": 3.4819406507336095,
      "objective/train/tokens_used": 217066976,
      "theoretical_loss": 3.4819406507336095,
      "tokens_seen": 1846607872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044487462387161485,
      "loss": 1.454,
      "theoretical_loss": 3.4819406507336095,
      "tokens_seen": 1846607872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044485456369107326,
      "loss": 1.3995,
      "theoretical_loss": 3.4819302708327173,
      "tokens_seen": 1846673408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448345035105316,
      "loss": 1.3099,
      "theoretical_loss": 3.4819198914033267,
      "tokens_seen": 1846738944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044481444332999,
      "loss": 1.3943,
      "theoretical_loss": 3.4819095124454,
      "tokens_seen": 1846804480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044479438314944835,
      "loss": 1.3385,
      "theoretical_loss": 3.4818991339588976,
      "tokens_seen": 1846870016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044477432296890677,
      "loss": 1.2927,
      "theoretical_loss": 3.4818887559437828,
      "tokens_seen": 1846935552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044475426278836513,
      "loss": 1.4671,
      "theoretical_loss": 3.481878378400017,
      "tokens_seen": 1847001088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004447342026078235,
      "loss": 1.402,
      "theoretical_loss": 3.4818680013275625,
      "tokens_seen": 1847066624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044471414242728185,
      "loss": 1.3762,
      "theoretical_loss": 3.4818576247263806,
      "tokens_seen": 1847132160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446940822467402,
      "loss": 1.477,
      "theoretical_loss": 3.481847248596433,
      "tokens_seen": 1847197696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044467402206619863,
      "loss": 1.2732,
      "theoretical_loss": 3.4818368729376816,
      "tokens_seen": 1847263232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444653961885657,
      "loss": 1.4239,
      "theoretical_loss": 3.4818264977500895,
      "tokens_seen": 1847328768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044463390170511536,
      "loss": 1.3003,
      "theoretical_loss": 3.4818161230336173,
      "tokens_seen": 1847394304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446138415245737,
      "loss": 1.3105,
      "theoretical_loss": 3.481805748788227,
      "tokens_seen": 1847459840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044459378134403214,
      "loss": 1.3835,
      "theoretical_loss": 3.4817953750138813,
      "tokens_seen": 1847525376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445737211634905,
      "loss": 1.3508,
      "theoretical_loss": 3.481785001710541,
      "tokens_seen": 1847590912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044455366098294886,
      "loss": 1.2737,
      "theoretical_loss": 3.481774628878169,
      "tokens_seen": 1847656448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445336008024072,
      "loss": 1.4498,
      "theoretical_loss": 3.481764256516726,
      "tokens_seen": 1847721984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445135406218656,
      "loss": 1.4485,
      "theoretical_loss": 3.4817538846261757,
      "tokens_seen": 1847787520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444493480441324,
      "loss": 1.422,
      "theoretical_loss": 3.481743513206479,
      "tokens_seen": 1847853056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044447342026078236,
      "loss": 1.3907,
      "theoretical_loss": 3.4817331422575974,
      "tokens_seen": 1847918592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444533600802407,
      "loss": 1.4349,
      "theoretical_loss": 3.4817227717794936,
      "tokens_seen": 1847984128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444332998996991,
      "loss": 1.3739,
      "theoretical_loss": 3.4817124017721293,
      "tokens_seen": 1848049664
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444132397191575,
      "loss": 1.3609,
      "theoretical_loss": 3.481702032235466,
      "tokens_seen": 1848115200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044439317953861587,
      "loss": 1.3702,
      "theoretical_loss": 3.4816916631694665,
      "tokens_seen": 1848180736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044437311935807423,
      "loss": 1.3909,
      "theoretical_loss": 3.481681294574092,
      "tokens_seen": 1848246272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443530591775326,
      "loss": 1.339,
      "theoretical_loss": 3.4816709264493046,
      "tokens_seen": 1848311808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044433299899699095,
      "loss": 1.4149,
      "theoretical_loss": 3.481660558795067,
      "tokens_seen": 1848377344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044431293881644937,
      "loss": 1.3107,
      "theoretical_loss": 3.4816501916113394,
      "tokens_seen": 1848442880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044429287863590773,
      "loss": 1.386,
      "theoretical_loss": 3.4816398248980858,
      "tokens_seen": 1848508416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004442728184553661,
      "loss": 1.4043,
      "theoretical_loss": 3.481629458655267,
      "tokens_seen": 1848573952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044425275827482446,
      "loss": 1.3636,
      "theoretical_loss": 3.4816190928828457,
      "tokens_seen": 1848639488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044423269809428287,
      "loss": 1.3503,
      "theoretical_loss": 3.4816087275807828,
      "tokens_seen": 1848705024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044421263791374123,
      "loss": 1.4428,
      "theoretical_loss": 3.481598362749041,
      "tokens_seen": 1848770560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441925777331996,
      "loss": 1.4765,
      "theoretical_loss": 3.4815879983875826,
      "tokens_seen": 1848836096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044417251755265796,
      "loss": 1.331,
      "theoretical_loss": 3.481577634496369,
      "tokens_seen": 1848901632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441524573721163,
      "loss": 1.3547,
      "theoretical_loss": 3.4815672710753622,
      "tokens_seen": 1848967168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044413239719157474,
      "loss": 1.3534,
      "theoretical_loss": 3.4815569081245243,
      "tokens_seen": 1849032704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441123370110331,
      "loss": 1.3678,
      "theoretical_loss": 3.481546545643818,
      "tokens_seen": 1849098240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044409227683049146,
      "loss": 1.3229,
      "theoretical_loss": 3.4815361836332044,
      "tokens_seen": 1849163776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440722166499498,
      "loss": 1.3476,
      "theoretical_loss": 3.481525822092646,
      "tokens_seen": 1849229312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044405215646940824,
      "loss": 1.4214,
      "theoretical_loss": 3.481515461022104,
      "tokens_seen": 1849294848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440320962888666,
      "loss": 1.3804,
      "theoretical_loss": 3.4815051004215416,
      "tokens_seen": 1849360384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044401203610832497,
      "loss": 1.3375,
      "theoretical_loss": 3.4814947402909207,
      "tokens_seen": 1849425920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044399197592778333,
      "loss": 1.3461,
      "theoretical_loss": 3.481484380630202,
      "tokens_seen": 1849491456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004439719157472417,
      "loss": 1.2951,
      "theoretical_loss": 3.4814740214393494,
      "tokens_seen": 1849556992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004439518555667001,
      "loss": 1.5267,
      "theoretical_loss": 3.4814636627183235,
      "tokens_seen": 1849622528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044393179538615847,
      "loss": 1.3578,
      "theoretical_loss": 3.4814533044670872,
      "tokens_seen": 1849688064
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044391173520561683,
      "loss": 1.3945,
      "theoretical_loss": 3.481442946685602,
      "tokens_seen": 1849753600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438916750250752,
      "loss": 1.4038,
      "theoretical_loss": 3.4814325893738296,
      "tokens_seen": 1849819136
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 674404,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4075238704681396,
      "objective/train/theoretical_loss": 3.481422232531733,
      "objective/train/tokens_used": 220343776,
      "theoretical_loss": 3.481422232531733,
      "tokens_seen": 1849884672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438716148445336,
      "loss": 1.3272,
      "theoretical_loss": 3.481422232531733,
      "tokens_seen": 1849884672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044385155466399197,
      "loss": 1.3789,
      "theoretical_loss": 3.4814118761592745,
      "tokens_seen": 1849950208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044383149448345033,
      "loss": 1.3812,
      "theoretical_loss": 3.481401520256415,
      "tokens_seen": 1850015744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438114343029087,
      "loss": 1.2449,
      "theoretical_loss": 3.4813911648231173,
      "tokens_seen": 1850081280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437913741223671,
      "loss": 1.5921,
      "theoretical_loss": 3.481380809859343,
      "tokens_seen": 1850146816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437713139418255,
      "loss": 1.5209,
      "theoretical_loss": 3.4813704553650546,
      "tokens_seen": 1850212352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044375125376128384,
      "loss": 1.451,
      "theoretical_loss": 3.481360101340214,
      "tokens_seen": 1850277888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437311935807422,
      "loss": 1.3428,
      "theoretical_loss": 3.4813497477847837,
      "tokens_seen": 1850343424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044371113340020056,
      "loss": 1.337,
      "theoretical_loss": 3.4813393946987254,
      "tokens_seen": 1850408960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044369107321965903,
      "loss": 1.3824,
      "theoretical_loss": 3.481329042082001,
      "tokens_seen": 1850474496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436710130391174,
      "loss": 1.2999,
      "theoretical_loss": 3.481318689934573,
      "tokens_seen": 1850540032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044365095285857576,
      "loss": 1.2647,
      "theoretical_loss": 3.481308338256403,
      "tokens_seen": 1850605568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436308926780341,
      "loss": 1.2919,
      "theoretical_loss": 3.481297987047454,
      "tokens_seen": 1850671104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044361083249749254,
      "loss": 1.3866,
      "theoretical_loss": 3.4812876363076874,
      "tokens_seen": 1850736640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435907723169509,
      "loss": 1.2959,
      "theoretical_loss": 3.4812772860370655,
      "tokens_seen": 1850802176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044357071213640926,
      "loss": 1.4037,
      "theoretical_loss": 3.4812669362355506,
      "tokens_seen": 1850867712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435506519558676,
      "loss": 1.4122,
      "theoretical_loss": 3.4812565869031045,
      "tokens_seen": 1850933248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443530591775326,
      "loss": 1.3627,
      "theoretical_loss": 3.48124623803969,
      "tokens_seen": 1850998784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435105315947844,
      "loss": 1.3424,
      "theoretical_loss": 3.481235889645268,
      "tokens_seen": 1851064320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044349047141424276,
      "loss": 1.4296,
      "theoretical_loss": 3.4812255417198017,
      "tokens_seen": 1851129856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434704112337011,
      "loss": 1.3376,
      "theoretical_loss": 3.4812151942632523,
      "tokens_seen": 1851195392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434503510531595,
      "loss": 1.3155,
      "theoretical_loss": 3.4812048472755834,
      "tokens_seen": 1851260928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434302908726179,
      "loss": 1.3321,
      "theoretical_loss": 3.4811945007567564,
      "tokens_seen": 1851326464
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044341023069207627,
      "loss": 1.2365,
      "theoretical_loss": 3.4811841547067326,
      "tokens_seen": 1851392000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044339017051153463,
      "loss": 1.4327,
      "theoretical_loss": 3.4811738091254756,
      "tokens_seen": 1851457536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443370110330993,
      "loss": 1.3354,
      "theoretical_loss": 3.4811634640129467,
      "tokens_seen": 1851523072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044335005015045135,
      "loss": 1.2978,
      "theoretical_loss": 3.4811531193691083,
      "tokens_seen": 1851588608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044332998996990977,
      "loss": 1.2896,
      "theoretical_loss": 3.4811427751939226,
      "tokens_seen": 1851654144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044330992978936813,
      "loss": 1.3828,
      "theoretical_loss": 3.481132431487352,
      "tokens_seen": 1851719680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432898696088265,
      "loss": 1.3679,
      "theoretical_loss": 3.481122088249358,
      "tokens_seen": 1851785216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044326980942828486,
      "loss": 1.363,
      "theoretical_loss": 3.481111745479904,
      "tokens_seen": 1851850752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044324974924774327,
      "loss": 1.3166,
      "theoretical_loss": 3.4811014031789504,
      "tokens_seen": 1851916288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044322968906720163,
      "loss": 1.3442,
      "theoretical_loss": 3.4810910613464614,
      "tokens_seen": 1851981824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044320962888666,
      "loss": 1.4,
      "theoretical_loss": 3.4810807199823977,
      "tokens_seen": 1852047360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044318956870611836,
      "loss": 1.3621,
      "theoretical_loss": 3.481070379086722,
      "tokens_seen": 1852112896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431695085255767,
      "loss": 1.3715,
      "theoretical_loss": 3.481060038659397,
      "tokens_seen": 1852178432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044314944834503514,
      "loss": 1.4101,
      "theoretical_loss": 3.4810496987003843,
      "tokens_seen": 1852243968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431293881644935,
      "loss": 1.2859,
      "theoretical_loss": 3.481039359209646,
      "tokens_seen": 1852309504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044310932798395186,
      "loss": 1.3805,
      "theoretical_loss": 3.4810290201871448,
      "tokens_seen": 1852375040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004430892678034102,
      "loss": 1.4143,
      "theoretical_loss": 3.481018681632843,
      "tokens_seen": 1852440576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044306920762286864,
      "loss": 1.3947,
      "theoretical_loss": 3.4810083435467023,
      "tokens_seen": 1852506112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443049147442327,
      "loss": 1.4118,
      "theoretical_loss": 3.480998005928685,
      "tokens_seen": 1852571648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044302908726178537,
      "loss": 1.4132,
      "theoretical_loss": 3.480987668778754,
      "tokens_seen": 1852637184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044300902708124373,
      "loss": 1.3111,
      "theoretical_loss": 3.480977332096871,
      "tokens_seen": 1852702720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429889669007021,
      "loss": 1.3243,
      "theoretical_loss": 3.4809669958829983,
      "tokens_seen": 1852768256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429689067201605,
      "loss": 1.2679,
      "theoretical_loss": 3.4809566601370983,
      "tokens_seen": 1852833792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044294884653961887,
      "loss": 1.3461,
      "theoretical_loss": 3.4809463248591332,
      "tokens_seen": 1852899328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044292878635907723,
      "loss": 1.3692,
      "theoretical_loss": 3.4809359900490655,
      "tokens_seen": 1852964864
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429087261785356,
      "loss": 1.4018,
      "theoretical_loss": 3.480925655706857,
      "tokens_seen": 1853030400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442888665997994,
      "loss": 1.4192,
      "theoretical_loss": 3.48091532183247,
      "tokens_seen": 1853095936
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 675642,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4188053607940674,
      "objective/train/theoretical_loss": 3.480904988425867,
      "objective/train/tokens_used": 223620576,
      "theoretical_loss": 3.480904988425867,
      "tokens_seen": 1853161472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044286860581745237,
      "loss": 1.4056,
      "theoretical_loss": 3.480904988425867,
      "tokens_seen": 1853161472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044284854563691073,
      "loss": 1.3625,
      "theoretical_loss": 3.4808946554870106,
      "tokens_seen": 1853227008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428284854563691,
      "loss": 1.4041,
      "theoretical_loss": 3.4808843230158626,
      "tokens_seen": 1853292544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428084252758275,
      "loss": 1.3701,
      "theoretical_loss": 3.480873991012385,
      "tokens_seen": 1853358080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427883650952859,
      "loss": 1.3823,
      "theoretical_loss": 3.480863659476541,
      "tokens_seen": 1853423616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044276830491474424,
      "loss": 1.2795,
      "theoretical_loss": 3.4808533284082928,
      "tokens_seen": 1853489152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427482447342026,
      "loss": 1.417,
      "theoretical_loss": 3.480842997807602,
      "tokens_seen": 1853554688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044272818455366096,
      "loss": 1.3195,
      "theoretical_loss": 3.480832667674431,
      "tokens_seen": 1853620224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427081243731194,
      "loss": 1.3827,
      "theoretical_loss": 3.4808223380087426,
      "tokens_seen": 1853685760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044268806419257774,
      "loss": 1.3279,
      "theoretical_loss": 3.480812008810499,
      "tokens_seen": 1853751296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426680040120361,
      "loss": 1.2933,
      "theoretical_loss": 3.4808016800796624,
      "tokens_seen": 1853816832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044264794383149447,
      "loss": 1.4239,
      "theoretical_loss": 3.480791351816195,
      "tokens_seen": 1853882368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426278836509529,
      "loss": 1.386,
      "theoretical_loss": 3.480781024020059,
      "tokens_seen": 1853947904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044260782347041124,
      "loss": 1.3663,
      "theoretical_loss": 3.4807706966912173,
      "tokens_seen": 1854013440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425877632898696,
      "loss": 1.363,
      "theoretical_loss": 3.480760369829632,
      "tokens_seen": 1854078976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044256770310932797,
      "loss": 1.3051,
      "theoretical_loss": 3.4807500434352647,
      "tokens_seen": 1854144512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044254764292878633,
      "loss": 1.3871,
      "theoretical_loss": 3.4807397175080794,
      "tokens_seen": 1854210048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044252758274824475,
      "loss": 1.3827,
      "theoretical_loss": 3.480729392048037,
      "tokens_seen": 1854275584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425075225677031,
      "loss": 1.3754,
      "theoretical_loss": 3.4807190670551007,
      "tokens_seen": 1854341120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044248746238716147,
      "loss": 1.3318,
      "theoretical_loss": 3.480708742529232,
      "tokens_seen": 1854406656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044246740220661983,
      "loss": 1.3524,
      "theoretical_loss": 3.480698418470394,
      "tokens_seen": 1854472192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044244734202607825,
      "loss": 1.2482,
      "theoretical_loss": 3.480688094878549,
      "tokens_seen": 1854537728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004424272818455366,
      "loss": 1.3631,
      "theoretical_loss": 3.480677771753659,
      "tokens_seen": 1854603264
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442407221664995,
      "loss": 1.3305,
      "theoretical_loss": 3.4806674490956864,
      "tokens_seen": 1854668800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044238716148445334,
      "loss": 1.3104,
      "theoretical_loss": 3.4806571269045943,
      "tokens_seen": 1854734336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423671013039117,
      "loss": 1.4065,
      "theoretical_loss": 3.4806468051803443,
      "tokens_seen": 1854799872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423470411233701,
      "loss": 1.2656,
      "theoretical_loss": 3.480636483922899,
      "tokens_seen": 1854865408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423269809428285,
      "loss": 1.3492,
      "theoretical_loss": 3.480626163132221,
      "tokens_seen": 1854930944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044230692076228684,
      "loss": 1.4317,
      "theoretical_loss": 3.4806158428082723,
      "tokens_seen": 1854996480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422868605817452,
      "loss": 1.3879,
      "theoretical_loss": 3.480605522951016,
      "tokens_seen": 1855062016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422668004012036,
      "loss": 1.2738,
      "theoretical_loss": 3.480595203560414,
      "tokens_seen": 1855127552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442246740220662,
      "loss": 1.3911,
      "theoretical_loss": 3.4805848846364285,
      "tokens_seen": 1855193088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044222668004012034,
      "loss": 1.382,
      "theoretical_loss": 3.480574566179022,
      "tokens_seen": 1855258624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422066198595787,
      "loss": 1.4112,
      "theoretical_loss": 3.4805642481881574,
      "tokens_seen": 1855324160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421865596790371,
      "loss": 1.2419,
      "theoretical_loss": 3.480553930663797,
      "tokens_seen": 1855389696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044216649949849554,
      "loss": 1.3281,
      "theoretical_loss": 3.480543613605903,
      "tokens_seen": 1855455232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421464393179539,
      "loss": 1.4681,
      "theoretical_loss": 3.4805332970144383,
      "tokens_seen": 1855520768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044212637913741226,
      "loss": 1.3449,
      "theoretical_loss": 3.480522980889364,
      "tokens_seen": 1855586304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421063189568706,
      "loss": 1.3734,
      "theoretical_loss": 3.4805126652306444,
      "tokens_seen": 1855651840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044208625877632904,
      "loss": 1.2939,
      "theoretical_loss": 3.4805023500382406,
      "tokens_seen": 1855717376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420661985957874,
      "loss": 1.4606,
      "theoretical_loss": 3.480492035312116,
      "tokens_seen": 1855782912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044204613841524577,
      "loss": 1.287,
      "theoretical_loss": 3.480481721052232,
      "tokens_seen": 1855848448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044202607823470413,
      "loss": 1.2447,
      "theoretical_loss": 3.480471407258552,
      "tokens_seen": 1855913984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420060180541625,
      "loss": 1.2924,
      "theoretical_loss": 3.480461093931037,
      "tokens_seen": 1855979520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004419859578736209,
      "loss": 1.4143,
      "theoretical_loss": 3.480450781069652,
      "tokens_seen": 1856045056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044196589769307927,
      "loss": 1.2944,
      "theoretical_loss": 3.4804404686743577,
      "tokens_seen": 1856110592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044194583751253763,
      "loss": 1.315,
      "theoretical_loss": 3.4804301567451166,
      "tokens_seen": 1856176128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000441925777331996,
      "loss": 1.3002,
      "theoretical_loss": 3.4804198452818915,
      "tokens_seen": 1856241664
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004419057171514544,
      "loss": 1.3368,
      "theoretical_loss": 3.480409534284645,
      "tokens_seen": 1856307200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044188565697091277,
      "loss": 1.3622,
      "theoretical_loss": 3.4803992237533397,
      "tokens_seen": 1856372736
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 676855,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.236203908920288,
      "objective/train/theoretical_loss": 3.4803889136879373,
      "objective/train/tokens_used": 226897376,
      "theoretical_loss": 3.4803889136879373,
      "tokens_seen": 1856438272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044186559679037113,
      "loss": 1.2612,
      "theoretical_loss": 3.4803889136879373,
      "tokens_seen": 1856438272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418455366098295,
      "loss": 1.4207,
      "theoretical_loss": 3.4803786040884015,
      "tokens_seen": 1856503808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418254764292879,
      "loss": 1.4158,
      "theoretical_loss": 3.480368294954694,
      "tokens_seen": 1856569344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418054162487463,
      "loss": 1.3158,
      "theoretical_loss": 3.4803579862867773,
      "tokens_seen": 1856634880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044178535606820464,
      "loss": 1.3681,
      "theoretical_loss": 3.4803476780846143,
      "tokens_seen": 1856700416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441765295887663,
      "loss": 1.3378,
      "theoretical_loss": 3.4803373703481673,
      "tokens_seen": 1856765952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044174523570712136,
      "loss": 1.3602,
      "theoretical_loss": 3.4803270630773984,
      "tokens_seen": 1856831488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417251755265798,
      "loss": 1.3324,
      "theoretical_loss": 3.480316756272271,
      "tokens_seen": 1856897024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044170511534603814,
      "loss": 1.3608,
      "theoretical_loss": 3.4803064499327476,
      "tokens_seen": 1856962560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416850551654965,
      "loss": 1.3962,
      "theoretical_loss": 3.48029614405879,
      "tokens_seen": 1857028096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044166499498495487,
      "loss": 1.3728,
      "theoretical_loss": 3.480285838650361,
      "tokens_seen": 1857093632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416449348044133,
      "loss": 1.3567,
      "theoretical_loss": 3.480275533707423,
      "tokens_seen": 1857159168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044162487462387164,
      "loss": 1.4274,
      "theoretical_loss": 3.480265229229939,
      "tokens_seen": 1857224704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044160481444333,
      "loss": 1.3359,
      "theoretical_loss": 3.4802549252178716,
      "tokens_seen": 1857290240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044158475426278837,
      "loss": 1.3228,
      "theoretical_loss": 3.4802446216711824,
      "tokens_seen": 1857355776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044156469408224673,
      "loss": 1.3424,
      "theoretical_loss": 3.4802343185898352,
      "tokens_seen": 1857421312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044154463390170515,
      "loss": 1.4352,
      "theoretical_loss": 3.480224015973792,
      "tokens_seen": 1857486848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415245737211635,
      "loss": 1.4175,
      "theoretical_loss": 3.480213713823015,
      "tokens_seen": 1857552384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044150451354062187,
      "loss": 1.374,
      "theoretical_loss": 3.4802034121374676,
      "tokens_seen": 1857617920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044148445336008023,
      "loss": 1.4289,
      "theoretical_loss": 3.480193110917112,
      "tokens_seen": 1857683456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044146439317953865,
      "loss": 1.4383,
      "theoretical_loss": 3.48018281016191,
      "tokens_seen": 1857748992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441444332998997,
      "loss": 1.2962,
      "theoretical_loss": 3.4801725098718257,
      "tokens_seen": 1857814528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414242728184554,
      "loss": 1.294,
      "theoretical_loss": 3.480162210046821,
      "tokens_seen": 1857880064
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044140421263791374,
      "loss": 1.3548,
      "theoretical_loss": 3.4801519106868577,
      "tokens_seen": 1857945600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413841524573721,
      "loss": 1.3539,
      "theoretical_loss": 3.480141611791899,
      "tokens_seen": 1858011136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413640922768305,
      "loss": 1.3828,
      "theoretical_loss": 3.480131313361908,
      "tokens_seen": 1858076672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413440320962889,
      "loss": 1.3488,
      "theoretical_loss": 3.480121015396847,
      "tokens_seen": 1858142208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044132397191574724,
      "loss": 1.4505,
      "theoretical_loss": 3.480110717896679,
      "tokens_seen": 1858207744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413039117352056,
      "loss": 1.3774,
      "theoretical_loss": 3.480100420861365,
      "tokens_seen": 1858273280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441283851554664,
      "loss": 1.4,
      "theoretical_loss": 3.4800901242908693,
      "tokens_seen": 1858338816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412637913741224,
      "loss": 1.3801,
      "theoretical_loss": 3.480079828185154,
      "tokens_seen": 1858404352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044124373119358074,
      "loss": 1.3493,
      "theoretical_loss": 3.480069532544182,
      "tokens_seen": 1858469888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412236710130391,
      "loss": 1.4411,
      "theoretical_loss": 3.4800592373679153,
      "tokens_seen": 1858535424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044120361083249747,
      "loss": 1.3594,
      "theoretical_loss": 3.4800489426563166,
      "tokens_seen": 1858600960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411835506519559,
      "loss": 1.2677,
      "theoretical_loss": 3.4800386484093497,
      "tokens_seen": 1858666496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044116349047141425,
      "loss": 1.2797,
      "theoretical_loss": 3.4800283546269757,
      "tokens_seen": 1858732032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411434302908726,
      "loss": 1.3471,
      "theoretical_loss": 3.480018061309158,
      "tokens_seen": 1858797568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044112337011033097,
      "loss": 1.3357,
      "theoretical_loss": 3.4800077684558595,
      "tokens_seen": 1858863104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411033099297894,
      "loss": 1.4405,
      "theoretical_loss": 3.4799974760670427,
      "tokens_seen": 1858928640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044108324974924775,
      "loss": 1.2756,
      "theoretical_loss": 3.4799871841426695,
      "tokens_seen": 1858994176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410631895687061,
      "loss": 1.2834,
      "theoretical_loss": 3.4799768926827035,
      "tokens_seen": 1859059712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410431293881645,
      "loss": 1.5183,
      "theoretical_loss": 3.4799666016871074,
      "tokens_seen": 1859125248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044102306920762284,
      "loss": 1.3528,
      "theoretical_loss": 3.479956311155843,
      "tokens_seen": 1859190784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044100300902708125,
      "loss": 1.2681,
      "theoretical_loss": 3.4799460210888737,
      "tokens_seen": 1859256320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004409829488465396,
      "loss": 1.2271,
      "theoretical_loss": 3.4799357314861625,
      "tokens_seen": 1859321856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440962888665998,
      "loss": 1.4465,
      "theoretical_loss": 3.479925442347671,
      "tokens_seen": 1859387392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044094282848545634,
      "loss": 1.2832,
      "theoretical_loss": 3.479915153673363,
      "tokens_seen": 1859452928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044092276830491476,
      "loss": 1.2743,
      "theoretical_loss": 3.4799048654632,
      "tokens_seen": 1859518464
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004409027081243731,
      "loss": 1.3313,
      "theoretical_loss": 3.479894577717146,
      "tokens_seen": 1859584000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408826479438315,
      "loss": 1.4087,
      "theoretical_loss": 3.479884290435163,
      "tokens_seen": 1859649536
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 678160,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2900208234786987,
      "objective/train/theoretical_loss": 3.4798740036172138,
      "objective/train/tokens_used": 230174176,
      "theoretical_loss": 3.4798740036172138,
      "tokens_seen": 1859715072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044086258776328984,
      "loss": 1.308,
      "theoretical_loss": 3.4798740036172138,
      "tokens_seen": 1859715072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408425275827482,
      "loss": 1.3766,
      "theoretical_loss": 3.479863717263261,
      "tokens_seen": 1859780608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408224674022066,
      "loss": 1.3822,
      "theoretical_loss": 3.4798534313732676,
      "tokens_seen": 1859846144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440802407221665,
      "loss": 1.3408,
      "theoretical_loss": 3.4798431459471963,
      "tokens_seen": 1859911680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044078234704112335,
      "loss": 1.3509,
      "theoretical_loss": 3.4798328609850095,
      "tokens_seen": 1859977216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407622868605817,
      "loss": 1.3909,
      "theoretical_loss": 3.4798225764866704,
      "tokens_seen": 1860042752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407422266800401,
      "loss": 1.3062,
      "theoretical_loss": 3.4798122924521415,
      "tokens_seen": 1860108288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407221664994985,
      "loss": 1.3854,
      "theoretical_loss": 3.4798020088813852,
      "tokens_seen": 1860173824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044070210631895685,
      "loss": 1.3777,
      "theoretical_loss": 3.4797917257743647,
      "tokens_seen": 1860239360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044068204613841527,
      "loss": 1.3238,
      "theoretical_loss": 3.479781443131043,
      "tokens_seen": 1860304896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406619859578737,
      "loss": 1.3782,
      "theoretical_loss": 3.479771160951382,
      "tokens_seen": 1860370432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044064192577733204,
      "loss": 1.3323,
      "theoretical_loss": 3.479760879235345,
      "tokens_seen": 1860435968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406218655967904,
      "loss": 1.3067,
      "theoretical_loss": 3.4797505979828953,
      "tokens_seen": 1860501504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044060180541624877,
      "loss": 1.3041,
      "theoretical_loss": 3.4797403171939942,
      "tokens_seen": 1860567040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044058174523570713,
      "loss": 1.4958,
      "theoretical_loss": 3.479730036868606,
      "tokens_seen": 1860632576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044056168505516555,
      "loss": 1.3577,
      "theoretical_loss": 3.479719757006693,
      "tokens_seen": 1860698112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405416248746239,
      "loss": 1.318,
      "theoretical_loss": 3.4797094776082176,
      "tokens_seen": 1860763648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044052156469408227,
      "loss": 1.375,
      "theoretical_loss": 3.4796991986731425,
      "tokens_seen": 1860829184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044050150451354063,
      "loss": 1.304,
      "theoretical_loss": 3.479688920201431,
      "tokens_seen": 1860894720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044048144433299905,
      "loss": 1.4008,
      "theoretical_loss": 3.4796786421930457,
      "tokens_seen": 1860960256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404613841524574,
      "loss": 1.2228,
      "theoretical_loss": 3.4796683646479494,
      "tokens_seen": 1861025792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404413239719158,
      "loss": 1.3021,
      "theoretical_loss": 3.479658087566105,
      "tokens_seen": 1861091328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044042126379137414,
      "loss": 1.3481,
      "theoretical_loss": 3.4796478109474753,
      "tokens_seen": 1861156864
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404012036108325,
      "loss": 1.4006,
      "theoretical_loss": 3.4796375347920225,
      "tokens_seen": 1861222400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403811434302909,
      "loss": 1.3697,
      "theoretical_loss": 3.4796272590997104,
      "tokens_seen": 1861287936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403610832497493,
      "loss": 1.3966,
      "theoretical_loss": 3.4796169838705016,
      "tokens_seen": 1861353472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044034102306920764,
      "loss": 1.2961,
      "theoretical_loss": 3.479606709104358,
      "tokens_seen": 1861419008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440320962888666,
      "loss": 1.464,
      "theoretical_loss": 3.4795964348012434,
      "tokens_seen": 1861484544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403009027081244,
      "loss": 1.3555,
      "theoretical_loss": 3.4795861609611203,
      "tokens_seen": 1861550080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402808425275828,
      "loss": 1.3636,
      "theoretical_loss": 3.4795758875839518,
      "tokens_seen": 1861615616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044026078234704114,
      "loss": 1.3916,
      "theoretical_loss": 3.4795656146697,
      "tokens_seen": 1861681152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402407221664995,
      "loss": 1.39,
      "theoretical_loss": 3.4795553422183287,
      "tokens_seen": 1861746688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044022066198595787,
      "loss": 1.4329,
      "theoretical_loss": 3.4795450702298005,
      "tokens_seen": 1861812224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402006018054163,
      "loss": 1.3499,
      "theoretical_loss": 3.4795347987040777,
      "tokens_seen": 1861877760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044018054162487465,
      "loss": 1.3222,
      "theoretical_loss": 3.4795245276411233,
      "tokens_seen": 1861943296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440160481444333,
      "loss": 1.5041,
      "theoretical_loss": 3.4795142570409006,
      "tokens_seen": 1862008832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044014042126379137,
      "loss": 1.4517,
      "theoretical_loss": 3.4795039869033726,
      "tokens_seen": 1862074368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004401203610832498,
      "loss": 1.2425,
      "theoretical_loss": 3.4794937172285016,
      "tokens_seen": 1862139904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044010030090270815,
      "loss": 1.423,
      "theoretical_loss": 3.479483448016251,
      "tokens_seen": 1862205440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400802407221665,
      "loss": 1.4235,
      "theoretical_loss": 3.479473179266583,
      "tokens_seen": 1862270976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400601805416249,
      "loss": 1.3433,
      "theoretical_loss": 3.4794629109794606,
      "tokens_seen": 1862336512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044004012036108324,
      "loss": 1.3064,
      "theoretical_loss": 3.479452643154848,
      "tokens_seen": 1862402048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044002006018054165,
      "loss": 1.3943,
      "theoretical_loss": 3.479442375792706,
      "tokens_seen": 1862467584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044,
      "loss": 1.4319,
      "theoretical_loss": 3.4794321088929987,
      "tokens_seen": 1862533120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004399799398194584,
      "loss": 1.2236,
      "theoretical_loss": 3.4794218424556895,
      "tokens_seen": 1862598656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043995987963891674,
      "loss": 1.3413,
      "theoretical_loss": 3.4794115764807403,
      "tokens_seen": 1862664192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043993981945837516,
      "loss": 1.37,
      "theoretical_loss": 3.479401310968114,
      "tokens_seen": 1862729728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004399197592778335,
      "loss": 1.3581,
      "theoretical_loss": 3.479391045917774,
      "tokens_seen": 1862795264
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398996990972919,
      "loss": 1.3499,
      "theoretical_loss": 3.4793807813296835,
      "tokens_seen": 1862860800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043987963891675024,
      "loss": 1.2575,
      "theoretical_loss": 3.479370517203805,
      "tokens_seen": 1862926336
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 679393,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3746061325073242,
      "objective/train/theoretical_loss": 3.4793602535401016,
      "objective/train/tokens_used": 233450976,
      "theoretical_loss": 3.4793602535401016,
      "tokens_seen": 1862991872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398595787362086,
      "loss": 1.2845,
      "theoretical_loss": 3.4793602535401016,
      "tokens_seen": 1862991872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439839518555667,
      "loss": 1.2539,
      "theoretical_loss": 3.4793499903385356,
      "tokens_seen": 1863057408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398194583751254,
      "loss": 1.4335,
      "theoretical_loss": 3.4793397275990703,
      "tokens_seen": 1863122944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043979939819458375,
      "loss": 1.2686,
      "theoretical_loss": 3.479329465321669,
      "tokens_seen": 1863188480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397793380140421,
      "loss": 1.3696,
      "theoretical_loss": 3.4793192035062948,
      "tokens_seen": 1863254016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397592778335005,
      "loss": 1.309,
      "theoretical_loss": 3.47930894215291,
      "tokens_seen": 1863319552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397392176529589,
      "loss": 1.3723,
      "theoretical_loss": 3.4792986812614775,
      "tokens_seen": 1863385088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043971915747241725,
      "loss": 1.2588,
      "theoretical_loss": 3.479288420831961,
      "tokens_seen": 1863450624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396990972918756,
      "loss": 1.3809,
      "theoretical_loss": 3.479278160864323,
      "tokens_seen": 1863516160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043967903711133403,
      "loss": 1.4304,
      "theoretical_loss": 3.4792679013585266,
      "tokens_seen": 1863581696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396589769307924,
      "loss": 1.3769,
      "theoretical_loss": 3.4792576423145345,
      "tokens_seen": 1863647232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043963891675025075,
      "loss": 1.4156,
      "theoretical_loss": 3.4792473837323095,
      "tokens_seen": 1863712768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396188565697091,
      "loss": 1.3513,
      "theoretical_loss": 3.4792371256118155,
      "tokens_seen": 1863778304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395987963891675,
      "loss": 1.3958,
      "theoretical_loss": 3.479226867953014,
      "tokens_seen": 1863843840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395787362086259,
      "loss": 1.3839,
      "theoretical_loss": 3.47921661075587,
      "tokens_seen": 1863909376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043955867602808426,
      "loss": 1.3333,
      "theoretical_loss": 3.4792063540203446,
      "tokens_seen": 1863974912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395386158475426,
      "loss": 1.3939,
      "theoretical_loss": 3.479196097746402,
      "tokens_seen": 1864040448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439518555667001,
      "loss": 1.373,
      "theoretical_loss": 3.4791858419340045,
      "tokens_seen": 1864105984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394984954864594,
      "loss": 1.3925,
      "theoretical_loss": 3.479175586583116,
      "tokens_seen": 1864171520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043947843530591776,
      "loss": 1.4448,
      "theoretical_loss": 3.4791653316936983,
      "tokens_seen": 1864237056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394583751253761,
      "loss": 1.3389,
      "theoretical_loss": 3.479155077265715,
      "tokens_seen": 1864302592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394383149448345,
      "loss": 1.3119,
      "theoretical_loss": 3.479144823299129,
      "tokens_seen": 1864368128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043941825476429285,
      "loss": 1.4422,
      "theoretical_loss": 3.479134569793904,
      "tokens_seen": 1864433664
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043939819458375126,
      "loss": 1.214,
      "theoretical_loss": 3.479124316750002,
      "tokens_seen": 1864499200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393781344032096,
      "loss": 1.3549,
      "theoretical_loss": 3.479114064167387,
      "tokens_seen": 1864564736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439358074222668,
      "loss": 1.3359,
      "theoretical_loss": 3.479103812046021,
      "tokens_seen": 1864630272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043933801404212635,
      "loss": 1.336,
      "theoretical_loss": 3.4790935603858677,
      "tokens_seen": 1864695808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043931795386158477,
      "loss": 1.4317,
      "theoretical_loss": 3.4790833091868905,
      "tokens_seen": 1864761344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043929789368104313,
      "loss": 1.3956,
      "theoretical_loss": 3.4790730584490515,
      "tokens_seen": 1864826880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392778335005015,
      "loss": 1.2467,
      "theoretical_loss": 3.479062808172314,
      "tokens_seen": 1864892416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043925777331995985,
      "loss": 1.3515,
      "theoretical_loss": 3.479052558356642,
      "tokens_seen": 1864957952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392377131394182,
      "loss": 1.3135,
      "theoretical_loss": 3.479042309001997,
      "tokens_seen": 1865023488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043921765295887663,
      "loss": 1.4365,
      "theoretical_loss": 3.4790320601083433,
      "tokens_seen": 1865089024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439197592778335,
      "loss": 1.3232,
      "theoretical_loss": 3.4790218116756435,
      "tokens_seen": 1865154560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391775325977934,
      "loss": 1.3017,
      "theoretical_loss": 3.4790115637038608,
      "tokens_seen": 1865220096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043915747241725177,
      "loss": 1.2733,
      "theoretical_loss": 3.4790013161929583,
      "tokens_seen": 1865285632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391374122367102,
      "loss": 1.3725,
      "theoretical_loss": 3.478991069142899,
      "tokens_seen": 1865351168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043911735205616855,
      "loss": 1.3736,
      "theoretical_loss": 3.4789808225536456,
      "tokens_seen": 1865416704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390972918756269,
      "loss": 1.3882,
      "theoretical_loss": 3.478970576425162,
      "tokens_seen": 1865482240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390772316950853,
      "loss": 1.3529,
      "theoretical_loss": 3.4789603307574106,
      "tokens_seen": 1865547776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043905717151454364,
      "loss": 1.4199,
      "theoretical_loss": 3.478950085550355,
      "tokens_seen": 1865613312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043903711133400205,
      "loss": 1.3859,
      "theoretical_loss": 3.478939840803958,
      "tokens_seen": 1865678848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390170511534604,
      "loss": 1.3307,
      "theoretical_loss": 3.478929596518183,
      "tokens_seen": 1865744384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389969909729188,
      "loss": 1.3402,
      "theoretical_loss": 3.4789193526929925,
      "tokens_seen": 1865809920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043897693079237714,
      "loss": 1.3475,
      "theoretical_loss": 3.4789091093283497,
      "tokens_seen": 1865875456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043895687061183556,
      "loss": 1.2773,
      "theoretical_loss": 3.4788988664242186,
      "tokens_seen": 1865940992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389368104312939,
      "loss": 1.3979,
      "theoretical_loss": 3.478888623980562,
      "tokens_seen": 1866006528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389167502507523,
      "loss": 1.4352,
      "theoretical_loss": 3.478878381997342,
      "tokens_seen": 1866072064
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043889669007021064,
      "loss": 1.4133,
      "theoretical_loss": 3.478868140474523,
      "tokens_seen": 1866137600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438876629889669,
      "loss": 1.3842,
      "theoretical_loss": 3.478857899412067,
      "tokens_seen": 1866203136
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 681054,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4403936862945557,
      "objective/train/theoretical_loss": 3.478847658809938,
      "objective/train/tokens_used": 236727776,
      "theoretical_loss": 3.478847658809938,
      "tokens_seen": 1866268672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388565697091274,
      "loss": 1.3943,
      "theoretical_loss": 3.478847658809938,
      "tokens_seen": 1866268672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388365095285858,
      "loss": 1.262,
      "theoretical_loss": 3.4788374186680997,
      "tokens_seen": 1866334208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043881644934804415,
      "loss": 1.3017,
      "theoretical_loss": 3.4788271789865135,
      "tokens_seen": 1866399744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387963891675025,
      "loss": 1.387,
      "theoretical_loss": 3.478816939765144,
      "tokens_seen": 1866465280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387763289869609,
      "loss": 1.3761,
      "theoretical_loss": 3.478806701003954,
      "tokens_seen": 1866530816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387562688064193,
      "loss": 1.2479,
      "theoretical_loss": 3.478796462702906,
      "tokens_seen": 1866596352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043873620862587765,
      "loss": 1.3339,
      "theoretical_loss": 3.4787862248619645,
      "tokens_seen": 1866661888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438716148445336,
      "loss": 1.3381,
      "theoretical_loss": 3.4787759874810913,
      "tokens_seen": 1866727424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043869608826479443,
      "loss": 1.4082,
      "theoretical_loss": 3.4787657505602505,
      "tokens_seen": 1866792960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386760280842528,
      "loss": 1.3474,
      "theoretical_loss": 3.478755514099405,
      "tokens_seen": 1866858496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043865596790371115,
      "loss": 1.4598,
      "theoretical_loss": 3.4787452780985175,
      "tokens_seen": 1866924032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386359077231695,
      "loss": 1.3573,
      "theoretical_loss": 3.4787350425575516,
      "tokens_seen": 1866989568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386158475426279,
      "loss": 1.3373,
      "theoretical_loss": 3.478724807476471,
      "tokens_seen": 1867055104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385957873620863,
      "loss": 1.3677,
      "theoretical_loss": 3.478714572855238,
      "tokens_seen": 1867120640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043857572718154466,
      "loss": 1.3373,
      "theoretical_loss": 3.4787043386938166,
      "tokens_seen": 1867186176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438555667001003,
      "loss": 1.5246,
      "theoretical_loss": 3.4786941049921687,
      "tokens_seen": 1867251712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385356068204614,
      "loss": 1.3978,
      "theoretical_loss": 3.4786838717502593,
      "tokens_seen": 1867317248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385155466399198,
      "loss": 1.3118,
      "theoretical_loss": 3.4786736389680506,
      "tokens_seen": 1867382784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043849548645937816,
      "loss": 1.4362,
      "theoretical_loss": 3.4786634066455058,
      "tokens_seen": 1867448320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384754262788365,
      "loss": 1.395,
      "theoretical_loss": 3.478653174782588,
      "tokens_seen": 1867513856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384553660982949,
      "loss": 1.3358,
      "theoretical_loss": 3.4786429433792607,
      "tokens_seen": 1867579392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043843530591775325,
      "loss": 1.3989,
      "theoretical_loss": 3.4786327124354877,
      "tokens_seen": 1867644928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043841524573721166,
      "loss": 1.3759,
      "theoretical_loss": 3.4786224819512306,
      "tokens_seen": 1867710464
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043839518555667,
      "loss": 1.3507,
      "theoretical_loss": 3.478612251926455,
      "tokens_seen": 1867776000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383751253761284,
      "loss": 1.3864,
      "theoretical_loss": 3.478602022361122,
      "tokens_seen": 1867841536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043835506519558675,
      "loss": 1.4664,
      "theoretical_loss": 3.4785917932551955,
      "tokens_seen": 1867907072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043833500501504517,
      "loss": 1.3423,
      "theoretical_loss": 3.478581564608639,
      "tokens_seen": 1867972608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043831494483450353,
      "loss": 1.4152,
      "theoretical_loss": 3.4785713364214166,
      "tokens_seen": 1868038144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382948846539619,
      "loss": 1.4306,
      "theoretical_loss": 3.47856110869349,
      "tokens_seen": 1868103680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043827482447342025,
      "loss": 1.3792,
      "theoretical_loss": 3.4785508814248227,
      "tokens_seen": 1868169216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382547642928786,
      "loss": 1.4546,
      "theoretical_loss": 3.478540654615379,
      "tokens_seen": 1868234752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043823470411233703,
      "loss": 1.3498,
      "theoretical_loss": 3.4785304282651213,
      "tokens_seen": 1868300288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382146439317954,
      "loss": 1.2888,
      "theoretical_loss": 3.478520202374013,
      "tokens_seen": 1868365824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043819458375125376,
      "loss": 1.4168,
      "theoretical_loss": 3.478509976942018,
      "tokens_seen": 1868431360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381745235707121,
      "loss": 1.397,
      "theoretical_loss": 3.478499751969098,
      "tokens_seen": 1868496896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043815446339017053,
      "loss": 1.4325,
      "theoretical_loss": 3.4784895274552183,
      "tokens_seen": 1868562432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381344032096289,
      "loss": 1.4241,
      "theoretical_loss": 3.478479303400341,
      "tokens_seen": 1868627968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043811434302908726,
      "loss": 1.4264,
      "theoretical_loss": 3.4784690798044298,
      "tokens_seen": 1868693504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380942828485456,
      "loss": 1.37,
      "theoretical_loss": 3.478458856667448,
      "tokens_seen": 1868759040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438074222668004,
      "loss": 1.4807,
      "theoretical_loss": 3.478448633989358,
      "tokens_seen": 1868824576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380541624874624,
      "loss": 1.5314,
      "theoretical_loss": 3.4784384117701244,
      "tokens_seen": 1868890112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043803410230692076,
      "loss": 1.3582,
      "theoretical_loss": 3.4784281900097103,
      "tokens_seen": 1868955648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380140421263791,
      "loss": 1.4331,
      "theoretical_loss": 3.4784179687080785,
      "tokens_seen": 1869021184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004379939819458375,
      "loss": 1.316,
      "theoretical_loss": 3.4784077478651922,
      "tokens_seen": 1869086720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004379739217652959,
      "loss": 1.4006,
      "theoretical_loss": 3.4783975274810155,
      "tokens_seen": 1869152256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043795386158475427,
      "loss": 1.3135,
      "theoretical_loss": 3.478387307555511,
      "tokens_seen": 1869217792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043793380140421263,
      "loss": 1.374,
      "theoretical_loss": 3.478377088088642,
      "tokens_seen": 1869283328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437913741223671,
      "loss": 1.3584,
      "theoretical_loss": 3.4783668690803733,
      "tokens_seen": 1869348864
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043789368104312935,
      "loss": 1.3697,
      "theoretical_loss": 3.478356650530666,
      "tokens_seen": 1869414400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043787362086258777,
      "loss": 1.4704,
      "theoretical_loss": 3.4783464324394853,
      "tokens_seen": 1869479936
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 682171,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3038382530212402,
      "objective/train/theoretical_loss": 3.4783362148067933,
      "objective/train/tokens_used": 240004576,
      "theoretical_loss": 3.4783362148067933,
      "tokens_seen": 1869545472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043785356068204613,
      "loss": 1.372,
      "theoretical_loss": 3.4783362148067933,
      "tokens_seen": 1869545472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004378335005015045,
      "loss": 1.3688,
      "theoretical_loss": 3.4783259976325542,
      "tokens_seen": 1869611008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043781344032096286,
      "loss": 1.4204,
      "theoretical_loss": 3.4783157809167307,
      "tokens_seen": 1869676544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043779338014042127,
      "loss": 1.5168,
      "theoretical_loss": 3.4783055646592866,
      "tokens_seen": 1869742080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043777331995987963,
      "loss": 1.4738,
      "theoretical_loss": 3.4782953488601853,
      "tokens_seen": 1869807616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437753259779338,
      "loss": 1.3557,
      "theoretical_loss": 3.47828513351939,
      "tokens_seen": 1869873152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043773319959879636,
      "loss": 1.4439,
      "theoretical_loss": 3.478274918636864,
      "tokens_seen": 1869938688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377131394182547,
      "loss": 1.2031,
      "theoretical_loss": 3.478264704212571,
      "tokens_seen": 1870004224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043769307923771314,
      "loss": 1.3398,
      "theoretical_loss": 3.4782544902464743,
      "tokens_seen": 1870069760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043767301905717155,
      "loss": 1.3298,
      "theoretical_loss": 3.4782442767385366,
      "tokens_seen": 1870135296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376529588766299,
      "loss": 1.4249,
      "theoretical_loss": 3.4782340636887223,
      "tokens_seen": 1870200832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376328986960883,
      "loss": 1.3177,
      "theoretical_loss": 3.4782238510969945,
      "tokens_seen": 1870266368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376128385155467,
      "loss": 1.4201,
      "theoretical_loss": 3.4782136389633163,
      "tokens_seen": 1870331904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043759277833500506,
      "loss": 1.3452,
      "theoretical_loss": 3.478203427287651,
      "tokens_seen": 1870397440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375727181544634,
      "loss": 1.3268,
      "theoretical_loss": 3.4781932160699625,
      "tokens_seen": 1870462976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375526579739218,
      "loss": 1.261,
      "theoretical_loss": 3.478183005310214,
      "tokens_seen": 1870528512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375325977933802,
      "loss": 1.3436,
      "theoretical_loss": 3.478172795008369,
      "tokens_seen": 1870594048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043751253761283856,
      "loss": 1.2855,
      "theoretical_loss": 3.478162585164391,
      "tokens_seen": 1870659584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374924774322969,
      "loss": 1.3365,
      "theoretical_loss": 3.478152375778243,
      "tokens_seen": 1870725120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374724172517553,
      "loss": 1.5139,
      "theoretical_loss": 3.4781421668498886,
      "tokens_seen": 1870790656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043745235707121365,
      "loss": 1.3465,
      "theoretical_loss": 3.4781319583792913,
      "tokens_seen": 1870856192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043743229689067206,
      "loss": 1.3418,
      "theoretical_loss": 3.478121750366415,
      "tokens_seen": 1870921728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374122367101304,
      "loss": 1.3394,
      "theoretical_loss": 3.478111542811222,
      "tokens_seen": 1870987264
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373921765295888,
      "loss": 1.5021,
      "theoretical_loss": 3.4781013357136774,
      "tokens_seen": 1871052800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043737211634904715,
      "loss": 1.4359,
      "theoretical_loss": 3.4780911290737433,
      "tokens_seen": 1871118336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043735205616850557,
      "loss": 1.435,
      "theoretical_loss": 3.478080922891383,
      "tokens_seen": 1871183872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043733199598796393,
      "loss": 1.4357,
      "theoretical_loss": 3.4780707171665615,
      "tokens_seen": 1871249408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373119358074223,
      "loss": 1.4165,
      "theoretical_loss": 3.478060511899241,
      "tokens_seen": 1871314944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043729187562688065,
      "loss": 1.4591,
      "theoretical_loss": 3.478050307089385,
      "tokens_seen": 1871380480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437271815446339,
      "loss": 1.3839,
      "theoretical_loss": 3.4780401027369576,
      "tokens_seen": 1871446016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043725175526579743,
      "loss": 1.4561,
      "theoretical_loss": 3.478029898841922,
      "tokens_seen": 1871511552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372316950852558,
      "loss": 1.3793,
      "theoretical_loss": 3.4780196954042415,
      "tokens_seen": 1871577088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043721163490471416,
      "loss": 1.3833,
      "theoretical_loss": 3.4780094924238796,
      "tokens_seen": 1871642624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371915747241725,
      "loss": 1.4802,
      "theoretical_loss": 3.4779992899008,
      "tokens_seen": 1871708160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043717151454363093,
      "loss": 1.407,
      "theoretical_loss": 3.477989087834966,
      "tokens_seen": 1871773696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371514543630893,
      "loss": 1.2718,
      "theoretical_loss": 3.477978886226341,
      "tokens_seen": 1871839232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043713139418254766,
      "loss": 1.4083,
      "theoretical_loss": 3.477968685074889,
      "tokens_seen": 1871904768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437111334002006,
      "loss": 1.3741,
      "theoretical_loss": 3.477958484380573,
      "tokens_seen": 1871970304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370912738214644,
      "loss": 1.4443,
      "theoretical_loss": 3.477948284143357,
      "tokens_seen": 1872035840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370712136409228,
      "loss": 1.4117,
      "theoretical_loss": 3.477938084363204,
      "tokens_seen": 1872101376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043705115346038116,
      "loss": 1.3838,
      "theoretical_loss": 3.477927885040078,
      "tokens_seen": 1872166912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370310932798395,
      "loss": 1.346,
      "theoretical_loss": 3.477917686173942,
      "tokens_seen": 1872232448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370110330992979,
      "loss": 1.2985,
      "theoretical_loss": 3.4779074877647593,
      "tokens_seen": 1872297984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004369909729187563,
      "loss": 1.3435,
      "theoretical_loss": 3.4778972898124945,
      "tokens_seen": 1872363520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043697091273821467,
      "loss": 1.326,
      "theoretical_loss": 3.477887092317111,
      "tokens_seen": 1872429056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043695085255767303,
      "loss": 1.3718,
      "theoretical_loss": 3.4778768952785715,
      "tokens_seen": 1872494592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004369307923771314,
      "loss": 1.3345,
      "theoretical_loss": 3.4778666986968396,
      "tokens_seen": 1872560128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043691073219658975,
      "loss": 1.4466,
      "theoretical_loss": 3.4778565025718797,
      "tokens_seen": 1872625664
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043689067201604817,
      "loss": 1.3711,
      "theoretical_loss": 3.477846306903655,
      "tokens_seen": 1872691200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043687061183550653,
      "loss": 1.3178,
      "theoretical_loss": 3.4778361116921284,
      "tokens_seen": 1872756736
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 683408,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4890151023864746,
      "objective/train/theoretical_loss": 3.477825916937264,
      "objective/train/tokens_used": 243281376,
      "theoretical_loss": 3.477825916937264,
      "tokens_seen": 1872822272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004368505516549649,
      "loss": 1.5197,
      "theoretical_loss": 3.477825916937264,
      "tokens_seen": 1872822272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043683049147442326,
      "loss": 1.4975,
      "theoretical_loss": 3.4778157226390256,
      "tokens_seen": 1872887808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043681043129388167,
      "loss": 1.445,
      "theoretical_loss": 3.4778055287973766,
      "tokens_seen": 1872953344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043679037111334003,
      "loss": 1.3477,
      "theoretical_loss": 3.4777953354122806,
      "tokens_seen": 1873018880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367703109327984,
      "loss": 1.4355,
      "theoretical_loss": 3.4777851424837007,
      "tokens_seen": 1873084416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043675025075225676,
      "loss": 1.5064,
      "theoretical_loss": 3.477774950011601,
      "tokens_seen": 1873149952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367301905717151,
      "loss": 1.3944,
      "theoretical_loss": 3.4777647579959448,
      "tokens_seen": 1873215488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043671013039117354,
      "loss": 1.4303,
      "theoretical_loss": 3.477754566436696,
      "tokens_seen": 1873281024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366900702106319,
      "loss": 1.3826,
      "theoretical_loss": 3.4777443753338178,
      "tokens_seen": 1873346560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043667001003009026,
      "loss": 1.4101,
      "theoretical_loss": 3.4777341846872742,
      "tokens_seen": 1873412096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366499498495486,
      "loss": 1.3987,
      "theoretical_loss": 3.477723994497029,
      "tokens_seen": 1873477632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043662988966900704,
      "loss": 1.5339,
      "theoretical_loss": 3.477713804763045,
      "tokens_seen": 1873543168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366098294884654,
      "loss": 1.42,
      "theoretical_loss": 3.4777036154852863,
      "tokens_seen": 1873608704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043658976930792376,
      "loss": 1.3668,
      "theoretical_loss": 3.4776934266637167,
      "tokens_seen": 1873674240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043656970912738213,
      "loss": 1.4386,
      "theoretical_loss": 3.477683238298299,
      "tokens_seen": 1873739776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365496489468405,
      "loss": 1.3422,
      "theoretical_loss": 3.477673050388998,
      "tokens_seen": 1873805312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365295887662989,
      "loss": 1.332,
      "theoretical_loss": 3.4776628629357766,
      "tokens_seen": 1873870848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043650952858575727,
      "loss": 1.2139,
      "theoretical_loss": 3.4776526759385984,
      "tokens_seen": 1873936384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043648946840521563,
      "loss": 1.3377,
      "theoretical_loss": 3.477642489397428,
      "tokens_seen": 1874001920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436469408224674,
      "loss": 1.4418,
      "theoretical_loss": 3.477632303312227,
      "tokens_seen": 1874067456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364493480441324,
      "loss": 1.3544,
      "theoretical_loss": 3.4776221176829614,
      "tokens_seen": 1874132992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043642928786359077,
      "loss": 1.4147,
      "theoretical_loss": 3.477611932509593,
      "tokens_seen": 1874198528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043640922768304913,
      "loss": 1.4358,
      "theoretical_loss": 3.4776017477920864,
      "tokens_seen": 1874264064
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363891675025075,
      "loss": 1.3215,
      "theoretical_loss": 3.477591563530405,
      "tokens_seen": 1874329600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363691073219659,
      "loss": 1.4032,
      "theoretical_loss": 3.477581379724513,
      "tokens_seen": 1874395136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363490471414243,
      "loss": 1.4525,
      "theoretical_loss": 3.477571196374373,
      "tokens_seen": 1874460672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043632898696088264,
      "loss": 1.3883,
      "theoretical_loss": 3.4775610134799497,
      "tokens_seen": 1874526208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436308926780341,
      "loss": 1.2826,
      "theoretical_loss": 3.477550831041206,
      "tokens_seen": 1874591744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043628886659979936,
      "loss": 1.4635,
      "theoretical_loss": 3.4775406490581062,
      "tokens_seen": 1874657280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362688064192578,
      "loss": 1.4793,
      "theoretical_loss": 3.4775304675306136,
      "tokens_seen": 1874722816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043624874623871614,
      "loss": 1.3495,
      "theoretical_loss": 3.4775202864586916,
      "tokens_seen": 1874788352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362286860581745,
      "loss": 1.5164,
      "theoretical_loss": 3.477510105842305,
      "tokens_seen": 1874853888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043620862587763286,
      "loss": 1.3681,
      "theoretical_loss": 3.4774999256814163,
      "tokens_seen": 1874919424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361885656970913,
      "loss": 1.3664,
      "theoretical_loss": 3.4774897459759897,
      "tokens_seen": 1874984960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361685055165497,
      "loss": 1.322,
      "theoretical_loss": 3.477479566725989,
      "tokens_seen": 1875050496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043614844533600806,
      "loss": 1.3255,
      "theoretical_loss": 3.477469387931378,
      "tokens_seen": 1875116032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361283851554664,
      "loss": 1.485,
      "theoretical_loss": 3.4774592095921197,
      "tokens_seen": 1875181568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361083249749248,
      "loss": 1.3523,
      "theoretical_loss": 3.4774490317081783,
      "tokens_seen": 1875247104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360882647943832,
      "loss": 1.3344,
      "theoretical_loss": 3.477438854279518,
      "tokens_seen": 1875312640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043606820461384156,
      "loss": 1.3527,
      "theoretical_loss": 3.4774286773061016,
      "tokens_seen": 1875378176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360481444332999,
      "loss": 1.414,
      "theoretical_loss": 3.477418500787894,
      "tokens_seen": 1875443712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360280842527583,
      "loss": 1.3891,
      "theoretical_loss": 3.4774083247248573,
      "tokens_seen": 1875509248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360080240722167,
      "loss": 1.3888,
      "theoretical_loss": 3.477398149116957,
      "tokens_seen": 1875574784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043598796389167507,
      "loss": 1.3952,
      "theoretical_loss": 3.4773879739641553,
      "tokens_seen": 1875640320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043596790371113343,
      "loss": 1.3615,
      "theoretical_loss": 3.477377799266417,
      "tokens_seen": 1875705856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004359478435305918,
      "loss": 1.3103,
      "theoretical_loss": 3.4773676250237058,
      "tokens_seen": 1875771392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043592778335005015,
      "loss": 1.3074,
      "theoretical_loss": 3.477357451235984,
      "tokens_seen": 1875836928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043590772316950857,
      "loss": 1.4332,
      "theoretical_loss": 3.4773472779032177,
      "tokens_seen": 1875902464
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043588766298896693,
      "loss": 1.3871,
      "theoretical_loss": 3.477337105025369,
      "tokens_seen": 1875968000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358676028084253,
      "loss": 1.3467,
      "theoretical_loss": 3.477326932602402,
      "tokens_seen": 1876033536
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 684134,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3862347602844238,
      "objective/train/theoretical_loss": 3.477316760634281,
      "objective/train/tokens_used": 246558176,
      "theoretical_loss": 3.477316760634281,
      "tokens_seen": 1876099072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043584754262788366,
      "loss": 1.3679,
      "theoretical_loss": 3.477316760634281,
      "tokens_seen": 1876099072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043582748244734207,
      "loss": 1.4095,
      "theoretical_loss": 3.4773065891209693,
      "tokens_seen": 1876164608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043580742226680043,
      "loss": 1.3806,
      "theoretical_loss": 3.477296418062431,
      "tokens_seen": 1876230144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357873620862588,
      "loss": 1.401,
      "theoretical_loss": 3.477286247458629,
      "tokens_seen": 1876295680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043576730190571716,
      "loss": 1.3478,
      "theoretical_loss": 3.477276077309528,
      "tokens_seen": 1876361216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357472417251755,
      "loss": 1.4373,
      "theoretical_loss": 3.4772659076150916,
      "tokens_seen": 1876426752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043572718154463394,
      "loss": 1.3464,
      "theoretical_loss": 3.477255738375283,
      "tokens_seen": 1876492288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357071213640923,
      "loss": 1.4229,
      "theoretical_loss": 3.4772455695900675,
      "tokens_seen": 1876557824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043568706118355066,
      "loss": 1.2518,
      "theoretical_loss": 3.4772354012594073,
      "tokens_seen": 1876623360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000435667001003009,
      "loss": 1.4456,
      "theoretical_loss": 3.4772252333832667,
      "tokens_seen": 1876688896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043564694082246744,
      "loss": 1.3568,
      "theoretical_loss": 3.4772150659616097,
      "tokens_seen": 1876754432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356268806419258,
      "loss": 1.3992,
      "theoretical_loss": 3.4772048989944,
      "tokens_seen": 1876819968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043560682046138417,
      "loss": 1.4018,
      "theoretical_loss": 3.4771947324816024,
      "tokens_seen": 1876885504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043558676028084253,
      "loss": 1.4062,
      "theoretical_loss": 3.477184566423179,
      "tokens_seen": 1876951040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355667001003009,
      "loss": 1.4356,
      "theoretical_loss": 3.477174400819094,
      "tokens_seen": 1877016576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355466399197593,
      "loss": 1.4477,
      "theoretical_loss": 3.4771642356693127,
      "tokens_seen": 1877082112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043552657973921767,
      "loss": 1.4157,
      "theoretical_loss": 3.477154070973797,
      "tokens_seen": 1877147648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043550651955867603,
      "loss": 1.4087,
      "theoretical_loss": 3.477143906732512,
      "tokens_seen": 1877213184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354864593781344,
      "loss": 1.4263,
      "theoretical_loss": 3.477133742945421,
      "tokens_seen": 1877278720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354663991975928,
      "loss": 1.4682,
      "theoretical_loss": 3.4771235796124884,
      "tokens_seen": 1877344256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043544633901705117,
      "loss": 1.3881,
      "theoretical_loss": 3.4771134167336775,
      "tokens_seen": 1877409792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043542627883650953,
      "loss": 1.4303,
      "theoretical_loss": 3.477103254308952,
      "tokens_seen": 1877475328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354062186559679,
      "loss": 1.4407,
      "theoretical_loss": 3.4770930923382766,
      "tokens_seen": 1877540864
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353861584754263,
      "loss": 1.5102,
      "theoretical_loss": 3.4770829308216147,
      "tokens_seen": 1877606400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353660982948847,
      "loss": 1.4889,
      "theoretical_loss": 3.47707276975893,
      "tokens_seen": 1877671936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043534603811434304,
      "loss": 1.3806,
      "theoretical_loss": 3.4770626091501864,
      "tokens_seen": 1877737472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353259779338014,
      "loss": 1.4294,
      "theoretical_loss": 3.477052448995348,
      "tokens_seen": 1877803008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043530591775325976,
      "loss": 1.4653,
      "theoretical_loss": 3.4770422892943786,
      "tokens_seen": 1877868544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352858575727182,
      "loss": 1.381,
      "theoretical_loss": 3.4770321300472418,
      "tokens_seen": 1877934080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043526579739217654,
      "loss": 1.4214,
      "theoretical_loss": 3.477021971253902,
      "tokens_seen": 1877999616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352457372116349,
      "loss": 1.4503,
      "theoretical_loss": 3.4770118129143226,
      "tokens_seen": 1878065152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043522567703109326,
      "loss": 1.4693,
      "theoretical_loss": 3.4770016550284684,
      "tokens_seen": 1878130688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352056168505517,
      "loss": 1.3785,
      "theoretical_loss": 3.476991497596302,
      "tokens_seen": 1878196224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043518555667001004,
      "loss": 1.5234,
      "theoretical_loss": 3.476981340617788,
      "tokens_seen": 1878261760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004351654964894684,
      "loss": 1.2875,
      "theoretical_loss": 3.47697118409289,
      "tokens_seen": 1878327296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043514543630892677,
      "loss": 1.3326,
      "theoretical_loss": 3.476961028021573,
      "tokens_seen": 1878392832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043512537612838513,
      "loss": 1.3927,
      "theoretical_loss": 3.4769508724038,
      "tokens_seen": 1878458368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043510531594784355,
      "loss": 1.4017,
      "theoretical_loss": 3.4769407172395344,
      "tokens_seen": 1878523904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350852557673019,
      "loss": 1.4232,
      "theoretical_loss": 3.476930562528741,
      "tokens_seen": 1878589440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043506519558676027,
      "loss": 1.3257,
      "theoretical_loss": 3.4769204082713836,
      "tokens_seen": 1878654976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043504513540621863,
      "loss": 1.3627,
      "theoretical_loss": 3.476910254467426,
      "tokens_seen": 1878720512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043502507522567705,
      "loss": 1.3708,
      "theoretical_loss": 3.476900101116832,
      "tokens_seen": 1878786048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350050150451354,
      "loss": 1.2381,
      "theoretical_loss": 3.476889948219566,
      "tokens_seen": 1878851584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349849548645938,
      "loss": 1.2966,
      "theoretical_loss": 3.4768797957755915,
      "tokens_seen": 1878917120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043496489468405214,
      "loss": 1.4309,
      "theoretical_loss": 3.4768696437848723,
      "tokens_seen": 1878982656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349448345035105,
      "loss": 1.417,
      "theoretical_loss": 3.4768594922473732,
      "tokens_seen": 1879048192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349247743229689,
      "loss": 1.4358,
      "theoretical_loss": 3.4768493411630574,
      "tokens_seen": 1879113728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349047141424273,
      "loss": 1.3979,
      "theoretical_loss": 3.476839190531889,
      "tokens_seen": 1879179264
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043488465396188564,
      "loss": 1.3144,
      "theoretical_loss": 3.476829040353832,
      "tokens_seen": 1879244800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434864593781344,
      "loss": 1.311,
      "theoretical_loss": 3.4768188906288504,
      "tokens_seen": 1879310336
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 685672,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.300635576248169,
      "objective/train/theoretical_loss": 3.4768087413569084,
      "objective/train/tokens_used": 249834976,
      "theoretical_loss": 3.4768087413569084,
      "tokens_seen": 1879375872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348445336008024,
      "loss": 1.3267,
      "theoretical_loss": 3.4768087413569084,
      "tokens_seen": 1879375872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348244734202608,
      "loss": 1.5607,
      "theoretical_loss": 3.4767985925379703,
      "tokens_seen": 1879441408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043480441323971914,
      "loss": 1.2839,
      "theoretical_loss": 3.4767884441719987,
      "tokens_seen": 1879506944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347843530591775,
      "loss": 1.2734,
      "theoretical_loss": 3.4767782962589586,
      "tokens_seen": 1879572480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043476429287863587,
      "loss": 1.3421,
      "theoretical_loss": 3.476768148798814,
      "tokens_seen": 1879638016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347442326980943,
      "loss": 1.517,
      "theoretical_loss": 3.476758001791529,
      "tokens_seen": 1879703552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043472417251755265,
      "loss": 1.374,
      "theoretical_loss": 3.4767478552370674,
      "tokens_seen": 1879769088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434704112337011,
      "loss": 1.2775,
      "theoretical_loss": 3.4767377091353926,
      "tokens_seen": 1879834624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346840521564694,
      "loss": 1.416,
      "theoretical_loss": 3.4767275634864694,
      "tokens_seen": 1879900160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043466399197592784,
      "loss": 1.3652,
      "theoretical_loss": 3.476717418290262,
      "tokens_seen": 1879965696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346439317953862,
      "loss": 1.3781,
      "theoretical_loss": 3.476707273546733,
      "tokens_seen": 1880031232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043462387161484457,
      "loss": 1.3308,
      "theoretical_loss": 3.4766971292558484,
      "tokens_seen": 1880096768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043460381143430293,
      "loss": 1.4244,
      "theoretical_loss": 3.476686985417571,
      "tokens_seen": 1880162304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345837512537613,
      "loss": 1.4142,
      "theoretical_loss": 3.476676842031865,
      "tokens_seen": 1880227840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345636910732197,
      "loss": 1.3023,
      "theoretical_loss": 3.4766666990986947,
      "tokens_seen": 1880293376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043454363089267807,
      "loss": 1.3298,
      "theoretical_loss": 3.476656556618024,
      "tokens_seen": 1880358912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043452357071213643,
      "loss": 1.3911,
      "theoretical_loss": 3.4766464145898164,
      "tokens_seen": 1880424448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345035105315948,
      "loss": 1.4692,
      "theoretical_loss": 3.4766362730140368,
      "tokens_seen": 1880489984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344834503510532,
      "loss": 1.3192,
      "theoretical_loss": 3.476626131890649,
      "tokens_seen": 1880555520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043446339017051157,
      "loss": 1.291,
      "theoretical_loss": 3.476615991219617,
      "tokens_seen": 1880621056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043444332998996993,
      "loss": 1.37,
      "theoretical_loss": 3.4766058510009046,
      "tokens_seen": 1880686592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344232698094283,
      "loss": 1.3437,
      "theoretical_loss": 3.476595711234476,
      "tokens_seen": 1880752128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344032096288867,
      "loss": 1.3999,
      "theoretical_loss": 3.476585571920295,
      "tokens_seen": 1880817664
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343831494483451,
      "loss": 1.4353,
      "theoretical_loss": 3.476575433058327,
      "tokens_seen": 1880883200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043436308926780344,
      "loss": 1.2927,
      "theoretical_loss": 3.476565294648535,
      "tokens_seen": 1880948736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343430290872618,
      "loss": 1.4703,
      "theoretical_loss": 3.4765551566908823,
      "tokens_seen": 1881014272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043432296890672016,
      "loss": 1.3755,
      "theoretical_loss": 3.476545019185334,
      "tokens_seen": 1881079808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343029087261786,
      "loss": 1.328,
      "theoretical_loss": 3.4765348821318547,
      "tokens_seen": 1881145344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043428284854563694,
      "loss": 1.4415,
      "theoretical_loss": 3.4765247455304076,
      "tokens_seen": 1881210880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342627883650953,
      "loss": 1.2975,
      "theoretical_loss": 3.476514609380957,
      "tokens_seen": 1881276416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043424272818455367,
      "loss": 1.3733,
      "theoretical_loss": 3.476504473683467,
      "tokens_seen": 1881341952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342226680040121,
      "loss": 1.4108,
      "theoretical_loss": 3.4764943384379015,
      "tokens_seen": 1881407488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043420260782347044,
      "loss": 1.3032,
      "theoretical_loss": 3.476484203644225,
      "tokens_seen": 1881473024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341825476429288,
      "loss": 1.3831,
      "theoretical_loss": 3.4764740693024017,
      "tokens_seen": 1881538560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043416248746238717,
      "loss": 1.3583,
      "theoretical_loss": 3.4764639354123954,
      "tokens_seen": 1881604096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043414242728184553,
      "loss": 1.2341,
      "theoretical_loss": 3.47645380197417,
      "tokens_seen": 1881669632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043412236710130395,
      "loss": 1.376,
      "theoretical_loss": 3.4764436689876903,
      "tokens_seen": 1881735168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341023069207623,
      "loss": 1.3477,
      "theoretical_loss": 3.47643353645292,
      "tokens_seen": 1881800704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043408224674022067,
      "loss": 1.485,
      "theoretical_loss": 3.476423404369823,
      "tokens_seen": 1881866240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043406218655967903,
      "loss": 1.3975,
      "theoretical_loss": 3.4764132727383643,
      "tokens_seen": 1881931776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043404212637913745,
      "loss": 1.3323,
      "theoretical_loss": 3.4764031415585066,
      "tokens_seen": 1881997312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004340220661985958,
      "loss": 1.3039,
      "theoretical_loss": 3.4763930108302157,
      "tokens_seen": 1882062848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004340020060180542,
      "loss": 1.438,
      "theoretical_loss": 3.4763828805534547,
      "tokens_seen": 1882128384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043398194583751254,
      "loss": 1.446,
      "theoretical_loss": 3.4763727507281876,
      "tokens_seen": 1882193920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339618856569709,
      "loss": 1.5378,
      "theoretical_loss": 3.4763626213543795,
      "tokens_seen": 1882259456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339418254764293,
      "loss": 1.3626,
      "theoretical_loss": 3.476352492431994,
      "tokens_seen": 1882324992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339217652958877,
      "loss": 1.3032,
      "theoretical_loss": 3.4763423639609954,
      "tokens_seen": 1882390528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043390170511534604,
      "loss": 1.2521,
      "theoretical_loss": 3.4763322359413475,
      "tokens_seen": 1882456064
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338816449348044,
      "loss": 1.3163,
      "theoretical_loss": 3.4763221083730147,
      "tokens_seen": 1882521600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338615847542628,
      "loss": 1.4895,
      "theoretical_loss": 3.4763119812559613,
      "tokens_seen": 1882587136
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 686423,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5192041397094727,
      "objective/train/theoretical_loss": 3.4763018545901514,
      "objective/train/tokens_used": 253111776,
      "theoretical_loss": 3.4763018545901514,
      "tokens_seen": 1882652672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338415245737212,
      "loss": 1.4678,
      "theoretical_loss": 3.4763018545901514,
      "tokens_seen": 1882652672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043382146439317954,
      "loss": 1.3997,
      "theoretical_loss": 3.476291728375549,
      "tokens_seen": 1882718208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338014042126379,
      "loss": 1.3694,
      "theoretical_loss": 3.476281602612119,
      "tokens_seen": 1882783744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043378134403209627,
      "loss": 1.4464,
      "theoretical_loss": 3.476271477299825,
      "tokens_seen": 1882849280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337612838515547,
      "loss": 1.4271,
      "theoretical_loss": 3.476261352438631,
      "tokens_seen": 1882914816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043374122367101305,
      "loss": 1.4207,
      "theoretical_loss": 3.4762512280285014,
      "tokens_seen": 1882980352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337211634904714,
      "loss": 1.3115,
      "theoretical_loss": 3.476241104069401,
      "tokens_seen": 1883045888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043370110330992977,
      "loss": 1.39,
      "theoretical_loss": 3.4762309805612928,
      "tokens_seen": 1883111424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336810431293882,
      "loss": 1.4431,
      "theoretical_loss": 3.4762208575041424,
      "tokens_seen": 1883176960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043366098294884655,
      "loss": 1.4802,
      "theoretical_loss": 3.476210734897913,
      "tokens_seen": 1883242496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336409227683049,
      "loss": 1.4396,
      "theoretical_loss": 3.476200612742569,
      "tokens_seen": 1883308032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336208625877633,
      "loss": 1.4059,
      "theoretical_loss": 3.476190491038075,
      "tokens_seen": 1883373568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043360080240722164,
      "loss": 1.3619,
      "theoretical_loss": 3.476180369784395,
      "tokens_seen": 1883439104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043358074222668005,
      "loss": 1.4411,
      "theoretical_loss": 3.476170248981493,
      "tokens_seen": 1883504640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335606820461384,
      "loss": 1.4295,
      "theoretical_loss": 3.476160128629334,
      "tokens_seen": 1883570176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335406218655968,
      "loss": 1.3972,
      "theoretical_loss": 3.476150008727881,
      "tokens_seen": 1883635712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043352056168505514,
      "loss": 1.3767,
      "theoretical_loss": 3.4761398892771,
      "tokens_seen": 1883701248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043350050150451356,
      "loss": 1.2904,
      "theoretical_loss": 3.4761297702769536,
      "tokens_seen": 1883766784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334804413239719,
      "loss": 1.3371,
      "theoretical_loss": 3.476119651727407,
      "tokens_seen": 1883832320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334603811434303,
      "loss": 1.3201,
      "theoretical_loss": 3.476109533628424,
      "tokens_seen": 1883897856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043344032096288864,
      "loss": 1.4778,
      "theoretical_loss": 3.4760994159799687,
      "tokens_seen": 1883963392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433420260782347,
      "loss": 1.2988,
      "theoretical_loss": 3.4760892987820062,
      "tokens_seen": 1884028928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334002006018054,
      "loss": 1.2729,
      "theoretical_loss": 3.4760791820345,
      "tokens_seen": 1884094464
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333801404212638,
      "loss": 1.3424,
      "theoretical_loss": 3.4760690657374145,
      "tokens_seen": 1884160000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043336008024072215,
      "loss": 1.3418,
      "theoretical_loss": 3.476058949890714,
      "tokens_seen": 1884225536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333400200601805,
      "loss": 1.4191,
      "theoretical_loss": 3.476048834494363,
      "tokens_seen": 1884291072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333199598796389,
      "loss": 1.3216,
      "theoretical_loss": 3.4760387195483258,
      "tokens_seen": 1884356608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332998996990973,
      "loss": 1.381,
      "theoretical_loss": 3.4760286050525666,
      "tokens_seen": 1884422144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043327983951855565,
      "loss": 1.3322,
      "theoretical_loss": 3.47601849100705,
      "tokens_seen": 1884487680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433259779338014,
      "loss": 1.4154,
      "theoretical_loss": 3.4760083774117394,
      "tokens_seen": 1884553216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043323971915747243,
      "loss": 1.2989,
      "theoretical_loss": 3.4759982642666003,
      "tokens_seen": 1884618752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332196589769308,
      "loss": 1.3601,
      "theoretical_loss": 3.475988151571596,
      "tokens_seen": 1884684288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043319959879638915,
      "loss": 1.2676,
      "theoretical_loss": 3.475978039326691,
      "tokens_seen": 1884749824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043317953861584757,
      "loss": 1.554,
      "theoretical_loss": 3.4759679275318502,
      "tokens_seen": 1884815360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043315947843530593,
      "loss": 1.3906,
      "theoretical_loss": 3.4759578161870373,
      "tokens_seen": 1884880896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043313941825476435,
      "loss": 1.3675,
      "theoretical_loss": 3.475947705292217,
      "tokens_seen": 1884946432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331193580742227,
      "loss": 1.3803,
      "theoretical_loss": 3.4759375948473537,
      "tokens_seen": 1885011968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043309929789368107,
      "loss": 1.2959,
      "theoretical_loss": 3.475927484852411,
      "tokens_seen": 1885077504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043307923771313943,
      "loss": 1.3708,
      "theoretical_loss": 3.475917375307354,
      "tokens_seen": 1885143040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043305917753259785,
      "loss": 1.3631,
      "theoretical_loss": 3.475907266212147,
      "tokens_seen": 1885208576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330391173520562,
      "loss": 1.3508,
      "theoretical_loss": 3.475897157566754,
      "tokens_seen": 1885274112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330190571715146,
      "loss": 1.3081,
      "theoretical_loss": 3.4758870493711393,
      "tokens_seen": 1885339648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043299899699097294,
      "loss": 1.4033,
      "theoretical_loss": 3.475876941625268,
      "tokens_seen": 1885405184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329789368104313,
      "loss": 1.4372,
      "theoretical_loss": 3.4758668343291035,
      "tokens_seen": 1885470720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329588766298897,
      "loss": 1.3936,
      "theoretical_loss": 3.4758567274826104,
      "tokens_seen": 1885536256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329388164493481,
      "loss": 1.4228,
      "theoretical_loss": 3.475846621085754,
      "tokens_seen": 1885601792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043291875626880644,
      "loss": 1.4419,
      "theoretical_loss": 3.4758365151384973,
      "tokens_seen": 1885667328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328986960882648,
      "loss": 1.3192,
      "theoretical_loss": 3.4758264096408054,
      "tokens_seen": 1885732864
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328786359077232,
      "loss": 1.448,
      "theoretical_loss": 3.4758163045926427,
      "tokens_seen": 1885798400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328585757271816,
      "loss": 1.4919,
      "theoretical_loss": 3.475806199993973,
      "tokens_seen": 1885863936
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 687569,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3545880317687988,
      "objective/train/theoretical_loss": 3.475796095844762,
      "objective/train/tokens_used": 256388576,
      "theoretical_loss": 3.475796095844762,
      "tokens_seen": 1885929472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043283851554663994,
      "loss": 1.3714,
      "theoretical_loss": 3.475796095844762,
      "tokens_seen": 1885929472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328184553660983,
      "loss": 1.4924,
      "theoretical_loss": 3.4757859921449725,
      "tokens_seen": 1885995008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043279839518555667,
      "loss": 1.4168,
      "theoretical_loss": 3.47577588889457,
      "tokens_seen": 1886060544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327783350050151,
      "loss": 1.3255,
      "theoretical_loss": 3.4757657860935183,
      "tokens_seen": 1886126080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043275827482447345,
      "loss": 1.4571,
      "theoretical_loss": 3.475755683741782,
      "tokens_seen": 1886191616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327382146439318,
      "loss": 1.3323,
      "theoretical_loss": 3.4757455818393255,
      "tokens_seen": 1886257152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043271815446339017,
      "loss": 1.3504,
      "theoretical_loss": 3.4757354803861134,
      "tokens_seen": 1886322688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326980942828486,
      "loss": 1.4917,
      "theoretical_loss": 3.4757253793821095,
      "tokens_seen": 1886388224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043267803410230695,
      "loss": 1.4811,
      "theoretical_loss": 3.475715278827279,
      "tokens_seen": 1886453760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326579739217653,
      "loss": 1.3935,
      "theoretical_loss": 3.4757051787215865,
      "tokens_seen": 1886519296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326379137412237,
      "loss": 1.595,
      "theoretical_loss": 3.4756950790649954,
      "tokens_seen": 1886584832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043261785356068204,
      "loss": 1.3704,
      "theoretical_loss": 3.4756849798574705,
      "tokens_seen": 1886650368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043259779338014045,
      "loss": 1.3238,
      "theoretical_loss": 3.4756748810989766,
      "tokens_seen": 1886715904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325777331995988,
      "loss": 1.2713,
      "theoretical_loss": 3.4756647827894778,
      "tokens_seen": 1886781440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325576730190572,
      "loss": 1.4402,
      "theoretical_loss": 3.475654684928939,
      "tokens_seen": 1886846976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043253761283851554,
      "loss": 1.5006,
      "theoretical_loss": 3.475644587517324,
      "tokens_seen": 1886912512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043251755265797396,
      "loss": 1.2872,
      "theoretical_loss": 3.4756344905545977,
      "tokens_seen": 1886978048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324974924774323,
      "loss": 1.5122,
      "theoretical_loss": 3.4756243940407243,
      "tokens_seen": 1887043584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324774322968907,
      "loss": 1.3542,
      "theoretical_loss": 3.4756142979756683,
      "tokens_seen": 1887109120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043245737211634904,
      "loss": 1.3311,
      "theoretical_loss": 3.4756042023593947,
      "tokens_seen": 1887174656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324373119358074,
      "loss": 1.4956,
      "theoretical_loss": 3.475594107191867,
      "tokens_seen": 1887240192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324172517552658,
      "loss": 1.4333,
      "theoretical_loss": 3.4755840124730506,
      "tokens_seen": 1887305728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323971915747242,
      "loss": 1.3688,
      "theoretical_loss": 3.475573918202909,
      "tokens_seen": 1887371264
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043237713139418255,
      "loss": 1.3313,
      "theoretical_loss": 3.4755638243814078,
      "tokens_seen": 1887436800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323570712136409,
      "loss": 1.3202,
      "theoretical_loss": 3.475553731008511,
      "tokens_seen": 1887502336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323370110330993,
      "loss": 1.4015,
      "theoretical_loss": 3.475543638084183,
      "tokens_seen": 1887567872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323169508525577,
      "loss": 1.4399,
      "theoretical_loss": 3.4755335456083873,
      "tokens_seen": 1887633408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043229689067201605,
      "loss": 1.3367,
      "theoretical_loss": 3.4755234535810904,
      "tokens_seen": 1887698944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322768304914744,
      "loss": 1.3215,
      "theoretical_loss": 3.4755133620022556,
      "tokens_seen": 1887764480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043225677031093283,
      "loss": 1.3958,
      "theoretical_loss": 3.4755032708718474,
      "tokens_seen": 1887830016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322367101303912,
      "loss": 1.4355,
      "theoretical_loss": 3.4754931801898308,
      "tokens_seen": 1887895552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043221664994984955,
      "loss": 1.3877,
      "theoretical_loss": 3.4754830899561697,
      "tokens_seen": 1887961088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321965897693079,
      "loss": 1.2685,
      "theoretical_loss": 3.4754730001708287,
      "tokens_seen": 1888026624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321765295887663,
      "loss": 1.476,
      "theoretical_loss": 3.475462910833773,
      "tokens_seen": 1888092160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321564694082247,
      "loss": 1.3612,
      "theoretical_loss": 3.475452821944967,
      "tokens_seen": 1888157696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043213640922768306,
      "loss": 1.2024,
      "theoretical_loss": 3.4754427335043743,
      "tokens_seen": 1888223232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321163490471414,
      "loss": 1.3737,
      "theoretical_loss": 3.47543264551196,
      "tokens_seen": 1888288768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320962888665998,
      "loss": 1.2892,
      "theoretical_loss": 3.475422557967689,
      "tokens_seen": 1888354304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320762286860582,
      "loss": 1.4198,
      "theoretical_loss": 3.4754124708715257,
      "tokens_seen": 1888419840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043205616850551656,
      "loss": 1.338,
      "theoretical_loss": 3.4754023842234343,
      "tokens_seen": 1888485376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320361083249749,
      "loss": 1.325,
      "theoretical_loss": 3.4753922980233796,
      "tokens_seen": 1888550912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320160481444333,
      "loss": 1.442,
      "theoretical_loss": 3.4753822122713256,
      "tokens_seen": 1888616448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043199598796389165,
      "loss": 1.4417,
      "theoretical_loss": 3.4753721269672373,
      "tokens_seen": 1888681984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043197592778335006,
      "loss": 1.3321,
      "theoretical_loss": 3.4753620421110796,
      "tokens_seen": 1888747520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319558676028084,
      "loss": 1.2704,
      "theoretical_loss": 3.4753519577028165,
      "tokens_seen": 1888813056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319358074222668,
      "loss": 1.354,
      "theoretical_loss": 3.475341873742413,
      "tokens_seen": 1888878592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043191574724172515,
      "loss": 1.4117,
      "theoretical_loss": 3.4753317902298337,
      "tokens_seen": 1888944128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043189568706118357,
      "loss": 1.3963,
      "theoretical_loss": 3.4753217071650426,
      "tokens_seen": 1889009664
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043187562688064193,
      "loss": 1.4011,
      "theoretical_loss": 3.475311624548005,
      "tokens_seen": 1889075200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318555667001003,
      "loss": 1.4191,
      "theoretical_loss": 3.4753015423786846,
      "tokens_seen": 1889140736
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 689100,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.424290418624878,
      "objective/train/theoretical_loss": 3.475291460657047,
      "objective/train/tokens_used": 259665376,
      "theoretical_loss": 3.475291460657047,
      "tokens_seen": 1889206272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043183550651955865,
      "loss": 1.3274,
      "theoretical_loss": 3.475291460657047,
      "tokens_seen": 1889206272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000431815446339017,
      "loss": 1.3744,
      "theoretical_loss": 3.475281379383056,
      "tokens_seen": 1889271808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043179538615847543,
      "loss": 1.3243,
      "theoretical_loss": 3.4752712985566765,
      "tokens_seen": 1889337344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317753259779338,
      "loss": 1.2941,
      "theoretical_loss": 3.475261218177873,
      "tokens_seen": 1889402880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043175526579739215,
      "loss": 1.3798,
      "theoretical_loss": 3.4752511382466107,
      "tokens_seen": 1889468416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317352056168505,
      "loss": 1.2578,
      "theoretical_loss": 3.4752410587628533,
      "tokens_seen": 1889533952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043171514543630893,
      "loss": 1.3794,
      "theoretical_loss": 3.475230979726566,
      "tokens_seen": 1889599488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316950852557673,
      "loss": 1.4381,
      "theoretical_loss": 3.475220901137713,
      "tokens_seen": 1889665024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316750250752257,
      "loss": 1.2917,
      "theoretical_loss": 3.4752108229962593,
      "tokens_seen": 1889730560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316549648946841,
      "loss": 1.3257,
      "theoretical_loss": 3.4752007453021694,
      "tokens_seen": 1889796096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043163490471414244,
      "loss": 1.3897,
      "theoretical_loss": 3.475190668055408,
      "tokens_seen": 1889861632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043161484453360085,
      "loss": 1.2737,
      "theoretical_loss": 3.4751805912559393,
      "tokens_seen": 1889927168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315947843530592,
      "loss": 1.415,
      "theoretical_loss": 3.475170514903729,
      "tokens_seen": 1889992704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315747241725176,
      "loss": 1.3807,
      "theoretical_loss": 3.4751604389987403,
      "tokens_seen": 1890058240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043155466399197594,
      "loss": 1.4118,
      "theoretical_loss": 3.475150363540939,
      "tokens_seen": 1890123776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043153460381143436,
      "loss": 1.3708,
      "theoretical_loss": 3.475140288530289,
      "tokens_seen": 1890189312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315145436308927,
      "loss": 1.305,
      "theoretical_loss": 3.4751302139667555,
      "tokens_seen": 1890254848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314944834503511,
      "loss": 1.4094,
      "theoretical_loss": 3.4751201398503024,
      "tokens_seen": 1890320384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043147442326980944,
      "loss": 1.3743,
      "theoretical_loss": 3.4751100661808954,
      "tokens_seen": 1890385920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314543630892678,
      "loss": 1.4374,
      "theoretical_loss": 3.4750999929584987,
      "tokens_seen": 1890451456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314343029087262,
      "loss": 1.4821,
      "theoretical_loss": 3.475089920183077,
      "tokens_seen": 1890516992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314142427281846,
      "loss": 1.3348,
      "theoretical_loss": 3.4750798478545946,
      "tokens_seen": 1890582528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043139418254764295,
      "loss": 1.336,
      "theoretical_loss": 3.4750697759730165,
      "tokens_seen": 1890648064
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313741223671013,
      "loss": 1.3409,
      "theoretical_loss": 3.4750597045383076,
      "tokens_seen": 1890713600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313540621865597,
      "loss": 1.3092,
      "theoretical_loss": 3.475049633550432,
      "tokens_seen": 1890779136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313340020060181,
      "loss": 1.3651,
      "theoretical_loss": 3.475039563009355,
      "tokens_seen": 1890844672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043131394182547645,
      "loss": 1.3418,
      "theoretical_loss": 3.4750294929150405,
      "tokens_seen": 1890910208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312938816449348,
      "loss": 1.3331,
      "theoretical_loss": 3.4750194232674545,
      "tokens_seen": 1890975744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043127382146439323,
      "loss": 1.4589,
      "theoretical_loss": 3.4750093540665605,
      "tokens_seen": 1891041280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312537612838516,
      "loss": 1.3404,
      "theoretical_loss": 3.4749992853123235,
      "tokens_seen": 1891106816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043123370110330995,
      "loss": 1.2945,
      "theoretical_loss": 3.4749892170047083,
      "tokens_seen": 1891172352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312136409227683,
      "loss": 1.4336,
      "theoretical_loss": 3.47497914914368,
      "tokens_seen": 1891237888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311935807422267,
      "loss": 1.2902,
      "theoretical_loss": 3.474969081729203,
      "tokens_seen": 1891303424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311735205616851,
      "loss": 1.3762,
      "theoretical_loss": 3.4749590147612417,
      "tokens_seen": 1891368960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043115346038114346,
      "loss": 1.4193,
      "theoretical_loss": 3.4749489482397613,
      "tokens_seen": 1891434496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311334002006018,
      "loss": 1.4225,
      "theoretical_loss": 3.474938882164726,
      "tokens_seen": 1891500032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311133400200602,
      "loss": 1.3227,
      "theoretical_loss": 3.474928816536101,
      "tokens_seen": 1891565568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310932798395186,
      "loss": 1.3073,
      "theoretical_loss": 3.4749187513538513,
      "tokens_seen": 1891631104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043107321965897696,
      "loss": 1.2453,
      "theoretical_loss": 3.4749086866179413,
      "tokens_seen": 1891696640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310531594784353,
      "loss": 1.2515,
      "theoretical_loss": 3.474898622328335,
      "tokens_seen": 1891762176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310330992978937,
      "loss": 1.4127,
      "theoretical_loss": 3.4748885584849987,
      "tokens_seen": 1891827712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043101303911735205,
      "loss": 1.2902,
      "theoretical_loss": 3.4748784950878955,
      "tokens_seen": 1891893248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043099297893681046,
      "loss": 1.4339,
      "theoretical_loss": 3.474868432136991,
      "tokens_seen": 1891958784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004309729187562688,
      "loss": 1.294,
      "theoretical_loss": 3.4748583696322504,
      "tokens_seen": 1892024320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004309528585757272,
      "loss": 1.3107,
      "theoretical_loss": 3.474848307573638,
      "tokens_seen": 1892089856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043093279839518555,
      "loss": 1.4287,
      "theoretical_loss": 3.4748382459611182,
      "tokens_seen": 1892155392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043091273821464397,
      "loss": 1.2769,
      "theoretical_loss": 3.474828184794656,
      "tokens_seen": 1892220928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043089267803410233,
      "loss": 1.3889,
      "theoretical_loss": 3.4748181240742166,
      "tokens_seen": 1892286464
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308726178535607,
      "loss": 1.3995,
      "theoretical_loss": 3.474808063799764,
      "tokens_seen": 1892352000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043085255767301905,
      "loss": 1.3339,
      "theoretical_loss": 3.474798003971264,
      "tokens_seen": 1892417536
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 690244,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.412577509880066,
      "objective/train/theoretical_loss": 3.474787944588681,
      "objective/train/tokens_used": 262942176,
      "theoretical_loss": 3.474787944588681,
      "tokens_seen": 1892483072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308324974924774,
      "loss": 1.4505,
      "theoretical_loss": 3.474787944588681,
      "tokens_seen": 1892483072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043081243731193583,
      "loss": 1.3993,
      "theoretical_loss": 3.4747778856519793,
      "tokens_seen": 1892548608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307923771313942,
      "loss": 1.4911,
      "theoretical_loss": 3.4747678271611244,
      "tokens_seen": 1892614144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043077231695085256,
      "loss": 1.2409,
      "theoretical_loss": 3.4747577691160805,
      "tokens_seen": 1892679680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307522567703109,
      "loss": 1.4645,
      "theoretical_loss": 3.4747477115168124,
      "tokens_seen": 1892745216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043073219658976933,
      "loss": 1.4092,
      "theoretical_loss": 3.4747376543632855,
      "tokens_seen": 1892810752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307121364092277,
      "loss": 1.306,
      "theoretical_loss": 3.474727597655464,
      "tokens_seen": 1892876288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043069207622868606,
      "loss": 1.3555,
      "theoretical_loss": 3.4747175413933133,
      "tokens_seen": 1892941824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306720160481444,
      "loss": 1.3298,
      "theoretical_loss": 3.4747074855767974,
      "tokens_seen": 1893007360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306519558676028,
      "loss": 1.398,
      "theoretical_loss": 3.4746974302058824,
      "tokens_seen": 1893072896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306318956870612,
      "loss": 1.3687,
      "theoretical_loss": 3.4746873752805314,
      "tokens_seen": 1893138432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043061183550651956,
      "loss": 1.4131,
      "theoretical_loss": 3.474677320800711,
      "tokens_seen": 1893203968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305917753259779,
      "loss": 1.3469,
      "theoretical_loss": 3.474667266766385,
      "tokens_seen": 1893269504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305717151454363,
      "loss": 1.2067,
      "theoretical_loss": 3.4746572131775184,
      "tokens_seen": 1893335040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305516549648947,
      "loss": 1.4596,
      "theoretical_loss": 3.4746471600340763,
      "tokens_seen": 1893400576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043053159478435306,
      "loss": 1.3052,
      "theoretical_loss": 3.474637107336023,
      "tokens_seen": 1893466112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305115346038114,
      "loss": 1.3396,
      "theoretical_loss": 3.474627055083324,
      "tokens_seen": 1893531648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304914744232698,
      "loss": 1.4196,
      "theoretical_loss": 3.4746170032759442,
      "tokens_seen": 1893597184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043047141424272815,
      "loss": 1.4317,
      "theoretical_loss": 3.474606951913848,
      "tokens_seen": 1893662720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043045135406218657,
      "loss": 1.4878,
      "theoretical_loss": 3.4745969009969997,
      "tokens_seen": 1893728256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043043129388164493,
      "loss": 1.3817,
      "theoretical_loss": 3.4745868505253656,
      "tokens_seen": 1893793792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304112337011033,
      "loss": 1.2922,
      "theoretical_loss": 3.4745768004989097,
      "tokens_seen": 1893859328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043039117352056165,
      "loss": 1.3906,
      "theoretical_loss": 3.474566750917597,
      "tokens_seen": 1893924864
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043037111334002007,
      "loss": 1.3697,
      "theoretical_loss": 3.474556701781392,
      "tokens_seen": 1893990400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043035105315947843,
      "loss": 1.3732,
      "theoretical_loss": 3.4745466530902607,
      "tokens_seen": 1894055936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004303309929789368,
      "loss": 1.474,
      "theoretical_loss": 3.474536604844167,
      "tokens_seen": 1894121472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043031093279839516,
      "loss": 1.3413,
      "theoretical_loss": 3.4745265570430757,
      "tokens_seen": 1894187008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302908726178535,
      "loss": 1.4142,
      "theoretical_loss": 3.4745165096869526,
      "tokens_seen": 1894252544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043027081243731194,
      "loss": 1.3226,
      "theoretical_loss": 3.4745064627757616,
      "tokens_seen": 1894318080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302507522567703,
      "loss": 1.3723,
      "theoretical_loss": 3.4744964163094685,
      "tokens_seen": 1894383616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043023069207622866,
      "loss": 1.3663,
      "theoretical_loss": 3.4744863702880378,
      "tokens_seen": 1894449152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430210631895687,
      "loss": 1.4078,
      "theoretical_loss": 3.474476324711434,
      "tokens_seen": 1894514688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043019057171514544,
      "loss": 1.3855,
      "theoretical_loss": 3.474466279579623,
      "tokens_seen": 1894580224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043017051153460386,
      "loss": 1.3748,
      "theoretical_loss": 3.4744562348925685,
      "tokens_seen": 1894645760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301504513540622,
      "loss": 1.3102,
      "theoretical_loss": 3.4744461906502364,
      "tokens_seen": 1894711296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301303911735206,
      "loss": 1.3159,
      "theoretical_loss": 3.4744361468525913,
      "tokens_seen": 1894776832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430110330992979,
      "loss": 1.3533,
      "theoretical_loss": 3.474426103499598,
      "tokens_seen": 1894842368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043009027081243736,
      "loss": 1.4138,
      "theoretical_loss": 3.4744160605912215,
      "tokens_seen": 1894907904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300702106318957,
      "loss": 1.364,
      "theoretical_loss": 3.4744060181274268,
      "tokens_seen": 1894973440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300501504513541,
      "loss": 1.4546,
      "theoretical_loss": 3.474395976108179,
      "tokens_seen": 1895038976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043003009027081245,
      "loss": 1.3582,
      "theoretical_loss": 3.474385934533443,
      "tokens_seen": 1895104512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043001003009027086,
      "loss": 1.3999,
      "theoretical_loss": 3.4743758934031836,
      "tokens_seen": 1895170048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004299899699097292,
      "loss": 1.4197,
      "theoretical_loss": 3.4743658527173658,
      "tokens_seen": 1895235584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004299699097291876,
      "loss": 1.3912,
      "theoretical_loss": 3.474355812475954,
      "tokens_seen": 1895301120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042994984954864595,
      "loss": 1.3552,
      "theoretical_loss": 3.4743457726789146,
      "tokens_seen": 1895366656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042992978936810437,
      "loss": 1.4356,
      "theoretical_loss": 3.4743357333262113,
      "tokens_seen": 1895432192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042990972918756273,
      "loss": 1.4076,
      "theoretical_loss": 3.4743256944178094,
      "tokens_seen": 1895497728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298896690070211,
      "loss": 1.5283,
      "theoretical_loss": 3.4743156559536743,
      "tokens_seen": 1895563264
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042986960882647945,
      "loss": 1.3122,
      "theoretical_loss": 3.47430561793377,
      "tokens_seen": 1895628800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298495486459378,
      "loss": 1.3786,
      "theoretical_loss": 3.4742955803580626,
      "tokens_seen": 1895694336
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 691404,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3427211046218872,
      "objective/train/theoretical_loss": 3.4742855432265163,
      "objective/train/tokens_used": 266218976,
      "theoretical_loss": 3.4742855432265163,
      "tokens_seen": 1895759872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042982948846539623,
      "loss": 1.3634,
      "theoretical_loss": 3.4742855432265163,
      "tokens_seen": 1895759872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298094282848546,
      "loss": 1.3741,
      "theoretical_loss": 3.4742755065390964,
      "tokens_seen": 1895825408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042978936810431296,
      "loss": 1.4185,
      "theoretical_loss": 3.4742654702957685,
      "tokens_seen": 1895890944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297693079237713,
      "loss": 1.432,
      "theoretical_loss": 3.4742554344964964,
      "tokens_seen": 1895956480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042974924774322973,
      "loss": 1.3078,
      "theoretical_loss": 3.474245399141246,
      "tokens_seen": 1896022016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297291875626881,
      "loss": 1.403,
      "theoretical_loss": 3.4742353642299815,
      "tokens_seen": 1896087552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042970912738214646,
      "loss": 1.3919,
      "theoretical_loss": 3.474225329762669,
      "tokens_seen": 1896153088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296890672016048,
      "loss": 1.3869,
      "theoretical_loss": 3.4742152957392722,
      "tokens_seen": 1896218624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296690070210632,
      "loss": 1.2555,
      "theoretical_loss": 3.4742052621597574,
      "tokens_seen": 1896284160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296489468405216,
      "loss": 1.3245,
      "theoretical_loss": 3.4741952290240894,
      "tokens_seen": 1896349696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042962888665997996,
      "loss": 1.2693,
      "theoretical_loss": 3.4741851963322326,
      "tokens_seen": 1896415232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296088264794383,
      "loss": 1.3253,
      "theoretical_loss": 3.4741751640841523,
      "tokens_seen": 1896480768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295887662988967,
      "loss": 1.4115,
      "theoretical_loss": 3.474165132279813,
      "tokens_seen": 1896546304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295687061183551,
      "loss": 1.3543,
      "theoretical_loss": 3.474155100919181,
      "tokens_seen": 1896611840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042954864593781347,
      "loss": 1.4173,
      "theoretical_loss": 3.47414507000222,
      "tokens_seen": 1896677376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042952858575727183,
      "loss": 1.3931,
      "theoretical_loss": 3.4741350395288966,
      "tokens_seen": 1896742912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295085255767302,
      "loss": 1.3902,
      "theoretical_loss": 3.4741250094991747,
      "tokens_seen": 1896808448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042948846539618855,
      "loss": 1.3042,
      "theoretical_loss": 3.4741149799130193,
      "tokens_seen": 1896873984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042946840521564697,
      "loss": 1.3126,
      "theoretical_loss": 3.474104950770396,
      "tokens_seen": 1896939520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042944834503510533,
      "loss": 1.3503,
      "theoretical_loss": 3.474094922071269,
      "tokens_seen": 1897005056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004294282848545637,
      "loss": 1.3756,
      "theoretical_loss": 3.474084893815605,
      "tokens_seen": 1897070592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042940822467402206,
      "loss": 1.3653,
      "theoretical_loss": 3.4740748660033676,
      "tokens_seen": 1897136128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042938816449348047,
      "loss": 1.3881,
      "theoretical_loss": 3.474064838634522,
      "tokens_seen": 1897201664
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042936810431293883,
      "loss": 1.4373,
      "theoretical_loss": 3.474054811709034,
      "tokens_seen": 1897267200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293480441323972,
      "loss": 1.3754,
      "theoretical_loss": 3.4740447852268685,
      "tokens_seen": 1897332736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042932798395185556,
      "loss": 1.3489,
      "theoretical_loss": 3.4740347591879903,
      "tokens_seen": 1897398272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293079237713139,
      "loss": 1.4002,
      "theoretical_loss": 3.4740247335923646,
      "tokens_seen": 1897463808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042928786359077234,
      "loss": 1.2348,
      "theoretical_loss": 3.474014708439956,
      "tokens_seen": 1897529344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004292678034102307,
      "loss": 1.4235,
      "theoretical_loss": 3.4740046837307306,
      "tokens_seen": 1897594880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042924774322968906,
      "loss": 1.3492,
      "theoretical_loss": 3.473994659464653,
      "tokens_seen": 1897660416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004292276830491474,
      "loss": 1.3499,
      "theoretical_loss": 3.4739846356416875,
      "tokens_seen": 1897725952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042920762286860584,
      "loss": 1.342,
      "theoretical_loss": 3.473974612261801,
      "tokens_seen": 1897791488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291875626880642,
      "loss": 1.2957,
      "theoretical_loss": 3.4739645893249573,
      "tokens_seen": 1897857024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042916750250752256,
      "loss": 1.349,
      "theoretical_loss": 3.4739545668311216,
      "tokens_seen": 1897922560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291474423269809,
      "loss": 1.3224,
      "theoretical_loss": 3.4739445447802595,
      "tokens_seen": 1897988096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291273821464393,
      "loss": 1.4146,
      "theoretical_loss": 3.473934523172336,
      "tokens_seen": 1898053632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291073219658977,
      "loss": 1.2691,
      "theoretical_loss": 3.473924502007316,
      "tokens_seen": 1898119168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042908726178535607,
      "loss": 1.4653,
      "theoretical_loss": 3.4739144812851643,
      "tokens_seen": 1898184704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042906720160481443,
      "loss": 1.3429,
      "theoretical_loss": 3.473904461005847,
      "tokens_seen": 1898250240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290471414242728,
      "loss": 1.2645,
      "theoretical_loss": 3.473894441169329,
      "tokens_seen": 1898315776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290270812437312,
      "loss": 1.448,
      "theoretical_loss": 3.4738844217755744,
      "tokens_seen": 1898381312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042900702106318957,
      "loss": 1.3709,
      "theoretical_loss": 3.47387440282455,
      "tokens_seen": 1898446848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042898696088264793,
      "loss": 1.3676,
      "theoretical_loss": 3.4738643843162196,
      "tokens_seen": 1898512384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289669007021063,
      "loss": 1.3592,
      "theoretical_loss": 3.4738543662505488,
      "tokens_seen": 1898577920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289468405215647,
      "loss": 1.3627,
      "theoretical_loss": 3.473844348627503,
      "tokens_seen": 1898643456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289267803410231,
      "loss": 1.4196,
      "theoretical_loss": 3.473834331447047,
      "tokens_seen": 1898708992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042890672016048144,
      "loss": 1.3327,
      "theoretical_loss": 3.4738243147091463,
      "tokens_seen": 1898774528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288866599799398,
      "loss": 1.3173,
      "theoretical_loss": 3.4738142984137657,
      "tokens_seen": 1898840064
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042886659979939816,
      "loss": 1.3653,
      "theoretical_loss": 3.473804282560871,
      "tokens_seen": 1898905600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288465396188566,
      "loss": 1.3874,
      "theoretical_loss": 3.473794267150427,
      "tokens_seen": 1898971136
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 692658,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4199891090393066,
      "objective/train/theoretical_loss": 3.4737842521823983,
      "objective/train/tokens_used": 269495776,
      "theoretical_loss": 3.4737842521823983,
      "tokens_seen": 1899036672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042882647943831494,
      "loss": 1.3691,
      "theoretical_loss": 3.4737842521823983,
      "tokens_seen": 1899036672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288064192577733,
      "loss": 1.4167,
      "theoretical_loss": 3.473774237656751,
      "tokens_seen": 1899102208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042878635907723166,
      "loss": 1.4092,
      "theoretical_loss": 3.47376422357345,
      "tokens_seen": 1899167744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287662988966901,
      "loss": 1.3841,
      "theoretical_loss": 3.4737542099324603,
      "tokens_seen": 1899233280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042874623871614844,
      "loss": 1.337,
      "theoretical_loss": 3.4737441967337475,
      "tokens_seen": 1899298816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287261785356068,
      "loss": 1.3721,
      "theoretical_loss": 3.4737341839772764,
      "tokens_seen": 1899364352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042870611835506517,
      "loss": 1.3862,
      "theoretical_loss": 3.4737241716630125,
      "tokens_seen": 1899429888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042868605817452353,
      "loss": 1.3537,
      "theoretical_loss": 3.4737141597909207,
      "tokens_seen": 1899495424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000428665997993982,
      "loss": 1.3901,
      "theoretical_loss": 3.4737041483609663,
      "tokens_seen": 1899560960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042864593781344036,
      "loss": 1.3544,
      "theoretical_loss": 3.4736941373731147,
      "tokens_seen": 1899626496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286258776328987,
      "loss": 1.2165,
      "theoretical_loss": 3.4736841268273313,
      "tokens_seen": 1899692032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286058174523571,
      "loss": 1.3403,
      "theoretical_loss": 3.473674116723581,
      "tokens_seen": 1899757568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285857572718155,
      "loss": 1.3762,
      "theoretical_loss": 3.473664107061829,
      "tokens_seen": 1899823104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042856569709127387,
      "loss": 1.3103,
      "theoretical_loss": 3.4736540978420405,
      "tokens_seen": 1899888640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042854563691073223,
      "loss": 1.3381,
      "theoretical_loss": 3.473644089064181,
      "tokens_seen": 1899954176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285255767301906,
      "loss": 1.2995,
      "theoretical_loss": 3.473634080728216,
      "tokens_seen": 1900019712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042850551654964895,
      "loss": 1.3195,
      "theoretical_loss": 3.47362407283411,
      "tokens_seen": 1900085248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042848545636910737,
      "loss": 1.4056,
      "theoretical_loss": 3.4736140653818284,
      "tokens_seen": 1900150784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042846539618856573,
      "loss": 1.362,
      "theoretical_loss": 3.473604058371337,
      "tokens_seen": 1900216320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284453360080241,
      "loss": 1.349,
      "theoretical_loss": 3.4735940518026007,
      "tokens_seen": 1900281856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042842527582748246,
      "loss": 1.3205,
      "theoretical_loss": 3.473584045675585,
      "tokens_seen": 1900347392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042840521564694087,
      "loss": 1.3586,
      "theoretical_loss": 3.473574039990255,
      "tokens_seen": 1900412928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042838515546639923,
      "loss": 1.3669,
      "theoretical_loss": 3.473564034746576,
      "tokens_seen": 1900478464
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283650952858576,
      "loss": 1.2914,
      "theoretical_loss": 3.4735540299445127,
      "tokens_seen": 1900544000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042834503510531596,
      "loss": 1.3597,
      "theoretical_loss": 3.4735440255840313,
      "tokens_seen": 1900609536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283249749247743,
      "loss": 1.4428,
      "theoretical_loss": 3.4735340216650967,
      "tokens_seen": 1900675072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042830491474423274,
      "loss": 1.3922,
      "theoretical_loss": 3.4735240181876743,
      "tokens_seen": 1900740608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282848545636911,
      "loss": 1.3481,
      "theoretical_loss": 3.473514015151729,
      "tokens_seen": 1900806144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042826479438314946,
      "loss": 1.3261,
      "theoretical_loss": 3.4735040125572265,
      "tokens_seen": 1900871680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282447342026078,
      "loss": 1.2631,
      "theoretical_loss": 3.4734940104041323,
      "tokens_seen": 1900937216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042822467402206624,
      "loss": 1.3716,
      "theoretical_loss": 3.473484008692411,
      "tokens_seen": 1901002752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282046138415246,
      "loss": 1.3907,
      "theoretical_loss": 3.4734740074220283,
      "tokens_seen": 1901068288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042818455366098296,
      "loss": 1.3052,
      "theoretical_loss": 3.4734640065929496,
      "tokens_seen": 1901133824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281644934804413,
      "loss": 1.3997,
      "theoretical_loss": 3.47345400620514,
      "tokens_seen": 1901199360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281444332998997,
      "loss": 1.4308,
      "theoretical_loss": 3.473444006258565,
      "tokens_seen": 1901264896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281243731193581,
      "loss": 1.3591,
      "theoretical_loss": 3.47343400675319,
      "tokens_seen": 1901330432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042810431293881647,
      "loss": 1.4344,
      "theoretical_loss": 3.47342400768898,
      "tokens_seen": 1901395968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042808425275827483,
      "loss": 1.4546,
      "theoretical_loss": 3.473414009065901,
      "tokens_seen": 1901461504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004280641925777332,
      "loss": 1.3709,
      "theoretical_loss": 3.4734040108839173,
      "tokens_seen": 1901527040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004280441323971916,
      "loss": 1.317,
      "theoretical_loss": 3.473394013142995,
      "tokens_seen": 1901592576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042802407221664997,
      "loss": 1.4808,
      "theoretical_loss": 3.4733840158430995,
      "tokens_seen": 1901658112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042800401203610833,
      "loss": 1.419,
      "theoretical_loss": 3.4733740189841953,
      "tokens_seen": 1901723648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279839518555667,
      "loss": 1.3145,
      "theoretical_loss": 3.4733640225662485,
      "tokens_seen": 1901789184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279638916750251,
      "loss": 1.3979,
      "theoretical_loss": 3.4733540265892247,
      "tokens_seen": 1901854720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279438314944835,
      "loss": 1.3897,
      "theoretical_loss": 3.473344031053089,
      "tokens_seen": 1901920256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042792377131394184,
      "loss": 1.405,
      "theoretical_loss": 3.473334035957806,
      "tokens_seen": 1901985792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279037111334002,
      "loss": 1.4781,
      "theoretical_loss": 3.473324041303342,
      "tokens_seen": 1902051328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042788365095285856,
      "loss": 1.3574,
      "theoretical_loss": 3.4733140470896617,
      "tokens_seen": 1902116864
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427863590772317,
      "loss": 1.4382,
      "theoretical_loss": 3.473304053316731,
      "tokens_seen": 1902182400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042784353059177534,
      "loss": 1.3338,
      "theoretical_loss": 3.4732940599845152,
      "tokens_seen": 1902247936
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 693391,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4137303829193115,
      "objective/train/theoretical_loss": 3.4732840670929797,
      "objective/train/tokens_used": 272772576,
      "theoretical_loss": 3.4732840670929797,
      "tokens_seen": 1902313472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278234704112337,
      "loss": 1.3666,
      "theoretical_loss": 3.4732840670929797,
      "tokens_seen": 1902313472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042780341023069206,
      "loss": 1.3747,
      "theoretical_loss": 3.47327407464209,
      "tokens_seen": 1902379008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277833500501505,
      "loss": 1.3595,
      "theoretical_loss": 3.4732640826318106,
      "tokens_seen": 1902444544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042776328986960884,
      "loss": 1.4031,
      "theoretical_loss": 3.473254091062108,
      "tokens_seen": 1902510080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277432296890672,
      "loss": 1.3706,
      "theoretical_loss": 3.473244099932947,
      "tokens_seen": 1902575616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042772316950852557,
      "loss": 1.2708,
      "theoretical_loss": 3.4732341092442933,
      "tokens_seen": 1902641152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042770310932798393,
      "loss": 1.216,
      "theoretical_loss": 3.4732241189961117,
      "tokens_seen": 1902706688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042768304914744235,
      "loss": 1.3736,
      "theoretical_loss": 3.473214129188369,
      "tokens_seen": 1902772224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276629889669007,
      "loss": 1.475,
      "theoretical_loss": 3.473204139821029,
      "tokens_seen": 1902837760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042764292878635907,
      "loss": 1.3918,
      "theoretical_loss": 3.473194150894058,
      "tokens_seen": 1902903296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042762286860581743,
      "loss": 1.3059,
      "theoretical_loss": 3.473184162407421,
      "tokens_seen": 1902968832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042760280842527585,
      "loss": 1.2221,
      "theoretical_loss": 3.473174174361084,
      "tokens_seen": 1903034368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275827482447342,
      "loss": 1.3464,
      "theoretical_loss": 3.473164186755012,
      "tokens_seen": 1903099904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275626880641926,
      "loss": 1.3136,
      "theoretical_loss": 3.47315419958917,
      "tokens_seen": 1903165440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042754262788365094,
      "loss": 1.4245,
      "theoretical_loss": 3.4731442128635246,
      "tokens_seen": 1903230976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275225677031093,
      "loss": 1.3853,
      "theoretical_loss": 3.4731342265780407,
      "tokens_seen": 1903296512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275025075225677,
      "loss": 1.3413,
      "theoretical_loss": 3.473124240732683,
      "tokens_seen": 1903362048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274824473420261,
      "loss": 1.5591,
      "theoretical_loss": 3.473114255327418,
      "tokens_seen": 1903427584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042746238716148444,
      "loss": 1.324,
      "theoretical_loss": 3.473104270362211,
      "tokens_seen": 1903493120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274423269809428,
      "loss": 1.3618,
      "theoretical_loss": 3.4730942858370266,
      "tokens_seen": 1903558656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274222668004012,
      "loss": 1.3494,
      "theoretical_loss": 3.473084301751831,
      "tokens_seen": 1903624192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274022066198596,
      "loss": 1.2978,
      "theoretical_loss": 3.47307431810659,
      "tokens_seen": 1903689728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042738214643931794,
      "loss": 1.2977,
      "theoretical_loss": 3.4730643349012684,
      "tokens_seen": 1903755264
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273620862587763,
      "loss": 1.3044,
      "theoretical_loss": 3.4730543521358315,
      "tokens_seen": 1903820800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042734202607823467,
      "loss": 1.2978,
      "theoretical_loss": 3.4730443698102453,
      "tokens_seen": 1903886336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273219658976931,
      "loss": 1.2835,
      "theoretical_loss": 3.473034387924475,
      "tokens_seen": 1903951872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042730190571715145,
      "loss": 1.4056,
      "theoretical_loss": 3.473024406478487,
      "tokens_seen": 1904017408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272818455366098,
      "loss": 1.3813,
      "theoretical_loss": 3.473014425472245,
      "tokens_seen": 1904082944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042726178535606817,
      "loss": 1.4001,
      "theoretical_loss": 3.4730044449057162,
      "tokens_seen": 1904148480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272417251755266,
      "loss": 1.2306,
      "theoretical_loss": 3.472994464778865,
      "tokens_seen": 1904214016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042722166499498495,
      "loss": 1.3758,
      "theoretical_loss": 3.472984485091657,
      "tokens_seen": 1904279552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272016048144433,
      "loss": 1.4079,
      "theoretical_loss": 3.4729745058440584,
      "tokens_seen": 1904345088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271815446339017,
      "loss": 1.2416,
      "theoretical_loss": 3.4729645270360345,
      "tokens_seen": 1904410624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271614844533601,
      "loss": 1.2875,
      "theoretical_loss": 3.47295454866755,
      "tokens_seen": 1904476160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271414242728185,
      "loss": 1.3996,
      "theoretical_loss": 3.472944570738571,
      "tokens_seen": 1904541696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042712136409227687,
      "loss": 1.4145,
      "theoretical_loss": 3.4729345932490636,
      "tokens_seen": 1904607232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042710130391173523,
      "loss": 1.1262,
      "theoretical_loss": 3.472924616198993,
      "tokens_seen": 1904672768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270812437311936,
      "loss": 1.2712,
      "theoretical_loss": 3.472914639588324,
      "tokens_seen": 1904738304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427061183550652,
      "loss": 1.2907,
      "theoretical_loss": 3.4729046634170224,
      "tokens_seen": 1904803840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042704112337011037,
      "loss": 1.3632,
      "theoretical_loss": 3.4728946876850544,
      "tokens_seen": 1904869376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042702106318956873,
      "loss": 1.2769,
      "theoretical_loss": 3.4728847123923847,
      "tokens_seen": 1904934912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270010030090271,
      "loss": 1.2793,
      "theoretical_loss": 3.4728747375389792,
      "tokens_seen": 1905000448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269809428284855,
      "loss": 1.4319,
      "theoretical_loss": 3.4728647631248037,
      "tokens_seen": 1905065984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269608826479439,
      "loss": 1.4494,
      "theoretical_loss": 3.4728547891498236,
      "tokens_seen": 1905131520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042694082246740224,
      "loss": 1.4156,
      "theoretical_loss": 3.4728448156140046,
      "tokens_seen": 1905197056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269207622868606,
      "loss": 1.3812,
      "theoretical_loss": 3.4728348425173117,
      "tokens_seen": 1905262592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042690070210631896,
      "loss": 1.3979,
      "theoretical_loss": 3.4728248698597106,
      "tokens_seen": 1905328128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268806419257774,
      "loss": 1.3651,
      "theoretical_loss": 3.4728148976411677,
      "tokens_seen": 1905393664
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042686058174523574,
      "loss": 1.3123,
      "theoretical_loss": 3.472804925861648,
      "tokens_seen": 1905459200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268405215646941,
      "loss": 1.2943,
      "theoretical_loss": 3.472794954521116,
      "tokens_seen": 1905524736
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 694687,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3013882637023926,
      "objective/train/theoretical_loss": 3.472784983619539,
      "objective/train/tokens_used": 276049376,
      "theoretical_loss": 3.472784983619539,
      "tokens_seen": 1905590272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042682046138415246,
      "loss": 1.3291,
      "theoretical_loss": 3.472784983619539,
      "tokens_seen": 1905590272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268004012036109,
      "loss": 1.4634,
      "theoretical_loss": 3.4727750131568818,
      "tokens_seen": 1905655808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042678034102306924,
      "loss": 1.2832,
      "theoretical_loss": 3.47276504313311,
      "tokens_seen": 1905721344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004267602808425276,
      "loss": 1.383,
      "theoretical_loss": 3.472755073548189,
      "tokens_seen": 1905786880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042674022066198597,
      "loss": 1.4233,
      "theoretical_loss": 3.472745104402085,
      "tokens_seen": 1905852416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042672016048144433,
      "loss": 1.3622,
      "theoretical_loss": 3.472735135694763,
      "tokens_seen": 1905917952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042670010030090275,
      "loss": 1.2575,
      "theoretical_loss": 3.472725167426189,
      "tokens_seen": 1905983488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266800401203611,
      "loss": 1.3806,
      "theoretical_loss": 3.472715199596329,
      "tokens_seen": 1906049024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042665997993981947,
      "loss": 1.4998,
      "theoretical_loss": 3.472705232205147,
      "tokens_seen": 1906114560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042663991975927783,
      "loss": 1.2589,
      "theoretical_loss": 3.4726952652526104,
      "tokens_seen": 1906180096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042661985957873625,
      "loss": 1.4269,
      "theoretical_loss": 3.472685298738684,
      "tokens_seen": 1906245632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265997993981946,
      "loss": 1.3553,
      "theoretical_loss": 3.472675332663333,
      "tokens_seen": 1906311168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000426579739217653,
      "loss": 1.304,
      "theoretical_loss": 3.472665367026524,
      "tokens_seen": 1906376704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042655967903711134,
      "loss": 1.2914,
      "theoretical_loss": 3.472655401828222,
      "tokens_seen": 1906442240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265396188565697,
      "loss": 1.3901,
      "theoretical_loss": 3.4726454370683926,
      "tokens_seen": 1906507776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265195586760281,
      "loss": 1.3213,
      "theoretical_loss": 3.472635472747002,
      "tokens_seen": 1906573312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264994984954865,
      "loss": 1.4238,
      "theoretical_loss": 3.4726255088640148,
      "tokens_seen": 1906638848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042647943831494484,
      "loss": 1.4514,
      "theoretical_loss": 3.472615545419398,
      "tokens_seen": 1906704384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264593781344032,
      "loss": 1.4761,
      "theoretical_loss": 3.472605582413116,
      "tokens_seen": 1906769920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264393179538616,
      "loss": 1.3466,
      "theoretical_loss": 3.4725956198451353,
      "tokens_seen": 1906835456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042641925777332,
      "loss": 1.4145,
      "theoretical_loss": 3.4725856577154213,
      "tokens_seen": 1906900992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042639919759277834,
      "loss": 1.5121,
      "theoretical_loss": 3.47257569602394,
      "tokens_seen": 1906966528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263791374122367,
      "loss": 1.4003,
      "theoretical_loss": 3.4725657347706558,
      "tokens_seen": 1907032064
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042635907723169507,
      "loss": 1.4397,
      "theoretical_loss": 3.4725557739555355,
      "tokens_seen": 1907097600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263390170511535,
      "loss": 1.355,
      "theoretical_loss": 3.4725458135785447,
      "tokens_seen": 1907163136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042631895687061185,
      "loss": 1.3257,
      "theoretical_loss": 3.4725358536396485,
      "tokens_seen": 1907228672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262988966900702,
      "loss": 1.3299,
      "theoretical_loss": 3.4725258941388137,
      "tokens_seen": 1907294208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042627883650952857,
      "loss": 1.3987,
      "theoretical_loss": 3.4725159350760046,
      "tokens_seen": 1907359744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000426258776328987,
      "loss": 1.3068,
      "theoretical_loss": 3.472505976451188,
      "tokens_seen": 1907425280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042623871614844535,
      "loss": 1.2822,
      "theoretical_loss": 3.472496018264329,
      "tokens_seen": 1907490816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262186559679037,
      "loss": 1.4213,
      "theoretical_loss": 3.4724860605153935,
      "tokens_seen": 1907556352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261985957873621,
      "loss": 1.375,
      "theoretical_loss": 3.4724761032043467,
      "tokens_seen": 1907621888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042617853560682044,
      "loss": 1.2884,
      "theoretical_loss": 3.472466146331155,
      "tokens_seen": 1907687424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042615847542627885,
      "loss": 1.3896,
      "theoretical_loss": 3.4724561898957838,
      "tokens_seen": 1907752960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261384152457372,
      "loss": 1.4266,
      "theoretical_loss": 3.4724462338981983,
      "tokens_seen": 1907818496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261183550651956,
      "loss": 1.4355,
      "theoretical_loss": 3.4724362783383658,
      "tokens_seen": 1907884032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042609829488465394,
      "loss": 1.4033,
      "theoretical_loss": 3.47242632321625,
      "tokens_seen": 1907949568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042607823470411236,
      "loss": 1.3222,
      "theoretical_loss": 3.4724163685318183,
      "tokens_seen": 1908015104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260581745235707,
      "loss": 1.3804,
      "theoretical_loss": 3.472406414285035,
      "tokens_seen": 1908080640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260381143430291,
      "loss": 1.2951,
      "theoretical_loss": 3.472396460475867,
      "tokens_seen": 1908146176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042601805416248744,
      "loss": 1.4139,
      "theoretical_loss": 3.4723865071042797,
      "tokens_seen": 1908211712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259979939819458,
      "loss": 1.3795,
      "theoretical_loss": 3.4723765541702383,
      "tokens_seen": 1908277248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259779338014042,
      "loss": 1.2519,
      "theoretical_loss": 3.472366601673709,
      "tokens_seen": 1908342784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259578736208626,
      "loss": 1.237,
      "theoretical_loss": 3.4723566496146576,
      "tokens_seen": 1908408320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042593781344032095,
      "loss": 1.3796,
      "theoretical_loss": 3.47234669799305,
      "tokens_seen": 1908473856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259177532597793,
      "loss": 1.4262,
      "theoretical_loss": 3.4723367468088515,
      "tokens_seen": 1908539392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258976930792377,
      "loss": 1.49,
      "theoretical_loss": 3.4723267960620277,
      "tokens_seen": 1908604928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258776328986961,
      "loss": 1.3709,
      "theoretical_loss": 3.4723168457525446,
      "tokens_seen": 1908670464
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042585757271815445,
      "loss": 1.3193,
      "theoretical_loss": 3.4723068958803687,
      "tokens_seen": 1908736000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258375125376128,
      "loss": 1.3829,
      "theoretical_loss": 3.4722969464454647,
      "tokens_seen": 1908801536
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 695871,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2853118181228638,
      "objective/train/theoretical_loss": 3.472286997447799,
      "objective/train/tokens_used": 279326176,
      "theoretical_loss": 3.472286997447799,
      "tokens_seen": 1908867072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258174523570712,
      "loss": 1.3275,
      "theoretical_loss": 3.472286997447799,
      "tokens_seen": 1908867072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257973921765296,
      "loss": 1.3316,
      "theoretical_loss": 3.472277048887337,
      "tokens_seen": 1908932608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042577733199598795,
      "loss": 1.3028,
      "theoretical_loss": 3.4722671007640447,
      "tokens_seen": 1908998144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257572718154463,
      "loss": 1.389,
      "theoretical_loss": 3.472257153077888,
      "tokens_seen": 1909063680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257372116349047,
      "loss": 1.4206,
      "theoretical_loss": 3.4722472058288325,
      "tokens_seen": 1909129216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257171514543631,
      "loss": 1.455,
      "theoretical_loss": 3.472237259016844,
      "tokens_seen": 1909194752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042569709127382145,
      "loss": 1.3542,
      "theoretical_loss": 3.472227312641888,
      "tokens_seen": 1909260288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256770310932798,
      "loss": 1.3504,
      "theoretical_loss": 3.472217366703931,
      "tokens_seen": 1909325824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042565697091273823,
      "loss": 1.2731,
      "theoretical_loss": 3.4722074212029383,
      "tokens_seen": 1909391360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042563691073219665,
      "loss": 1.4055,
      "theoretical_loss": 3.4721974761388754,
      "tokens_seen": 1909456896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425616850551655,
      "loss": 1.3072,
      "theoretical_loss": 3.4721875315117092,
      "tokens_seen": 1909522432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255967903711134,
      "loss": 1.3071,
      "theoretical_loss": 3.4721775873214042,
      "tokens_seen": 1909587968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042557673019057174,
      "loss": 1.403,
      "theoretical_loss": 3.472167643567927,
      "tokens_seen": 1909653504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255566700100301,
      "loss": 1.3652,
      "theoretical_loss": 3.472157700251244,
      "tokens_seen": 1909719040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255366098294885,
      "loss": 1.3595,
      "theoretical_loss": 3.472147757371319,
      "tokens_seen": 1909784576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255165496489469,
      "loss": 1.2875,
      "theoretical_loss": 3.47213781492812,
      "tokens_seen": 1909850112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042549648946840524,
      "loss": 1.2068,
      "theoretical_loss": 3.472127872921612,
      "tokens_seen": 1909915648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254764292878636,
      "loss": 1.3615,
      "theoretical_loss": 3.4721179313517605,
      "tokens_seen": 1909981184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425456369107322,
      "loss": 1.4361,
      "theoretical_loss": 3.4721079902185314,
      "tokens_seen": 1910046720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254363089267804,
      "loss": 1.4119,
      "theoretical_loss": 3.4720980495218914,
      "tokens_seen": 1910112256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042541624874623874,
      "loss": 1.4695,
      "theoretical_loss": 3.4720881092618052,
      "tokens_seen": 1910177792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253961885656971,
      "loss": 1.3584,
      "theoretical_loss": 3.472078169438239,
      "tokens_seen": 1910243328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042537612838515547,
      "loss": 1.348,
      "theoretical_loss": 3.4720682300511596,
      "tokens_seen": 1910308864
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253560682046139,
      "loss": 1.3381,
      "theoretical_loss": 3.4720582911005318,
      "tokens_seen": 1910374400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042533600802407225,
      "loss": 1.2519,
      "theoretical_loss": 3.4720483525863215,
      "tokens_seen": 1910439936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253159478435306,
      "loss": 1.4232,
      "theoretical_loss": 3.472038414508495,
      "tokens_seen": 1910505472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042529588766298897,
      "loss": 1.3168,
      "theoretical_loss": 3.4720284768670178,
      "tokens_seen": 1910571008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252758274824474,
      "loss": 1.2948,
      "theoretical_loss": 3.4720185396618564,
      "tokens_seen": 1910636544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042525576730190575,
      "loss": 1.3708,
      "theoretical_loss": 3.472008602892976,
      "tokens_seen": 1910702080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252357071213641,
      "loss": 1.3437,
      "theoretical_loss": 3.4719986665603426,
      "tokens_seen": 1910767616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252156469408225,
      "loss": 1.3479,
      "theoretical_loss": 3.471988730663922,
      "tokens_seen": 1910833152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042519558676028084,
      "loss": 1.3587,
      "theoretical_loss": 3.4719787952036807,
      "tokens_seen": 1910898688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042517552657973925,
      "loss": 1.3383,
      "theoretical_loss": 3.471968860179584,
      "tokens_seen": 1910964224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004251554663991976,
      "loss": 1.4451,
      "theoretical_loss": 3.4719589255915984,
      "tokens_seen": 1911029760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425135406218656,
      "loss": 1.325,
      "theoretical_loss": 3.471948991439689,
      "tokens_seen": 1911095296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042511534603811434,
      "loss": 1.2898,
      "theoretical_loss": 3.471939057723822,
      "tokens_seen": 1911160832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042509528585757276,
      "loss": 1.3863,
      "theoretical_loss": 3.471929124443964,
      "tokens_seen": 1911226368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250752256770311,
      "loss": 1.3278,
      "theoretical_loss": 3.4719191916000796,
      "tokens_seen": 1911291904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250551654964895,
      "loss": 1.23,
      "theoretical_loss": 3.471909259192136,
      "tokens_seen": 1911357440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042503510531594784,
      "loss": 1.3665,
      "theoretical_loss": 3.471899327220098,
      "tokens_seen": 1911422976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250150451354062,
      "loss": 1.3678,
      "theoretical_loss": 3.4718893956839327,
      "tokens_seen": 1911488512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249949849548646,
      "loss": 1.3816,
      "theoretical_loss": 3.471879464583605,
      "tokens_seen": 1911554048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424974924774323,
      "loss": 1.4047,
      "theoretical_loss": 3.4718695339190813,
      "tokens_seen": 1911619584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042495486459378135,
      "loss": 1.2681,
      "theoretical_loss": 3.4718596036903273,
      "tokens_seen": 1911685120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249348044132397,
      "loss": 1.345,
      "theoretical_loss": 3.4718496738973097,
      "tokens_seen": 1911750656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249147442326981,
      "loss": 1.377,
      "theoretical_loss": 3.4718397445399933,
      "tokens_seen": 1911816192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248946840521565,
      "loss": 1.4166,
      "theoretical_loss": 3.471829815618345,
      "tokens_seen": 1911881728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042487462387161485,
      "loss": 1.438,
      "theoretical_loss": 3.4718198871323303,
      "tokens_seen": 1911947264
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248545636910732,
      "loss": 1.4252,
      "theoretical_loss": 3.471809959081915,
      "tokens_seen": 1912012800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248345035105316,
      "loss": 1.384,
      "theoretical_loss": 3.4718000314670654,
      "tokens_seen": 1912078336
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 697086,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3212376832962036,
      "objective/train/theoretical_loss": 3.4717901042877477,
      "objective/train/tokens_used": 282602976,
      "theoretical_loss": 3.4717901042877477,
      "tokens_seen": 1912143872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042481444332999,
      "loss": 1.3844,
      "theoretical_loss": 3.4717901042877477,
      "tokens_seen": 1912143872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042479438314944835,
      "loss": 1.4276,
      "theoretical_loss": 3.471780177543927,
      "tokens_seen": 1912209408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247743229689067,
      "loss": 1.4065,
      "theoretical_loss": 3.47177025123557,
      "tokens_seen": 1912274944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247542627883651,
      "loss": 1.3849,
      "theoretical_loss": 3.4717603253626423,
      "tokens_seen": 1912340480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247342026078235,
      "loss": 1.3053,
      "theoretical_loss": 3.47175039992511,
      "tokens_seen": 1912406016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042471414242728186,
      "loss": 1.3719,
      "theoretical_loss": 3.4717404749229397,
      "tokens_seen": 1912471552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246940822467402,
      "loss": 1.3186,
      "theoretical_loss": 3.471730550356096,
      "tokens_seen": 1912537088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246740220661986,
      "loss": 1.3231,
      "theoretical_loss": 3.4717206262245464,
      "tokens_seen": 1912602624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424653961885657,
      "loss": 1.2723,
      "theoretical_loss": 3.471710702528256,
      "tokens_seen": 1912668160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042463390170511536,
      "loss": 1.4056,
      "theoretical_loss": 3.4717007792671906,
      "tokens_seen": 1912733696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246138415245737,
      "loss": 1.3863,
      "theoretical_loss": 3.4716908564413167,
      "tokens_seen": 1912799232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245937813440321,
      "loss": 1.3216,
      "theoretical_loss": 3.4716809340506,
      "tokens_seen": 1912864768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042457372116349044,
      "loss": 1.4053,
      "theoretical_loss": 3.4716710120950074,
      "tokens_seen": 1912930304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042455366098294886,
      "loss": 1.3346,
      "theoretical_loss": 3.4716610905745036,
      "tokens_seen": 1912995840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245336008024072,
      "loss": 1.3701,
      "theoretical_loss": 3.4716511694890553,
      "tokens_seen": 1913061376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245135406218656,
      "loss": 1.311,
      "theoretical_loss": 3.471641248838629,
      "tokens_seen": 1913126912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042449348044132395,
      "loss": 1.3466,
      "theoretical_loss": 3.4716313286231895,
      "tokens_seen": 1913192448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042447342026078236,
      "loss": 1.4149,
      "theoretical_loss": 3.4716214088427035,
      "tokens_seen": 1913257984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244533600802407,
      "loss": 1.3819,
      "theoretical_loss": 3.4716114894971373,
      "tokens_seen": 1913323520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244332998996991,
      "loss": 1.2682,
      "theoretical_loss": 3.471601570586456,
      "tokens_seen": 1913389056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042441323971915745,
      "loss": 1.3924,
      "theoretical_loss": 3.471591652110627,
      "tokens_seen": 1913454592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243931795386158,
      "loss": 1.3623,
      "theoretical_loss": 3.471581734069616,
      "tokens_seen": 1913520128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042437311935807423,
      "loss": 1.3461,
      "theoretical_loss": 3.471571816463388,
      "tokens_seen": 1913585664
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243530591775326,
      "loss": 1.3284,
      "theoretical_loss": 3.4715618992919097,
      "tokens_seen": 1913651200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042433299899699095,
      "loss": 1.3253,
      "theoretical_loss": 3.4715519825551477,
      "tokens_seen": 1913716736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243129388164493,
      "loss": 1.2873,
      "theoretical_loss": 3.4715420662530674,
      "tokens_seen": 1913782272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042429287863590773,
      "loss": 1.3215,
      "theoretical_loss": 3.471532150385635,
      "tokens_seen": 1913847808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242728184553661,
      "loss": 1.3909,
      "theoretical_loss": 3.4715222349528165,
      "tokens_seen": 1913913344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042425275827482446,
      "loss": 1.3988,
      "theoretical_loss": 3.471512319954578,
      "tokens_seen": 1913978880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242326980942828,
      "loss": 1.2377,
      "theoretical_loss": 3.4715024053908854,
      "tokens_seen": 1914044416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242126379137412,
      "loss": 1.3359,
      "theoretical_loss": 3.4714924912617056,
      "tokens_seen": 1914109952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241925777331996,
      "loss": 1.1767,
      "theoretical_loss": 3.471482577567004,
      "tokens_seen": 1914175488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424172517552658,
      "loss": 1.3653,
      "theoretical_loss": 3.4714726643067464,
      "tokens_seen": 1914241024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241524573721164,
      "loss": 1.3332,
      "theoretical_loss": 3.4714627514808996,
      "tokens_seen": 1914306560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042413239719157474,
      "loss": 1.3313,
      "theoretical_loss": 3.4714528390894293,
      "tokens_seen": 1914372096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042411233701103316,
      "loss": 1.491,
      "theoretical_loss": 3.4714429271323013,
      "tokens_seen": 1914437632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240922768304915,
      "loss": 1.3816,
      "theoretical_loss": 3.471433015609482,
      "tokens_seen": 1914503168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240722166499499,
      "loss": 1.4036,
      "theoretical_loss": 3.4714231045209383,
      "tokens_seen": 1914568704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042405215646940824,
      "loss": 1.3297,
      "theoretical_loss": 3.471413193866635,
      "tokens_seen": 1914634240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240320962888666,
      "loss": 1.1832,
      "theoretical_loss": 3.4714032836465387,
      "tokens_seen": 1914699776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424012036108325,
      "loss": 1.3462,
      "theoretical_loss": 3.4713933738606157,
      "tokens_seen": 1914765312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239919759277834,
      "loss": 1.3834,
      "theoretical_loss": 3.4713834645088317,
      "tokens_seen": 1914830848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042397191574724175,
      "loss": 1.3982,
      "theoretical_loss": 3.471373555591154,
      "tokens_seen": 1914896384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239518555667001,
      "loss": 1.4738,
      "theoretical_loss": 3.4713636471075473,
      "tokens_seen": 1914961920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239317953861585,
      "loss": 1.3269,
      "theoretical_loss": 3.471353739057978,
      "tokens_seen": 1915027456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239117352056169,
      "loss": 1.4749,
      "theoretical_loss": 3.4713438314424128,
      "tokens_seen": 1915092992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042389167502507525,
      "loss": 1.3876,
      "theoretical_loss": 3.4713339242608177,
      "tokens_seen": 1915158528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238716148445336,
      "loss": 1.4143,
      "theoretical_loss": 3.471324017513159,
      "tokens_seen": 1915224064
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423851554663992,
      "loss": 1.3773,
      "theoretical_loss": 3.471314111199402,
      "tokens_seen": 1915289600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238314944834504,
      "loss": 1.3953,
      "theoretical_loss": 3.4713042053195133,
      "tokens_seen": 1915355136
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 698281,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3410608768463135,
      "objective/train/theoretical_loss": 3.4712942998734597,
      "objective/train/tokens_used": 285879776,
      "theoretical_loss": 3.4712942998734597,
      "tokens_seen": 1915420672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042381143430290875,
      "loss": 1.3491,
      "theoretical_loss": 3.4712942998734597,
      "tokens_seen": 1915420672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237913741223671,
      "loss": 1.322,
      "theoretical_loss": 3.4712843948612067,
      "tokens_seen": 1915486208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237713139418255,
      "loss": 1.4388,
      "theoretical_loss": 3.47127449028272,
      "tokens_seen": 1915551744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237512537612839,
      "loss": 1.297,
      "theoretical_loss": 3.4712645861379667,
      "tokens_seen": 1915617280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042373119358074226,
      "loss": 1.4093,
      "theoretical_loss": 3.4712546824269124,
      "tokens_seen": 1915682816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237111334002006,
      "loss": 1.3321,
      "theoretical_loss": 3.471244779149524,
      "tokens_seen": 1915748352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423691073219659,
      "loss": 1.4864,
      "theoretical_loss": 3.4712348763057665,
      "tokens_seen": 1915813888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236710130391174,
      "loss": 1.4675,
      "theoretical_loss": 3.471224973895607,
      "tokens_seen": 1915879424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042365095285857576,
      "loss": 1.3899,
      "theoretical_loss": 3.471215071919012,
      "tokens_seen": 1915944960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236308926780341,
      "loss": 1.4491,
      "theoretical_loss": 3.4712051703759466,
      "tokens_seen": 1916010496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236108324974925,
      "loss": 1.2167,
      "theoretical_loss": 3.4711952692663774,
      "tokens_seen": 1916076032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042359077231695085,
      "loss": 1.3526,
      "theoretical_loss": 3.471185368590271,
      "tokens_seen": 1916141568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042357071213640926,
      "loss": 1.4873,
      "theoretical_loss": 3.4711754683475933,
      "tokens_seen": 1916207104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004235506519558676,
      "loss": 1.2723,
      "theoretical_loss": 3.47116556853831,
      "tokens_seen": 1916272640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423530591775326,
      "loss": 1.5336,
      "theoretical_loss": 3.471155669162388,
      "tokens_seen": 1916338176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042351053159478435,
      "loss": 1.3566,
      "theoretical_loss": 3.4711457702197936,
      "tokens_seen": 1916403712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042349047141424276,
      "loss": 1.2315,
      "theoretical_loss": 3.4711358717104925,
      "tokens_seen": 1916469248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042347041123370113,
      "loss": 1.4048,
      "theoretical_loss": 3.471125973634451,
      "tokens_seen": 1916534784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234503510531595,
      "loss": 1.3125,
      "theoretical_loss": 3.4711160759916355,
      "tokens_seen": 1916600320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042343029087261785,
      "loss": 1.3724,
      "theoretical_loss": 3.471106178782012,
      "tokens_seen": 1916665856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234102306920762,
      "loss": 1.3469,
      "theoretical_loss": 3.4710962820055475,
      "tokens_seen": 1916731392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042339017051153463,
      "loss": 1.301,
      "theoretical_loss": 3.4710863856622076,
      "tokens_seen": 1916796928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423370110330993,
      "loss": 1.3064,
      "theoretical_loss": 3.471076489751958,
      "tokens_seen": 1916862464
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042335005015045135,
      "loss": 1.3603,
      "theoretical_loss": 3.471066594274766,
      "tokens_seen": 1916928000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233299899699097,
      "loss": 1.4267,
      "theoretical_loss": 3.4710566992305973,
      "tokens_seen": 1916993536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042330992978936813,
      "loss": 1.264,
      "theoretical_loss": 3.4710468046194185,
      "tokens_seen": 1917059072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232898696088265,
      "loss": 1.3519,
      "theoretical_loss": 3.471036910441195,
      "tokens_seen": 1917124608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042326980942828486,
      "loss": 1.3924,
      "theoretical_loss": 3.471027016695894,
      "tokens_seen": 1917190144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232497492477432,
      "loss": 1.3352,
      "theoretical_loss": 3.471017123383481,
      "tokens_seen": 1917255680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232296890672016,
      "loss": 1.3834,
      "theoretical_loss": 3.4710072305039232,
      "tokens_seen": 1917321216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042320962888666,
      "loss": 1.2678,
      "theoretical_loss": 3.470997338057186,
      "tokens_seen": 1917386752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042318956870611836,
      "loss": 1.3665,
      "theoretical_loss": 3.470987446043236,
      "tokens_seen": 1917452288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231695085255767,
      "loss": 1.445,
      "theoretical_loss": 3.4709775544620394,
      "tokens_seen": 1917517824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231494483450351,
      "loss": 1.3184,
      "theoretical_loss": 3.4709676633135627,
      "tokens_seen": 1917583360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231293881644935,
      "loss": 1.3992,
      "theoretical_loss": 3.470957772597772,
      "tokens_seen": 1917648896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042310932798395186,
      "loss": 1.3922,
      "theoretical_loss": 3.4709478823146336,
      "tokens_seen": 1917714432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230892678034102,
      "loss": 1.4292,
      "theoretical_loss": 3.4709379924641137,
      "tokens_seen": 1917779968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230692076228686,
      "loss": 1.3808,
      "theoretical_loss": 3.470928103046179,
      "tokens_seen": 1917845504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042304914744232695,
      "loss": 1.2741,
      "theoretical_loss": 3.470918214060795,
      "tokens_seen": 1917911040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042302908726178537,
      "loss": 1.2935,
      "theoretical_loss": 3.4709083255079287,
      "tokens_seen": 1917976576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042300902708124373,
      "loss": 1.36,
      "theoretical_loss": 3.4708984373875467,
      "tokens_seen": 1918042112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004229889669007021,
      "loss": 1.4353,
      "theoretical_loss": 3.4708885496996142,
      "tokens_seen": 1918107648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042296890672016045,
      "loss": 1.3358,
      "theoretical_loss": 3.4708786624440986,
      "tokens_seen": 1918173184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042294884653961887,
      "loss": 1.3376,
      "theoretical_loss": 3.470868775620965,
      "tokens_seen": 1918238720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042292878635907723,
      "loss": 1.3963,
      "theoretical_loss": 3.470858889230181,
      "tokens_seen": 1918304256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004229087261785356,
      "loss": 1.3852,
      "theoretical_loss": 3.470849003271712,
      "tokens_seen": 1918369792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042288866599799396,
      "loss": 1.3698,
      "theoretical_loss": 3.470839117745525,
      "tokens_seen": 1918435328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228686058174523,
      "loss": 1.3956,
      "theoretical_loss": 3.470829232651586,
      "tokens_seen": 1918500864
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042284854563691074,
      "loss": 1.2971,
      "theoretical_loss": 3.4708193479898615,
      "tokens_seen": 1918566400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228284854563691,
      "loss": 1.4123,
      "theoretical_loss": 3.4708094637603177,
      "tokens_seen": 1918631936
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 699453,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3871700763702393,
      "objective/train/theoretical_loss": 3.4707995799629208,
      "objective/train/tokens_used": 289156576,
      "theoretical_loss": 3.4707995799629208,
      "tokens_seen": 1918697472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042280842527582746,
      "loss": 1.3768,
      "theoretical_loss": 3.4707995799629208,
      "tokens_seen": 1918697472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227883650952858,
      "loss": 1.3481,
      "theoretical_loss": 3.4707896965976373,
      "tokens_seen": 1918763008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042276830491474424,
      "loss": 1.3204,
      "theoretical_loss": 3.470779813664434,
      "tokens_seen": 1918828544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227482447342026,
      "loss": 1.325,
      "theoretical_loss": 3.4707699311632765,
      "tokens_seen": 1918894080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042272818455366096,
      "loss": 1.3812,
      "theoretical_loss": 3.4707600490941313,
      "tokens_seen": 1918959616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227081243731193,
      "loss": 1.3722,
      "theoretical_loss": 3.470750167456965,
      "tokens_seen": 1919025152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226880641925777,
      "loss": 1.3694,
      "theoretical_loss": 3.4707402862517442,
      "tokens_seen": 1919090688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042266800401203616,
      "loss": 1.4158,
      "theoretical_loss": 3.4707304054784345,
      "tokens_seen": 1919156224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226479438314945,
      "loss": 1.3443,
      "theoretical_loss": 3.470720525137003,
      "tokens_seen": 1919221760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226278836509529,
      "loss": 1.3175,
      "theoretical_loss": 3.4707106452274155,
      "tokens_seen": 1919287296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042260782347041125,
      "loss": 1.4202,
      "theoretical_loss": 3.4707007657496387,
      "tokens_seen": 1919352832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042258776328986966,
      "loss": 1.388,
      "theoretical_loss": 3.4706908867036397,
      "tokens_seen": 1919418368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422567703109328,
      "loss": 1.3134,
      "theoretical_loss": 3.4706810080893833,
      "tokens_seen": 1919483904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225476429287864,
      "loss": 1.2193,
      "theoretical_loss": 3.4706711299068376,
      "tokens_seen": 1919549440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042252758274824475,
      "loss": 1.3715,
      "theoretical_loss": 3.4706612521559674,
      "tokens_seen": 1919614976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042250752256770317,
      "loss": 1.3441,
      "theoretical_loss": 3.47065137483674,
      "tokens_seen": 1919680512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042248746238716153,
      "loss": 1.2731,
      "theoretical_loss": 3.4706414979491216,
      "tokens_seen": 1919746048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224674022066199,
      "loss": 1.4185,
      "theoretical_loss": 3.4706316214930792,
      "tokens_seen": 1919811584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042244734202607825,
      "loss": 1.2818,
      "theoretical_loss": 3.470621745468578,
      "tokens_seen": 1919877120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224272818455366,
      "loss": 1.374,
      "theoretical_loss": 3.4706118698755857,
      "tokens_seen": 1919942656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042240722166499503,
      "loss": 1.3041,
      "theoretical_loss": 3.470601994714068,
      "tokens_seen": 1920008192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223871614844534,
      "loss": 1.377,
      "theoretical_loss": 3.4705921199839906,
      "tokens_seen": 1920073728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042236710130391176,
      "loss": 1.4402,
      "theoretical_loss": 3.4705822456853213,
      "tokens_seen": 1920139264
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223470411233701,
      "loss": 1.3111,
      "theoretical_loss": 3.4705723718180264,
      "tokens_seen": 1920204800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042232698094282853,
      "loss": 1.3272,
      "theoretical_loss": 3.4705624983820718,
      "tokens_seen": 1920270336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223069207622869,
      "loss": 1.3412,
      "theoretical_loss": 3.4705526253774233,
      "tokens_seen": 1920335872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042228686058174526,
      "loss": 1.259,
      "theoretical_loss": 3.470542752804049,
      "tokens_seen": 1920401408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222668004012036,
      "loss": 1.1954,
      "theoretical_loss": 3.4705328806619136,
      "tokens_seen": 1920466944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422246740220662,
      "loss": 1.2623,
      "theoretical_loss": 3.470523008950985,
      "tokens_seen": 1920532480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222266800401204,
      "loss": 1.4743,
      "theoretical_loss": 3.470513137671229,
      "tokens_seen": 1920598016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042220661985957876,
      "loss": 1.2877,
      "theoretical_loss": 3.470503266822612,
      "tokens_seen": 1920663552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221865596790371,
      "loss": 1.4402,
      "theoretical_loss": 3.4704933964051,
      "tokens_seen": 1920729088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221664994984955,
      "loss": 1.4002,
      "theoretical_loss": 3.470483526418661,
      "tokens_seen": 1920794624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221464393179539,
      "loss": 1.2645,
      "theoretical_loss": 3.47047365686326,
      "tokens_seen": 1920860160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042212637913741226,
      "loss": 1.3769,
      "theoretical_loss": 3.470463787738864,
      "tokens_seen": 1920925696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221063189568706,
      "loss": 1.3736,
      "theoretical_loss": 3.470453919045439,
      "tokens_seen": 1920991232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422086258776329,
      "loss": 1.2686,
      "theoretical_loss": 3.4704440507829526,
      "tokens_seen": 1921056768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042206619859578735,
      "loss": 1.3846,
      "theoretical_loss": 3.4704341829513696,
      "tokens_seen": 1921122304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042204613841524577,
      "loss": 1.3452,
      "theoretical_loss": 3.4704243155506584,
      "tokens_seen": 1921187840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042202607823470413,
      "loss": 1.3725,
      "theoretical_loss": 3.470414448580784,
      "tokens_seen": 1921253376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004220060180541625,
      "loss": 1.3418,
      "theoretical_loss": 3.4704045820417138,
      "tokens_seen": 1921318912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042198595787362085,
      "loss": 1.4191,
      "theoretical_loss": 3.470394715933414,
      "tokens_seen": 1921384448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042196589769307927,
      "loss": 1.3412,
      "theoretical_loss": 3.4703848502558508,
      "tokens_seen": 1921449984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042194583751253763,
      "loss": 1.366,
      "theoretical_loss": 3.4703749850089913,
      "tokens_seen": 1921515520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000421925777331996,
      "loss": 1.4247,
      "theoretical_loss": 3.4703651201928016,
      "tokens_seen": 1921581056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042190571715145436,
      "loss": 1.3447,
      "theoretical_loss": 3.470355255807248,
      "tokens_seen": 1921646592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218856569709127,
      "loss": 1.2791,
      "theoretical_loss": 3.4703453918522973,
      "tokens_seen": 1921712128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042186559679037114,
      "loss": 1.3157,
      "theoretical_loss": 3.470335528327916,
      "tokens_seen": 1921777664
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218455366098295,
      "loss": 1.2867,
      "theoretical_loss": 3.470325665234071,
      "tokens_seen": 1921843200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042182547642928786,
      "loss": 1.4806,
      "theoretical_loss": 3.4703158025707284,
      "tokens_seen": 1921908736
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 700743,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3846075534820557,
      "objective/train/theoretical_loss": 3.4703059403378544,
      "objective/train/tokens_used": 292433376,
      "theoretical_loss": 3.4703059403378544,
      "tokens_seen": 1921974272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218054162487462,
      "loss": 1.2258,
      "theoretical_loss": 3.4703059403378544,
      "tokens_seen": 1921974272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042178535606820464,
      "loss": 1.4184,
      "theoretical_loss": 3.4702960785354158,
      "tokens_seen": 1922039808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000421765295887663,
      "loss": 1.3426,
      "theoretical_loss": 3.4702862171633795,
      "tokens_seen": 1922105344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042174523570712136,
      "loss": 1.4345,
      "theoretical_loss": 3.4702763562217123,
      "tokens_seen": 1922170880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004217251755265797,
      "loss": 1.2874,
      "theoretical_loss": 3.47026649571038,
      "tokens_seen": 1922236416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004217051153460381,
      "loss": 1.3703,
      "theoretical_loss": 3.470256635629349,
      "tokens_seen": 1922301952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216850551654965,
      "loss": 1.3356,
      "theoretical_loss": 3.470246775978586,
      "tokens_seen": 1922367488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042166499498495487,
      "loss": 1.3424,
      "theoretical_loss": 3.4702369167580587,
      "tokens_seen": 1922433024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042164493480441323,
      "loss": 1.3373,
      "theoretical_loss": 3.4702270579677323,
      "tokens_seen": 1922498560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216248746238716,
      "loss": 1.3344,
      "theoretical_loss": 3.470217199607574,
      "tokens_seen": 1922564096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042160481444333,
      "loss": 1.3912,
      "theoretical_loss": 3.4702073416775496,
      "tokens_seen": 1922629632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042158475426278837,
      "loss": 1.496,
      "theoretical_loss": 3.470197484177627,
      "tokens_seen": 1922695168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042156469408224673,
      "loss": 1.318,
      "theoretical_loss": 3.470187627107772,
      "tokens_seen": 1922760704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215446339017051,
      "loss": 1.3642,
      "theoretical_loss": 3.470177770467951,
      "tokens_seen": 1922826240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215245737211635,
      "loss": 1.341,
      "theoretical_loss": 3.4701679142581305,
      "tokens_seen": 1922891776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215045135406219,
      "loss": 1.172,
      "theoretical_loss": 3.470158058478278,
      "tokens_seen": 1922957312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042148445336008024,
      "loss": 1.2682,
      "theoretical_loss": 3.4701482031283595,
      "tokens_seen": 1923022848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214643931795386,
      "loss": 1.4813,
      "theoretical_loss": 3.4701383482083408,
      "tokens_seen": 1923088384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042144433299899696,
      "loss": 1.4482,
      "theoretical_loss": 3.4701284937181898,
      "tokens_seen": 1923153920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214242728184554,
      "loss": 1.3991,
      "theoretical_loss": 3.4701186396578727,
      "tokens_seen": 1923219456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042140421263791374,
      "loss": 1.3467,
      "theoretical_loss": 3.470108786027356,
      "tokens_seen": 1923284992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213841524573721,
      "loss": 1.3614,
      "theoretical_loss": 3.4700989328266063,
      "tokens_seen": 1923350528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042136409227683046,
      "loss": 1.3238,
      "theoretical_loss": 3.4700890800555895,
      "tokens_seen": 1923416064
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213440320962889,
      "loss": 1.2781,
      "theoretical_loss": 3.470079227714274,
      "tokens_seen": 1923481600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042132397191574724,
      "loss": 1.2222,
      "theoretical_loss": 3.4700693758026246,
      "tokens_seen": 1923547136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213039117352056,
      "loss": 1.3536,
      "theoretical_loss": 3.470059524320609,
      "tokens_seen": 1923612672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042128385155466397,
      "loss": 1.3236,
      "theoretical_loss": 3.4700496732681936,
      "tokens_seen": 1923678208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042126379137412233,
      "loss": 1.2185,
      "theoretical_loss": 3.4700398226453446,
      "tokens_seen": 1923743744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042124373119358075,
      "loss": 1.356,
      "theoretical_loss": 3.470029972452029,
      "tokens_seen": 1923809280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212236710130391,
      "loss": 1.3047,
      "theoretical_loss": 3.4700201226882132,
      "tokens_seen": 1923874816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042120361083249747,
      "loss": 1.2991,
      "theoretical_loss": 3.4700102733538642,
      "tokens_seen": 1923940352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042118355065195583,
      "loss": 1.4352,
      "theoretical_loss": 3.470000424448949,
      "tokens_seen": 1924005888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211634904714143,
      "loss": 1.2492,
      "theoretical_loss": 3.4699905759734335,
      "tokens_seen": 1924071424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042114343029087266,
      "loss": 1.3506,
      "theoretical_loss": 3.4699807279272843,
      "tokens_seen": 1924136960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042112337011033103,
      "loss": 1.3672,
      "theoretical_loss": 3.4699708803104685,
      "tokens_seen": 1924202496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211033099297894,
      "loss": 1.3594,
      "theoretical_loss": 3.469961033122953,
      "tokens_seen": 1924268032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042108324974924775,
      "loss": 1.3249,
      "theoretical_loss": 3.4699511863647032,
      "tokens_seen": 1924333568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042106318956870617,
      "loss": 1.3542,
      "theoretical_loss": 3.469941340035687,
      "tokens_seen": 1924399104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042104312938816453,
      "loss": 1.3987,
      "theoretical_loss": 3.4699314941358708,
      "tokens_seen": 1924464640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004210230692076229,
      "loss": 1.4009,
      "theoretical_loss": 3.4699216486652213,
      "tokens_seen": 1924530176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042100300902708125,
      "loss": 1.298,
      "theoretical_loss": 3.469911803623705,
      "tokens_seen": 1924595712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042098294884653967,
      "loss": 1.3296,
      "theoretical_loss": 3.469901959011289,
      "tokens_seen": 1924661248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042096288866599803,
      "loss": 1.3959,
      "theoretical_loss": 3.4698921148279394,
      "tokens_seen": 1924726784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004209428284854564,
      "loss": 1.3748,
      "theoretical_loss": 3.4698822710736232,
      "tokens_seen": 1924792320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042092276830491476,
      "loss": 1.3945,
      "theoretical_loss": 3.4698724277483066,
      "tokens_seen": 1924857856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004209027081243731,
      "loss": 1.3703,
      "theoretical_loss": 3.469862584851957,
      "tokens_seen": 1924923392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042088264794383154,
      "loss": 1.2939,
      "theoretical_loss": 3.469852742384541,
      "tokens_seen": 1924988928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208625877632899,
      "loss": 1.359,
      "theoretical_loss": 3.469842900346025,
      "tokens_seen": 1925054464
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042084252758274826,
      "loss": 1.3594,
      "theoretical_loss": 3.469833058736376,
      "tokens_seen": 1925120000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208224674022066,
      "loss": 1.4239,
      "theoretical_loss": 3.4698232175555606,
      "tokens_seen": 1925185536
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 701973,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5264232158660889,
      "objective/train/theoretical_loss": 3.469813376803545,
      "objective/train/tokens_used": 295710176,
      "theoretical_loss": 3.469813376803545,
      "tokens_seen": 1925251072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042080240722166504,
      "loss": 1.4383,
      "theoretical_loss": 3.469813376803545,
      "tokens_seen": 1925251072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207823470411234,
      "loss": 1.3312,
      "theoretical_loss": 3.469803536480297,
      "tokens_seen": 1925316608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042076228686058176,
      "loss": 1.2398,
      "theoretical_loss": 3.469793696585782,
      "tokens_seen": 1925382144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207422266800401,
      "loss": 1.3864,
      "theoretical_loss": 3.4697838571199684,
      "tokens_seen": 1925447680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207221664994985,
      "loss": 1.3397,
      "theoretical_loss": 3.469774018082821,
      "tokens_seen": 1925513216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207021063189569,
      "loss": 1.4218,
      "theoretical_loss": 3.4697641794743084,
      "tokens_seen": 1925578752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042068204613841527,
      "loss": 1.3639,
      "theoretical_loss": 3.469754341294396,
      "tokens_seen": 1925644288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042066198595787363,
      "loss": 1.4777,
      "theoretical_loss": 3.469744503543051,
      "tokens_seen": 1925709824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420641925777332,
      "loss": 1.4203,
      "theoretical_loss": 3.46973466622024,
      "tokens_seen": 1925775360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004206218655967904,
      "loss": 1.4475,
      "theoretical_loss": 3.4697248293259304,
      "tokens_seen": 1925840896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042060180541624877,
      "loss": 1.4234,
      "theoretical_loss": 3.469714992860088,
      "tokens_seen": 1925906432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042058174523570713,
      "loss": 1.293,
      "theoretical_loss": 3.4697051568226804,
      "tokens_seen": 1925971968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205616850551655,
      "loss": 1.3534,
      "theoretical_loss": 3.4696953212136736,
      "tokens_seen": 1926037504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205416248746239,
      "loss": 1.3738,
      "theoretical_loss": 3.469685486033035,
      "tokens_seen": 1926103040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205215646940823,
      "loss": 1.3611,
      "theoretical_loss": 3.469675651280731,
      "tokens_seen": 1926168576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042050150451354064,
      "loss": 1.2836,
      "theoretical_loss": 3.4696658169567285,
      "tokens_seen": 1926234112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420481444332999,
      "loss": 1.3201,
      "theoretical_loss": 3.469655983060994,
      "tokens_seen": 1926299648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042046138415245736,
      "loss": 1.4302,
      "theoretical_loss": 3.469646149593495,
      "tokens_seen": 1926365184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204413239719158,
      "loss": 1.3677,
      "theoretical_loss": 3.4696363165541975,
      "tokens_seen": 1926430720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042042126379137414,
      "loss": 1.3277,
      "theoretical_loss": 3.4696264839430686,
      "tokens_seen": 1926496256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204012036108325,
      "loss": 1.2716,
      "theoretical_loss": 3.4696166517600755,
      "tokens_seen": 1926561792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042038114343029086,
      "loss": 1.4919,
      "theoretical_loss": 3.4696068200051844,
      "tokens_seen": 1926627328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004203610832497493,
      "loss": 1.3817,
      "theoretical_loss": 3.469596988678362,
      "tokens_seen": 1926692864
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042034102306920764,
      "loss": 1.3356,
      "theoretical_loss": 3.469587157779576,
      "tokens_seen": 1926758400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420320962888666,
      "loss": 1.3599,
      "theoretical_loss": 3.4695773273087918,
      "tokens_seen": 1926823936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042030090270812437,
      "loss": 1.4011,
      "theoretical_loss": 3.4695674972659774,
      "tokens_seen": 1926889472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042028084252758273,
      "loss": 1.3683,
      "theoretical_loss": 3.469557667651099,
      "tokens_seen": 1926955008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042026078234704115,
      "loss": 1.4777,
      "theoretical_loss": 3.4695478384641243,
      "tokens_seen": 1927020544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202407221664995,
      "loss": 1.3343,
      "theoretical_loss": 3.469538009705019,
      "tokens_seen": 1927086080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042022066198595787,
      "loss": 1.2653,
      "theoretical_loss": 3.46952818137375,
      "tokens_seen": 1927151616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042020060180541623,
      "loss": 1.498,
      "theoretical_loss": 3.4695183534702854,
      "tokens_seen": 1927217152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042018054162487465,
      "loss": 1.3042,
      "theoretical_loss": 3.4695085259945904,
      "tokens_seen": 1927282688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420160481444333,
      "loss": 1.4258,
      "theoretical_loss": 3.4694986989466328,
      "tokens_seen": 1927348224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201404212637914,
      "loss": 1.3078,
      "theoretical_loss": 3.4694888723263793,
      "tokens_seen": 1927413760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042012036108324974,
      "loss": 1.3834,
      "theoretical_loss": 3.469479046133797,
      "tokens_seen": 1927479296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201003009027081,
      "loss": 1.4132,
      "theoretical_loss": 3.469469220368852,
      "tokens_seen": 1927544832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200802407221665,
      "loss": 1.4725,
      "theoretical_loss": 3.4694593950315116,
      "tokens_seen": 1927610368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200601805416249,
      "loss": 1.4918,
      "theoretical_loss": 3.4694495701217427,
      "tokens_seen": 1927675904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042004012036108324,
      "loss": 1.3441,
      "theoretical_loss": 3.4694397456395123,
      "tokens_seen": 1927741440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200200601805416,
      "loss": 1.3741,
      "theoretical_loss": 3.469429921584787,
      "tokens_seen": 1927806976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042,
      "loss": 1.5306,
      "theoretical_loss": 3.469420097957533,
      "tokens_seen": 1927872512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004199799398194584,
      "loss": 1.2331,
      "theoretical_loss": 3.4694102747577187,
      "tokens_seen": 1927938048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041995987963891674,
      "loss": 1.3393,
      "theoretical_loss": 3.46940045198531,
      "tokens_seen": 1928003584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004199398194583751,
      "loss": 1.3895,
      "theoretical_loss": 3.4693906296402734,
      "tokens_seen": 1928069120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041991975927783347,
      "loss": 1.416,
      "theoretical_loss": 3.469380807722577,
      "tokens_seen": 1928134656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198996990972919,
      "loss": 1.3692,
      "theoretical_loss": 3.469370986232186,
      "tokens_seen": 1928200192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041987963891675025,
      "loss": 1.3991,
      "theoretical_loss": 3.4693611651690692,
      "tokens_seen": 1928265728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198595787362086,
      "loss": 1.2821,
      "theoretical_loss": 3.4693513445331923,
      "tokens_seen": 1928331264
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041983951855566697,
      "loss": 1.4471,
      "theoretical_loss": 3.4693415243245225,
      "tokens_seen": 1928396800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198194583751254,
      "loss": 1.37,
      "theoretical_loss": 3.4693317045430265,
      "tokens_seen": 1928462336
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 703126,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3585114479064941,
      "objective/train/theoretical_loss": 3.4693218851886716,
      "objective/train/tokens_used": 298986976,
      "theoretical_loss": 3.4693218851886716,
      "tokens_seen": 1928527872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041979939819458375,
      "loss": 1.3773,
      "theoretical_loss": 3.4693218851886716,
      "tokens_seen": 1928527872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004197793380140421,
      "loss": 1.3453,
      "theoretical_loss": 3.4693120662614243,
      "tokens_seen": 1928593408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041975927783350047,
      "loss": 1.295,
      "theoretical_loss": 3.4693022477612514,
      "tokens_seen": 1928658944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041973921765295883,
      "loss": 1.395,
      "theoretical_loss": 3.4692924296881205,
      "tokens_seen": 1928724480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041971915747241725,
      "loss": 1.4105,
      "theoretical_loss": 3.469282612041998,
      "tokens_seen": 1928790016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196990972918756,
      "loss": 1.35,
      "theoretical_loss": 3.469272794822851,
      "tokens_seen": 1928855552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419679037111334,
      "loss": 1.319,
      "theoretical_loss": 3.4692629780306463,
      "tokens_seen": 1928921088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196589769307924,
      "loss": 1.4431,
      "theoretical_loss": 3.4692531616653506,
      "tokens_seen": 1928986624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196389167502508,
      "loss": 1.425,
      "theoretical_loss": 3.469243345726931,
      "tokens_seen": 1929052160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041961885656970917,
      "loss": 1.3815,
      "theoretical_loss": 3.469233530215355,
      "tokens_seen": 1929117696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041959879638916753,
      "loss": 1.2496,
      "theoretical_loss": 3.4692237151305894,
      "tokens_seen": 1929183232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195787362086259,
      "loss": 1.3619,
      "theoretical_loss": 3.4692139004726004,
      "tokens_seen": 1929248768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195586760280843,
      "loss": 1.275,
      "theoretical_loss": 3.469204086241355,
      "tokens_seen": 1929314304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195386158475427,
      "loss": 1.376,
      "theoretical_loss": 3.4691942724368214,
      "tokens_seen": 1929379840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041951855566700104,
      "loss": 1.359,
      "theoretical_loss": 3.469184459058965,
      "tokens_seen": 1929445376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194984954864594,
      "loss": 1.4429,
      "theoretical_loss": 3.4691746461077537,
      "tokens_seen": 1929510912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041947843530591776,
      "loss": 1.5078,
      "theoretical_loss": 3.4691648335831538,
      "tokens_seen": 1929576448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194583751253762,
      "loss": 1.3917,
      "theoretical_loss": 3.4691550214851334,
      "tokens_seen": 1929641984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041943831494483454,
      "loss": 1.4125,
      "theoretical_loss": 3.469145209813658,
      "tokens_seen": 1929707520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194182547642929,
      "loss": 1.4025,
      "theoretical_loss": 3.469135398568696,
      "tokens_seen": 1929773056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041939819458375126,
      "loss": 1.363,
      "theoretical_loss": 3.469125587750213,
      "tokens_seen": 1929838592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193781344032097,
      "loss": 1.4099,
      "theoretical_loss": 3.469115777358177,
      "tokens_seen": 1929904128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041935807422266804,
      "loss": 1.4516,
      "theoretical_loss": 3.4691059673925544,
      "tokens_seen": 1929969664
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193380140421264,
      "loss": 1.3919,
      "theoretical_loss": 3.4690961578533126,
      "tokens_seen": 1930035200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041931795386158477,
      "loss": 1.3842,
      "theoretical_loss": 3.4690863487404187,
      "tokens_seen": 1930100736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041929789368104313,
      "loss": 1.4357,
      "theoretical_loss": 3.4690765400538393,
      "tokens_seen": 1930166272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041927783350050155,
      "loss": 1.3845,
      "theoretical_loss": 3.469066731793541,
      "tokens_seen": 1930231808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004192577733199599,
      "loss": 1.3076,
      "theoretical_loss": 3.4690569239594917,
      "tokens_seen": 1930297344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041923771313941827,
      "loss": 1.381,
      "theoretical_loss": 3.469047116551658,
      "tokens_seen": 1930362880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041921765295887663,
      "loss": 1.3423,
      "theoretical_loss": 3.469037309570007,
      "tokens_seen": 1930428416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041919759277833505,
      "loss": 1.4357,
      "theoretical_loss": 3.4690275030145052,
      "tokens_seen": 1930493952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191775325977934,
      "loss": 1.3848,
      "theoretical_loss": 3.4690176968851203,
      "tokens_seen": 1930559488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191574724172518,
      "loss": 1.4606,
      "theoretical_loss": 3.4690078911818194,
      "tokens_seen": 1930625024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041913741223671014,
      "loss": 1.3832,
      "theoretical_loss": 3.468998085904569,
      "tokens_seen": 1930690560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191173520561685,
      "loss": 1.396,
      "theoretical_loss": 3.4689882810533366,
      "tokens_seen": 1930756096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190972918756269,
      "loss": 1.3276,
      "theoretical_loss": 3.4689784766280884,
      "tokens_seen": 1930821632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190772316950853,
      "loss": 1.5123,
      "theoretical_loss": 3.4689686726287925,
      "tokens_seen": 1930887168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041905717151454364,
      "loss": 1.4276,
      "theoretical_loss": 3.468958869055415,
      "tokens_seen": 1930952704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419037111334002,
      "loss": 1.3639,
      "theoretical_loss": 3.468949065907924,
      "tokens_seen": 1931018240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190170511534604,
      "loss": 1.4026,
      "theoretical_loss": 3.468939263186285,
      "tokens_seen": 1931083776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189969909729188,
      "loss": 1.2994,
      "theoretical_loss": 3.4689294608904664,
      "tokens_seen": 1931149312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041897693079237714,
      "loss": 1.2572,
      "theoretical_loss": 3.468919659020435,
      "tokens_seen": 1931214848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189568706118355,
      "loss": 1.3661,
      "theoretical_loss": 3.4689098575761577,
      "tokens_seen": 1931280384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041893681043129387,
      "loss": 1.4139,
      "theoretical_loss": 3.4689000565576014,
      "tokens_seen": 1931345920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189167502507523,
      "loss": 1.3471,
      "theoretical_loss": 3.468890255964733,
      "tokens_seen": 1931411456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041889669007021065,
      "loss": 1.3739,
      "theoretical_loss": 3.46888045579752,
      "tokens_seen": 1931476992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418876629889669,
      "loss": 1.3559,
      "theoretical_loss": 3.46887065605593,
      "tokens_seen": 1931542528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041885656970912737,
      "loss": 1.3802,
      "theoretical_loss": 3.468860856739929,
      "tokens_seen": 1931608064
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188365095285858,
      "loss": 1.3959,
      "theoretical_loss": 3.468851057849484,
      "tokens_seen": 1931673600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041881644934804415,
      "loss": 1.4677,
      "theoretical_loss": 3.468841259384563,
      "tokens_seen": 1931739136
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 704289,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5385944843292236,
      "objective/train/theoretical_loss": 3.4688314613451325,
      "objective/train/tokens_used": 302263776,
      "theoretical_loss": 3.4688314613451325,
      "tokens_seen": 1931804672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187963891675025,
      "loss": 1.4941,
      "theoretical_loss": 3.4688314613451325,
      "tokens_seen": 1931804672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187763289869609,
      "loss": 1.4521,
      "theoretical_loss": 3.46882166373116,
      "tokens_seen": 1931870208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041875626880641924,
      "loss": 1.3902,
      "theoretical_loss": 3.4688118665426124,
      "tokens_seen": 1931935744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041873620862587765,
      "loss": 1.4057,
      "theoretical_loss": 3.4688020697794566,
      "tokens_seen": 1932001280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418716148445336,
      "loss": 1.272,
      "theoretical_loss": 3.46879227344166,
      "tokens_seen": 1932066816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186960882647944,
      "loss": 1.3272,
      "theoretical_loss": 3.4687824775291896,
      "tokens_seen": 1932132352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041867602808425274,
      "loss": 1.4092,
      "theoretical_loss": 3.4687726820420117,
      "tokens_seen": 1932197888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041865596790371115,
      "loss": 1.344,
      "theoretical_loss": 3.4687628869800946,
      "tokens_seen": 1932263424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186359077231695,
      "loss": 1.1831,
      "theoretical_loss": 3.4687530923434053,
      "tokens_seen": 1932328960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186158475426279,
      "loss": 1.4457,
      "theoretical_loss": 3.46874329813191,
      "tokens_seen": 1932394496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041859578736208624,
      "loss": 1.3865,
      "theoretical_loss": 3.468733504345577,
      "tokens_seen": 1932460032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185757271815446,
      "loss": 1.3186,
      "theoretical_loss": 3.468723710984373,
      "tokens_seen": 1932525568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418555667001003,
      "loss": 1.3784,
      "theoretical_loss": 3.4687139180482647,
      "tokens_seen": 1932591104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185356068204614,
      "loss": 1.2988,
      "theoretical_loss": 3.4687041255372195,
      "tokens_seen": 1932656640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041851554663991974,
      "loss": 1.314,
      "theoretical_loss": 3.4686943334512046,
      "tokens_seen": 1932722176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184954864593781,
      "loss": 1.3587,
      "theoretical_loss": 3.4686845417901866,
      "tokens_seen": 1932787712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184754262788365,
      "loss": 1.434,
      "theoretical_loss": 3.4686747505541335,
      "tokens_seen": 1932853248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184553660982949,
      "loss": 1.3355,
      "theoretical_loss": 3.4686649597430126,
      "tokens_seen": 1932918784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041843530591775325,
      "loss": 1.3835,
      "theoretical_loss": 3.46865516935679,
      "tokens_seen": 1932984320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184152457372116,
      "loss": 1.3459,
      "theoretical_loss": 3.4686453793954337,
      "tokens_seen": 1933049856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041839518555666997,
      "loss": 1.3748,
      "theoretical_loss": 3.46863558985891,
      "tokens_seen": 1933115392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183751253761284,
      "loss": 1.3542,
      "theoretical_loss": 3.468625800747187,
      "tokens_seen": 1933180928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041835506519558675,
      "loss": 1.4299,
      "theoretical_loss": 3.4686160120602314,
      "tokens_seen": 1933246464
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183350050150451,
      "loss": 1.3459,
      "theoretical_loss": 3.4686062237980106,
      "tokens_seen": 1933312000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183149448345035,
      "loss": 1.4708,
      "theoretical_loss": 3.468596435960491,
      "tokens_seen": 1933377536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182948846539619,
      "loss": 1.2687,
      "theoretical_loss": 3.468586648547641,
      "tokens_seen": 1933443072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041827482447342025,
      "loss": 1.3396,
      "theoretical_loss": 3.468576861559427,
      "tokens_seen": 1933508608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182547642928786,
      "loss": 1.3175,
      "theoretical_loss": 3.468567074995817,
      "tokens_seen": 1933574144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418234704112337,
      "loss": 1.3236,
      "theoretical_loss": 3.4685572888567764,
      "tokens_seen": 1933639680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182146439317954,
      "loss": 1.3742,
      "theoretical_loss": 3.4685475031422746,
      "tokens_seen": 1933705216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041819458375125376,
      "loss": 1.3708,
      "theoretical_loss": 3.468537717852277,
      "tokens_seen": 1933770752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181745235707121,
      "loss": 1.4306,
      "theoretical_loss": 3.4685279329867518,
      "tokens_seen": 1933836288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041815446339017054,
      "loss": 1.3214,
      "theoretical_loss": 3.468518148545666,
      "tokens_seen": 1933901824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181344032096289,
      "loss": 1.396,
      "theoretical_loss": 3.468508364528987,
      "tokens_seen": 1933967360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181143430290873,
      "loss": 1.4688,
      "theoretical_loss": 3.4684985809366813,
      "tokens_seen": 1934032896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180942828485457,
      "loss": 1.4803,
      "theoretical_loss": 3.4684887977687167,
      "tokens_seen": 1934098432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041807422266800404,
      "loss": 1.3986,
      "theoretical_loss": 3.4684790150250606,
      "tokens_seen": 1934163968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180541624874624,
      "loss": 1.3921,
      "theoretical_loss": 3.4684692327056794,
      "tokens_seen": 1934229504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180341023069208,
      "loss": 1.2771,
      "theoretical_loss": 3.468459450810541,
      "tokens_seen": 1934295040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180140421263792,
      "loss": 1.322,
      "theoretical_loss": 3.4684496693396127,
      "tokens_seen": 1934360576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041799398194583754,
      "loss": 1.4031,
      "theoretical_loss": 3.4684398882928615,
      "tokens_seen": 1934426112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179739217652959,
      "loss": 1.3434,
      "theoretical_loss": 3.4684301076702546,
      "tokens_seen": 1934491648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041795386158475427,
      "loss": 1.3041,
      "theoretical_loss": 3.468420327471759,
      "tokens_seen": 1934557184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179338014042127,
      "loss": 1.4664,
      "theoretical_loss": 3.4684105476973426,
      "tokens_seen": 1934622720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041791374122367105,
      "loss": 1.3646,
      "theoretical_loss": 3.4684007683469718,
      "tokens_seen": 1934688256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178936810431294,
      "loss": 1.3553,
      "theoretical_loss": 3.468390989420615,
      "tokens_seen": 1934753792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041787362086258777,
      "loss": 1.3907,
      "theoretical_loss": 3.4683812109182384,
      "tokens_seen": 1934819328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178535606820462,
      "loss": 1.286,
      "theoretical_loss": 3.4683714328398096,
      "tokens_seen": 1934884864
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041783350050150455,
      "loss": 1.314,
      "theoretical_loss": 3.4683616551852956,
      "tokens_seen": 1934950400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178134403209629,
      "loss": 1.4161,
      "theoretical_loss": 3.4683518779546647,
      "tokens_seen": 1935015936
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 705622,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3978188037872314,
      "objective/train/theoretical_loss": 3.4683421011478828,
      "objective/train/tokens_used": 305540576,
      "theoretical_loss": 3.4683421011478828,
      "tokens_seen": 1935081472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177933801404213,
      "loss": 1.3751,
      "theoretical_loss": 3.4683421011478828,
      "tokens_seen": 1935081472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041777331995987964,
      "loss": 1.2738,
      "theoretical_loss": 3.468332324764918,
      "tokens_seen": 1935147008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041775325977933805,
      "loss": 1.3688,
      "theoretical_loss": 3.4683225488057374,
      "tokens_seen": 1935212544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177331995987964,
      "loss": 1.3081,
      "theoretical_loss": 3.468312773270308,
      "tokens_seen": 1935278080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177131394182548,
      "loss": 1.3663,
      "theoretical_loss": 3.4683029981585975,
      "tokens_seen": 1935343616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041769307923771314,
      "loss": 1.432,
      "theoretical_loss": 3.468293223470573,
      "tokens_seen": 1935409152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041767301905717156,
      "loss": 1.2619,
      "theoretical_loss": 3.4682834492062016,
      "tokens_seen": 1935474688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176529588766299,
      "loss": 1.2516,
      "theoretical_loss": 3.4682736753654515,
      "tokens_seen": 1935540224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176328986960883,
      "loss": 1.3214,
      "theoretical_loss": 3.468263901948289,
      "tokens_seen": 1935605760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041761283851554664,
      "loss": 1.4344,
      "theoretical_loss": 3.468254128954681,
      "tokens_seen": 1935671296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417592778335005,
      "loss": 1.344,
      "theoretical_loss": 3.468244356384596,
      "tokens_seen": 1935736832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175727181544634,
      "loss": 1.3614,
      "theoretical_loss": 3.468234584238001,
      "tokens_seen": 1935802368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175526579739218,
      "loss": 1.2512,
      "theoretical_loss": 3.468224812514863,
      "tokens_seen": 1935867904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041753259779338015,
      "loss": 1.3703,
      "theoretical_loss": 3.4682150412151493,
      "tokens_seen": 1935933440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175125376128385,
      "loss": 1.2505,
      "theoretical_loss": 3.468205270338828,
      "tokens_seen": 1935998976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174924774322969,
      "loss": 1.3161,
      "theoretical_loss": 3.468195499885865,
      "tokens_seen": 1936064512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174724172517553,
      "loss": 1.3768,
      "theoretical_loss": 3.468185729856229,
      "tokens_seen": 1936130048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041745235707121365,
      "loss": 1.3987,
      "theoretical_loss": 3.4681759602498863,
      "tokens_seen": 1936195584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417432296890672,
      "loss": 1.4023,
      "theoretical_loss": 3.4681661910668047,
      "tokens_seen": 1936261120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041741223671013037,
      "loss": 1.3435,
      "theoretical_loss": 3.4681564223069516,
      "tokens_seen": 1936326656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173921765295888,
      "loss": 1.3387,
      "theoretical_loss": 3.4681466539702948,
      "tokens_seen": 1936392192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041737211634904715,
      "loss": 1.3121,
      "theoretical_loss": 3.4681368860568,
      "tokens_seen": 1936457728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173520561685055,
      "loss": 1.3161,
      "theoretical_loss": 3.4681271185664366,
      "tokens_seen": 1936523264
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173319959879639,
      "loss": 1.4292,
      "theoretical_loss": 3.4681173514991706,
      "tokens_seen": 1936588800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173119358074223,
      "loss": 1.3674,
      "theoretical_loss": 3.4681075848549696,
      "tokens_seen": 1936654336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041729187562688065,
      "loss": 1.2904,
      "theoretical_loss": 3.4680978186338014,
      "tokens_seen": 1936719872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417271815446339,
      "loss": 1.4372,
      "theoretical_loss": 3.4680880528356326,
      "tokens_seen": 1936785408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172517552657974,
      "loss": 1.3775,
      "theoretical_loss": 3.4680782874604317,
      "tokens_seen": 1936850944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172316950852558,
      "loss": 1.3468,
      "theoretical_loss": 3.4680685225081653,
      "tokens_seen": 1936916480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041721163490471416,
      "loss": 1.3314,
      "theoretical_loss": 3.4680587579788007,
      "tokens_seen": 1936982016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171915747241725,
      "loss": 1.4066,
      "theoretical_loss": 3.468048993872306,
      "tokens_seen": 1937047552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171715145436309,
      "loss": 1.3649,
      "theoretical_loss": 3.468039230188647,
      "tokens_seen": 1937113088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041715145436308924,
      "loss": 1.4352,
      "theoretical_loss": 3.468029466927793,
      "tokens_seen": 1937178624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041713139418254766,
      "loss": 1.2309,
      "theoretical_loss": 3.46801970408971,
      "tokens_seen": 1937244160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417111334002006,
      "loss": 1.4011,
      "theoretical_loss": 3.468009941674366,
      "tokens_seen": 1937309696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170912738214644,
      "loss": 1.4143,
      "theoretical_loss": 3.468000179681728,
      "tokens_seen": 1937375232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041707121364092275,
      "loss": 1.3763,
      "theoretical_loss": 3.4679904181117642,
      "tokens_seen": 1937440768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041705115346038116,
      "loss": 1.4496,
      "theoretical_loss": 3.467980656964442,
      "tokens_seen": 1937506304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170310932798395,
      "loss": 1.3204,
      "theoretical_loss": 3.467970896239727,
      "tokens_seen": 1937571840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170110330992979,
      "loss": 1.2882,
      "theoretical_loss": 3.467961135937589,
      "tokens_seen": 1937637376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041699097291875625,
      "loss": 1.4697,
      "theoretical_loss": 3.467951376057994,
      "tokens_seen": 1937702912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169709127382146,
      "loss": 1.4418,
      "theoretical_loss": 3.4679416166009096,
      "tokens_seen": 1937768448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041695085255767303,
      "loss": 1.3643,
      "theoretical_loss": 3.467931857566304,
      "tokens_seen": 1937833984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169307923771314,
      "loss": 1.277,
      "theoretical_loss": 3.4679220989541433,
      "tokens_seen": 1937899520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041691073219658975,
      "loss": 1.418,
      "theoretical_loss": 3.4679123407643955,
      "tokens_seen": 1937965056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168906720160481,
      "loss": 1.3672,
      "theoretical_loss": 3.467902582997029,
      "tokens_seen": 1938030592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041687061183550653,
      "loss": 1.383,
      "theoretical_loss": 3.46789282565201,
      "tokens_seen": 1938096128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168505516549649,
      "loss": 1.2926,
      "theoretical_loss": 3.467883068729306,
      "tokens_seen": 1938161664
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041683049147442326,
      "loss": 1.4197,
      "theoretical_loss": 3.467873312228885,
      "tokens_seen": 1938227200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168104312938816,
      "loss": 1.2633,
      "theoretical_loss": 3.4678635561507143,
      "tokens_seen": 1938292736
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 706849,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.241760015487671,
      "objective/train/theoretical_loss": 3.467853800494761,
      "objective/train/tokens_used": 308817376,
      "theoretical_loss": 3.467853800494761,
      "tokens_seen": 1938358272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041679037111334,
      "loss": 1.3329,
      "theoretical_loss": 3.467853800494761,
      "tokens_seen": 1938358272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167703109327984,
      "loss": 1.3804,
      "theoretical_loss": 3.467844045260993,
      "tokens_seen": 1938423808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041675025075225676,
      "loss": 1.4074,
      "theoretical_loss": 3.4678342904493777,
      "tokens_seen": 1938489344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167301905717151,
      "loss": 1.2663,
      "theoretical_loss": 3.467824536059882,
      "tokens_seen": 1938554880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167101303911735,
      "loss": 1.3218,
      "theoretical_loss": 3.4678147820924745,
      "tokens_seen": 1938620416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004166900702106319,
      "loss": 1.425,
      "theoretical_loss": 3.4678050285471214,
      "tokens_seen": 1938685952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041667001003009026,
      "loss": 1.3919,
      "theoretical_loss": 3.467795275423791,
      "tokens_seen": 1938751488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166499498495487,
      "loss": 1.3897,
      "theoretical_loss": 3.4677855227224508,
      "tokens_seen": 1938817024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041662988966900704,
      "loss": 1.3555,
      "theoretical_loss": 3.4677757704430676,
      "tokens_seen": 1938882560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166098294884654,
      "loss": 1.2896,
      "theoretical_loss": 3.4677660185856096,
      "tokens_seen": 1938948096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165897693079238,
      "loss": 1.3669,
      "theoretical_loss": 3.467756267150044,
      "tokens_seen": 1939013632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165697091273822,
      "loss": 1.3284,
      "theoretical_loss": 3.467746516136337,
      "tokens_seen": 1939079168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041654964894684055,
      "loss": 1.4591,
      "theoretical_loss": 3.467736765544459,
      "tokens_seen": 1939144704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165295887662989,
      "loss": 1.2747,
      "theoretical_loss": 3.467727015374375,
      "tokens_seen": 1939210240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165095285857573,
      "loss": 1.3812,
      "theoretical_loss": 3.467717265626054,
      "tokens_seen": 1939275776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164894684052157,
      "loss": 1.2393,
      "theoretical_loss": 3.4677075162994617,
      "tokens_seen": 1939341312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041646940822467405,
      "loss": 1.2968,
      "theoretical_loss": 3.4676977673945677,
      "tokens_seen": 1939406848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164493480441324,
      "loss": 1.4291,
      "theoretical_loss": 3.467688018911338,
      "tokens_seen": 1939472384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164292878635908,
      "loss": 1.287,
      "theoretical_loss": 3.467678270849741,
      "tokens_seen": 1939537920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164092276830492,
      "loss": 1.3923,
      "theoretical_loss": 3.467668523209744,
      "tokens_seen": 1939603456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041638916750250755,
      "loss": 1.4175,
      "theoretical_loss": 3.467658775991315,
      "tokens_seen": 1939668992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163691073219659,
      "loss": 1.3929,
      "theoretical_loss": 3.46764902919442,
      "tokens_seen": 1939734528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163490471414243,
      "loss": 1.3534,
      "theoretical_loss": 3.4676392828190274,
      "tokens_seen": 1939800064
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163289869608827,
      "loss": 1.3307,
      "theoretical_loss": 3.4676295368651053,
      "tokens_seen": 1939865600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041630892678034105,
      "loss": 1.4219,
      "theoretical_loss": 3.4676197913326208,
      "tokens_seen": 1939931136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162888665997994,
      "loss": 1.3602,
      "theoretical_loss": 3.467610046221541,
      "tokens_seen": 1939996672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162688064192578,
      "loss": 1.3338,
      "theoretical_loss": 3.467600301531834,
      "tokens_seen": 1940062208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162487462387162,
      "loss": 1.3719,
      "theoretical_loss": 3.467590557263467,
      "tokens_seen": 1940127744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041622868605817456,
      "loss": 1.3396,
      "theoretical_loss": 3.467580813416408,
      "tokens_seen": 1940193280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162086258776329,
      "loss": 1.3609,
      "theoretical_loss": 3.4675710699906244,
      "tokens_seen": 1940258816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161885656970913,
      "loss": 1.3998,
      "theoretical_loss": 3.4675613269860834,
      "tokens_seen": 1940324352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041616850551654964,
      "loss": 1.4225,
      "theoretical_loss": 3.4675515844027522,
      "tokens_seen": 1940389888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041614844533600806,
      "loss": 1.3909,
      "theoretical_loss": 3.4675418422405997,
      "tokens_seen": 1940455424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161283851554664,
      "loss": 1.3074,
      "theoretical_loss": 3.467532100499593,
      "tokens_seen": 1940520960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161083249749248,
      "loss": 1.568,
      "theoretical_loss": 3.4675223591796986,
      "tokens_seen": 1940586496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041608826479438315,
      "loss": 1.416,
      "theoretical_loss": 3.4675126182808853,
      "tokens_seen": 1940652032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041606820461384156,
      "loss": 1.3091,
      "theoretical_loss": 3.4675028778031205,
      "tokens_seen": 1940717568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160481444332999,
      "loss": 1.3716,
      "theoretical_loss": 3.467493137746371,
      "tokens_seen": 1940783104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160280842527583,
      "loss": 1.4581,
      "theoretical_loss": 3.4674833981106055,
      "tokens_seen": 1940848640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041600802407221665,
      "loss": 1.3903,
      "theoretical_loss": 3.4674736588957904,
      "tokens_seen": 1940914176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415987963891675,
      "loss": 1.3653,
      "theoretical_loss": 3.4674639201018946,
      "tokens_seen": 1940979712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041596790371113343,
      "loss": 1.3429,
      "theoretical_loss": 3.4674541817288844,
      "tokens_seen": 1941045248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004159478435305918,
      "loss": 1.2982,
      "theoretical_loss": 3.4674444437767282,
      "tokens_seen": 1941110784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041592778335005015,
      "loss": 1.4436,
      "theoretical_loss": 3.4674347062453936,
      "tokens_seen": 1941176320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004159077231695085,
      "loss": 1.2485,
      "theoretical_loss": 3.467424969134848,
      "tokens_seen": 1941241856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041588766298896693,
      "loss": 1.4267,
      "theoretical_loss": 3.4674152324450587,
      "tokens_seen": 1941307392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158676028084253,
      "loss": 1.3311,
      "theoretical_loss": 3.467405496175994,
      "tokens_seen": 1941372928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041584754262788366,
      "loss": 1.281,
      "theoretical_loss": 3.467395760327621,
      "tokens_seen": 1941438464
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415827482447342,
      "loss": 1.3445,
      "theoretical_loss": 3.4673860248999073,
      "tokens_seen": 1941504000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158074222668004,
      "loss": 1.3291,
      "theoretical_loss": 3.467376289892821,
      "tokens_seen": 1941569536
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 708064,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3450722694396973,
      "objective/train/theoretical_loss": 3.4673665553063295,
      "objective/train/tokens_used": 312094176,
      "theoretical_loss": 3.4673665553063295,
      "tokens_seen": 1941635072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157873620862588,
      "loss": 1.3436,
      "theoretical_loss": 3.4673665553063295,
      "tokens_seen": 1941635072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041576730190571716,
      "loss": 1.3548,
      "theoretical_loss": 3.4673568211404002,
      "tokens_seen": 1941700608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157472417251755,
      "loss": 1.3427,
      "theoretical_loss": 3.4673470873950007,
      "tokens_seen": 1941766144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157271815446339,
      "loss": 1.3958,
      "theoretical_loss": 3.467337354070099,
      "tokens_seen": 1941831680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157071213640923,
      "loss": 1.2572,
      "theoretical_loss": 3.467327621165663,
      "tokens_seen": 1941897216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041568706118355066,
      "loss": 1.349,
      "theoretical_loss": 3.46731788868166,
      "tokens_seen": 1941962752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415667001003009,
      "loss": 1.2885,
      "theoretical_loss": 3.4673081566180572,
      "tokens_seen": 1942028288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156469408224674,
      "loss": 1.3483,
      "theoretical_loss": 3.467298424974823,
      "tokens_seen": 1942093824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041562688064192575,
      "loss": 1.3482,
      "theoretical_loss": 3.4672886937519243,
      "tokens_seen": 1942159360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041560682046138417,
      "loss": 1.3247,
      "theoretical_loss": 3.467278962949329,
      "tokens_seen": 1942224896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041558676028084253,
      "loss": 1.2815,
      "theoretical_loss": 3.4672692325670056,
      "tokens_seen": 1942290432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155667001003009,
      "loss": 1.1718,
      "theoretical_loss": 3.4672595026049207,
      "tokens_seen": 1942355968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041554663991975925,
      "loss": 1.3279,
      "theoretical_loss": 3.467249773063042,
      "tokens_seen": 1942421504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041552657973921767,
      "loss": 1.3205,
      "theoretical_loss": 3.467240043941338,
      "tokens_seen": 1942487040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041550651955867603,
      "loss": 1.3592,
      "theoretical_loss": 3.4672303152397763,
      "tokens_seen": 1942552576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154864593781344,
      "loss": 1.4239,
      "theoretical_loss": 3.4672205869583235,
      "tokens_seen": 1942618112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041546639919759276,
      "loss": 1.422,
      "theoretical_loss": 3.467210859096948,
      "tokens_seen": 1942683648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154463390170511,
      "loss": 1.3735,
      "theoretical_loss": 3.467201131655618,
      "tokens_seen": 1942749184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041542627883650954,
      "loss": 1.3935,
      "theoretical_loss": 3.467191404634301,
      "tokens_seen": 1942814720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154062186559679,
      "loss": 1.3062,
      "theoretical_loss": 3.467181678032963,
      "tokens_seen": 1942880256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041538615847542626,
      "loss": 1.423,
      "theoretical_loss": 3.4671719518515745,
      "tokens_seen": 1942945792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153660982948846,
      "loss": 1.3051,
      "theoretical_loss": 3.4671622260901014,
      "tokens_seen": 1943011328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041534603811434304,
      "loss": 1.3078,
      "theoretical_loss": 3.4671525007485116,
      "tokens_seen": 1943076864
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153259779338014,
      "loss": 1.3397,
      "theoretical_loss": 3.4671427758267725,
      "tokens_seen": 1943142400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041530591775325976,
      "loss": 1.4405,
      "theoretical_loss": 3.4671330513248533,
      "tokens_seen": 1943207936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152858575727181,
      "loss": 1.3526,
      "theoretical_loss": 3.46712332724272,
      "tokens_seen": 1943273472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152657973921765,
      "loss": 1.307,
      "theoretical_loss": 3.4671136035803416,
      "tokens_seen": 1943339008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152457372116349,
      "loss": 1.3822,
      "theoretical_loss": 3.4671038803376852,
      "tokens_seen": 1943404544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041522567703109327,
      "loss": 1.3264,
      "theoretical_loss": 3.4670941575147185,
      "tokens_seen": 1943470080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041520561685055163,
      "loss": 1.3258,
      "theoretical_loss": 3.4670844351114094,
      "tokens_seen": 1943535616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041518555667001,
      "loss": 1.3334,
      "theoretical_loss": 3.4670747131277255,
      "tokens_seen": 1943601152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151654964894684,
      "loss": 1.4422,
      "theoretical_loss": 3.467064991563635,
      "tokens_seen": 1943666688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151454363089268,
      "loss": 1.3988,
      "theoretical_loss": 3.4670552704191047,
      "tokens_seen": 1943732224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151253761283852,
      "loss": 1.3525,
      "theoretical_loss": 3.4670455496941033,
      "tokens_seen": 1943797760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041510531594784355,
      "loss": 1.2827,
      "theoretical_loss": 3.467035829388598,
      "tokens_seen": 1943863296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041508525576730196,
      "loss": 1.4013,
      "theoretical_loss": 3.467026109502557,
      "tokens_seen": 1943928832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150651955867603,
      "loss": 1.4096,
      "theoretical_loss": 3.4670163900359476,
      "tokens_seen": 1943994368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150451354062187,
      "loss": 1.3927,
      "theoretical_loss": 3.4670066709887375,
      "tokens_seen": 1944059904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041502507522567705,
      "loss": 1.4027,
      "theoretical_loss": 3.466996952360895,
      "tokens_seen": 1944125440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150050150451354,
      "loss": 1.3256,
      "theoretical_loss": 3.4669872341523877,
      "tokens_seen": 1944190976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041498495486459383,
      "loss": 1.3302,
      "theoretical_loss": 3.4669775163631833,
      "tokens_seen": 1944256512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149648946840522,
      "loss": 1.3147,
      "theoretical_loss": 3.466967798993249,
      "tokens_seen": 1944322048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041494483450351055,
      "loss": 1.2997,
      "theoretical_loss": 3.466958082042553,
      "tokens_seen": 1944387584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149247743229689,
      "loss": 1.3472,
      "theoretical_loss": 3.466948365511064,
      "tokens_seen": 1944453120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041490471414242733,
      "loss": 1.3414,
      "theoretical_loss": 3.466938649398749,
      "tokens_seen": 1944518656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148846539618857,
      "loss": 1.3753,
      "theoretical_loss": 3.4669289337055753,
      "tokens_seen": 1944584192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041486459378134406,
      "loss": 1.3387,
      "theoretical_loss": 3.466919218431511,
      "tokens_seen": 1944649728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148445336008024,
      "loss": 1.333,
      "theoretical_loss": 3.4669095035765247,
      "tokens_seen": 1944715264
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148244734202608,
      "loss": 1.4065,
      "theoretical_loss": 3.466899789140583,
      "tokens_seen": 1944780800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148044132397192,
      "loss": 1.4217,
      "theoretical_loss": 3.466890075123654,
      "tokens_seen": 1944846336
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 709206,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4632246494293213,
      "objective/train/theoretical_loss": 3.466880361525707,
      "objective/train/tokens_used": 315370976,
      "theoretical_loss": 3.466880361525707,
      "tokens_seen": 1944911872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041478435305917756,
      "loss": 1.4598,
      "theoretical_loss": 3.466880361525707,
      "tokens_seen": 1944911872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147642928786359,
      "loss": 1.4358,
      "theoretical_loss": 3.466870648346707,
      "tokens_seen": 1944977408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147442326980943,
      "loss": 1.512,
      "theoretical_loss": 3.4668609355866247,
      "tokens_seen": 1945042944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147241725175527,
      "loss": 1.3632,
      "theoretical_loss": 3.466851223245426,
      "tokens_seen": 1945108480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041470411233701106,
      "loss": 1.3986,
      "theoretical_loss": 3.4668415113230795,
      "tokens_seen": 1945174016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146840521564694,
      "loss": 1.4201,
      "theoretical_loss": 3.466831799819553,
      "tokens_seen": 1945239552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146639919759278,
      "loss": 1.4438,
      "theoretical_loss": 3.466822088734814,
      "tokens_seen": 1945305088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041464393179538615,
      "loss": 1.3081,
      "theoretical_loss": 3.4668123780688305,
      "tokens_seen": 1945370624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041462387161484457,
      "loss": 1.4134,
      "theoretical_loss": 3.4668026678215704,
      "tokens_seen": 1945436160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041460381143430293,
      "loss": 1.2866,
      "theoretical_loss": 3.4667929579930012,
      "tokens_seen": 1945501696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145837512537613,
      "loss": 1.3598,
      "theoretical_loss": 3.466783248583092,
      "tokens_seen": 1945567232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041456369107321965,
      "loss": 1.3391,
      "theoretical_loss": 3.466773539591809,
      "tokens_seen": 1945632768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041454363089267807,
      "loss": 1.5092,
      "theoretical_loss": 3.4667638310191204,
      "tokens_seen": 1945698304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041452357071213643,
      "loss": 1.2782,
      "theoretical_loss": 3.4667541228649954,
      "tokens_seen": 1945763840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145035105315948,
      "loss": 1.2973,
      "theoretical_loss": 3.4667444151294005,
      "tokens_seen": 1945829376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041448345035105316,
      "loss": 1.4007,
      "theoretical_loss": 3.4667347078123036,
      "tokens_seen": 1945894912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144633901705115,
      "loss": 1.3627,
      "theoretical_loss": 3.466725000913673,
      "tokens_seen": 1945960448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041444332998996994,
      "loss": 1.3403,
      "theoretical_loss": 3.4667152944334765,
      "tokens_seen": 1946025984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144232698094283,
      "loss": 1.3026,
      "theoretical_loss": 3.466705588371682,
      "tokens_seen": 1946091520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041440320962888666,
      "loss": 1.367,
      "theoretical_loss": 3.4666958827282572,
      "tokens_seen": 1946157056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414383149448345,
      "loss": 1.3843,
      "theoretical_loss": 3.46668617750317,
      "tokens_seen": 1946222592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041436308926780344,
      "loss": 1.425,
      "theoretical_loss": 3.4666764726963883,
      "tokens_seen": 1946288128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143430290872618,
      "loss": 1.3519,
      "theoretical_loss": 3.4666667683078805,
      "tokens_seen": 1946353664
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041432296890672016,
      "loss": 1.4098,
      "theoretical_loss": 3.466657064337614,
      "tokens_seen": 1946419200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143029087261785,
      "loss": 1.415,
      "theoretical_loss": 3.466647360785556,
      "tokens_seen": 1946484736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142828485456369,
      "loss": 1.362,
      "theoretical_loss": 3.4666376576516758,
      "tokens_seen": 1946550272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142627883650953,
      "loss": 1.3715,
      "theoretical_loss": 3.4666279549359404,
      "tokens_seen": 1946615808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041424272818455367,
      "loss": 1.3768,
      "theoretical_loss": 3.4666182526383182,
      "tokens_seen": 1946681344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041422266800401203,
      "loss": 1.373,
      "theoretical_loss": 3.4666085507587767,
      "tokens_seen": 1946746880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142026078234704,
      "loss": 1.4301,
      "theoretical_loss": 3.466598849297284,
      "tokens_seen": 1946812416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141825476429288,
      "loss": 1.3681,
      "theoretical_loss": 3.466589148253808,
      "tokens_seen": 1946877952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041416248746238717,
      "loss": 1.333,
      "theoretical_loss": 3.466579447628316,
      "tokens_seen": 1946943488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041414242728184553,
      "loss": 1.4065,
      "theoretical_loss": 3.4665697474207775,
      "tokens_seen": 1947009024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141223671013039,
      "loss": 1.3339,
      "theoretical_loss": 3.466560047631159,
      "tokens_seen": 1947074560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141023069207623,
      "loss": 1.3279,
      "theoretical_loss": 3.466550348259428,
      "tokens_seen": 1947140096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140822467402207,
      "loss": 1.3322,
      "theoretical_loss": 3.4665406493055544,
      "tokens_seen": 1947205632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041406218655967904,
      "loss": 1.3926,
      "theoretical_loss": 3.4665309507695046,
      "tokens_seen": 1947271168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140421263791374,
      "loss": 1.2608,
      "theoretical_loss": 3.466521252651247,
      "tokens_seen": 1947336704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041402206619859576,
      "loss": 1.4738,
      "theoretical_loss": 3.46651155495075,
      "tokens_seen": 1947402240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140020060180542,
      "loss": 1.481,
      "theoretical_loss": 3.4665018576679802,
      "tokens_seen": 1947467776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041398194583751254,
      "loss": 1.4261,
      "theoretical_loss": 3.466492160802907,
      "tokens_seen": 1947533312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004139618856569709,
      "loss": 1.4166,
      "theoretical_loss": 3.466482464355498,
      "tokens_seen": 1947598848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041394182547642926,
      "loss": 1.3515,
      "theoretical_loss": 3.4664727683257204,
      "tokens_seen": 1947664384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004139217652958877,
      "loss": 1.3183,
      "theoretical_loss": 3.466463072713543,
      "tokens_seen": 1947729920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041390170511534604,
      "loss": 1.4433,
      "theoretical_loss": 3.466453377518933,
      "tokens_seen": 1947795456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138816449348044,
      "loss": 1.4355,
      "theoretical_loss": 3.466443682741859,
      "tokens_seen": 1947860992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041386158475426277,
      "loss": 1.458,
      "theoretical_loss": 3.466433988382289,
      "tokens_seen": 1947926528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041384152457372113,
      "loss": 1.3782,
      "theoretical_loss": 3.4664242944401904,
      "tokens_seen": 1947992064
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041382146439317954,
      "loss": 1.4586,
      "theoretical_loss": 3.466414600915532,
      "tokens_seen": 1948057600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138014042126379,
      "loss": 1.2852,
      "theoretical_loss": 3.466404907808281,
      "tokens_seen": 1948123136
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 710405,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4715192317962646,
      "objective/train/theoretical_loss": 3.466395215118406,
      "objective/train/tokens_used": 318647776,
      "theoretical_loss": 3.466395215118406,
      "tokens_seen": 1948188672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041378134403209627,
      "loss": 1.374,
      "theoretical_loss": 3.466395215118406,
      "tokens_seen": 1948188672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041376128385155463,
      "loss": 1.2738,
      "theoretical_loss": 3.466385522845874,
      "tokens_seen": 1948254208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041374122367101305,
      "loss": 1.4602,
      "theoretical_loss": 3.4663758309906543,
      "tokens_seen": 1948319744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004137211634904714,
      "loss": 1.4301,
      "theoretical_loss": 3.466366139552714,
      "tokens_seen": 1948385280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041370110330992977,
      "loss": 1.337,
      "theoretical_loss": 3.4663564485320215,
      "tokens_seen": 1948450816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041368104312938813,
      "loss": 1.3774,
      "theoretical_loss": 3.466346757928545,
      "tokens_seen": 1948516352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041366098294884655,
      "loss": 1.4442,
      "theoretical_loss": 3.466337067742252,
      "tokens_seen": 1948581888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041364092276830497,
      "loss": 1.3464,
      "theoretical_loss": 3.4663273779731103,
      "tokens_seen": 1948647424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041362086258776333,
      "loss": 1.5564,
      "theoretical_loss": 3.4663176886210887,
      "tokens_seen": 1948712960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136008024072217,
      "loss": 1.335,
      "theoretical_loss": 3.4663079996861548,
      "tokens_seen": 1948778496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041358074222668005,
      "loss": 1.3677,
      "theoretical_loss": 3.4662983111682766,
      "tokens_seen": 1948844032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041356068204613847,
      "loss": 1.2431,
      "theoretical_loss": 3.466288623067422,
      "tokens_seen": 1948909568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041354062186559683,
      "loss": 1.352,
      "theoretical_loss": 3.4662789353835595,
      "tokens_seen": 1948975104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135205616850552,
      "loss": 1.4645,
      "theoretical_loss": 3.4662692481166566,
      "tokens_seen": 1949040640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041350050150451356,
      "loss": 1.3282,
      "theoretical_loss": 3.466259561266682,
      "tokens_seen": 1949106176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134804413239719,
      "loss": 1.3414,
      "theoretical_loss": 3.466249874833603,
      "tokens_seen": 1949171712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041346038114343034,
      "loss": 1.3408,
      "theoretical_loss": 3.4662401888173875,
      "tokens_seen": 1949237248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134403209628887,
      "loss": 1.3645,
      "theoretical_loss": 3.4662305032180045,
      "tokens_seen": 1949302784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041342026078234706,
      "loss": 1.4125,
      "theoretical_loss": 3.4662208180354215,
      "tokens_seen": 1949368320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134002006018054,
      "loss": 1.415,
      "theoretical_loss": 3.4662111332696064,
      "tokens_seen": 1949433856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041338014042126384,
      "loss": 1.3519,
      "theoretical_loss": 3.4662014489205277,
      "tokens_seen": 1949499392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133600802407222,
      "loss": 1.3302,
      "theoretical_loss": 3.466191764988153,
      "tokens_seen": 1949564928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041334002006018056,
      "loss": 1.3472,
      "theoretical_loss": 3.466182081472451,
      "tokens_seen": 1949630464
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133199598796389,
      "loss": 1.3047,
      "theoretical_loss": 3.4661723983733888,
      "tokens_seen": 1949696000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132998996990973,
      "loss": 1.3471,
      "theoretical_loss": 3.4661627156909356,
      "tokens_seen": 1949761536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132798395185557,
      "loss": 1.2753,
      "theoretical_loss": 3.466153033425058,
      "tokens_seen": 1949827072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041325977933801407,
      "loss": 1.3179,
      "theoretical_loss": 3.4661433515757256,
      "tokens_seen": 1949892608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041323971915747243,
      "loss": 1.3378,
      "theoretical_loss": 3.4661336701429057,
      "tokens_seen": 1949958144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132196589769308,
      "loss": 1.3095,
      "theoretical_loss": 3.4661239891265665,
      "tokens_seen": 1950023680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131995987963892,
      "loss": 1.3829,
      "theoretical_loss": 3.4661143085266763,
      "tokens_seen": 1950089216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041317953861584757,
      "loss": 1.3426,
      "theoretical_loss": 3.466104628343203,
      "tokens_seen": 1950154752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041315947843530593,
      "loss": 1.4869,
      "theoretical_loss": 3.4660949485761146,
      "tokens_seen": 1950220288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131394182547643,
      "loss": 1.4083,
      "theoretical_loss": 3.466085269225379,
      "tokens_seen": 1950285824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131193580742227,
      "loss": 1.405,
      "theoretical_loss": 3.4660755902909646,
      "tokens_seen": 1950351360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130992978936811,
      "loss": 1.4002,
      "theoretical_loss": 3.46606591177284,
      "tokens_seen": 1950416896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041307923771313944,
      "loss": 1.3209,
      "theoretical_loss": 3.4660562336709724,
      "tokens_seen": 1950482432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130591775325978,
      "loss": 1.3634,
      "theoretical_loss": 3.4660465559853306,
      "tokens_seen": 1950547968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041303911735205616,
      "loss": 1.3572,
      "theoretical_loss": 3.466036878715882,
      "tokens_seen": 1950613504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130190571715146,
      "loss": 1.4311,
      "theoretical_loss": 3.4660272018625955,
      "tokens_seen": 1950679040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041299899699097294,
      "loss": 1.4459,
      "theoretical_loss": 3.466017525425439,
      "tokens_seen": 1950744576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129789368104313,
      "loss": 1.4468,
      "theoretical_loss": 3.4660078494043804,
      "tokens_seen": 1950810112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041295887662988966,
      "loss": 1.3474,
      "theoretical_loss": 3.465998173799388,
      "tokens_seen": 1950875648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129388164493481,
      "loss": 1.3252,
      "theoretical_loss": 3.4659884986104297,
      "tokens_seen": 1950941184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041291875626880644,
      "loss": 1.4643,
      "theoretical_loss": 3.4659788238374736,
      "tokens_seen": 1951006720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128986960882648,
      "loss": 1.3873,
      "theoretical_loss": 3.4659691494804887,
      "tokens_seen": 1951072256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041287863590772317,
      "loss": 1.2163,
      "theoretical_loss": 3.4659594755394423,
      "tokens_seen": 1951137792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041285857572718153,
      "loss": 1.3811,
      "theoretical_loss": 3.4659498020143023,
      "tokens_seen": 1951203328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041283851554663995,
      "loss": 1.4208,
      "theoretical_loss": 3.4659401289050376,
      "tokens_seen": 1951268864
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128184553660983,
      "loss": 1.4176,
      "theoretical_loss": 3.4659304562116158,
      "tokens_seen": 1951334400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041279839518555667,
      "loss": 1.379,
      "theoretical_loss": 3.4659207839340054,
      "tokens_seen": 1951399936
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 711575,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4758858680725098,
      "objective/train/theoretical_loss": 3.465911112072175,
      "objective/train/tokens_used": 321924576,
      "theoretical_loss": 3.465911112072175,
      "tokens_seen": 1951465472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041277833500501503,
      "loss": 1.3619,
      "theoretical_loss": 3.465911112072175,
      "tokens_seen": 1951465472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041275827482447345,
      "loss": 1.3404,
      "theoretical_loss": 3.4659014406260913,
      "tokens_seen": 1951531008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127382146439318,
      "loss": 1.4048,
      "theoretical_loss": 3.465891769595724,
      "tokens_seen": 1951596544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041271815446339017,
      "loss": 1.3414,
      "theoretical_loss": 3.465882098981041,
      "tokens_seen": 1951662080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041269809428284854,
      "loss": 1.4386,
      "theoretical_loss": 3.465872428782009,
      "tokens_seen": 1951727616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126780341023069,
      "loss": 1.308,
      "theoretical_loss": 3.465862758998598,
      "tokens_seen": 1951793152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126579739217653,
      "loss": 1.3508,
      "theoretical_loss": 3.4658530896307758,
      "tokens_seen": 1951858688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126379137412237,
      "loss": 1.2784,
      "theoretical_loss": 3.46584342067851,
      "tokens_seen": 1951924224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041261785356068204,
      "loss": 1.3221,
      "theoretical_loss": 3.465833752141769,
      "tokens_seen": 1951989760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125977933801404,
      "loss": 1.3897,
      "theoretical_loss": 3.465824084020521,
      "tokens_seen": 1952055296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125777331995988,
      "loss": 1.4931,
      "theoretical_loss": 3.465814416314734,
      "tokens_seen": 1952120832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125576730190572,
      "loss": 1.4631,
      "theoretical_loss": 3.465804749024377,
      "tokens_seen": 1952186368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041253761283851554,
      "loss": 1.3334,
      "theoretical_loss": 3.465795082149418,
      "tokens_seen": 1952251904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125175526579739,
      "loss": 1.3649,
      "theoretical_loss": 3.4657854156898242,
      "tokens_seen": 1952317440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041249749247743227,
      "loss": 1.4301,
      "theoretical_loss": 3.4657757496455646,
      "tokens_seen": 1952382976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124774322968907,
      "loss": 1.4002,
      "theoretical_loss": 3.4657660840166074,
      "tokens_seen": 1952448512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041245737211634904,
      "loss": 1.3878,
      "theoretical_loss": 3.4657564188029206,
      "tokens_seen": 1952514048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124373119358074,
      "loss": 1.3046,
      "theoretical_loss": 3.465746754004473,
      "tokens_seen": 1952579584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041241725175526577,
      "loss": 1.4153,
      "theoretical_loss": 3.465737089621232,
      "tokens_seen": 1952645120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123971915747242,
      "loss": 1.3442,
      "theoretical_loss": 3.4657274256531663,
      "tokens_seen": 1952710656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041237713139418255,
      "loss": 1.4558,
      "theoretical_loss": 3.465717762100244,
      "tokens_seen": 1952776192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123570712136409,
      "loss": 1.42,
      "theoretical_loss": 3.465708098962433,
      "tokens_seen": 1952841728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041233701103309927,
      "loss": 1.3981,
      "theoretical_loss": 3.465698436239702,
      "tokens_seen": 1952907264
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041231695085255763,
      "loss": 1.3289,
      "theoretical_loss": 3.4656887739320195,
      "tokens_seen": 1952972800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041229689067201605,
      "loss": 1.4061,
      "theoretical_loss": 3.465679112039353,
      "tokens_seen": 1953038336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122768304914744,
      "loss": 1.3814,
      "theoretical_loss": 3.4656694505616716,
      "tokens_seen": 1953103872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122567703109328,
      "loss": 1.4041,
      "theoretical_loss": 3.4656597894989427,
      "tokens_seen": 1953169408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041223671013039114,
      "loss": 1.2953,
      "theoretical_loss": 3.465650128851135,
      "tokens_seen": 1953234944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041221664994984955,
      "loss": 1.4445,
      "theoretical_loss": 3.4656404686182167,
      "tokens_seen": 1953300480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121965897693079,
      "loss": 1.3741,
      "theoretical_loss": 3.465630808800156,
      "tokens_seen": 1953366016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121765295887663,
      "loss": 1.446,
      "theoretical_loss": 3.4656211493969216,
      "tokens_seen": 1953431552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121564694082247,
      "loss": 1.409,
      "theoretical_loss": 3.465611490408481,
      "tokens_seen": 1953497088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121364092276831,
      "loss": 1.398,
      "theoretical_loss": 3.465601831834803,
      "tokens_seen": 1953562624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121163490471415,
      "loss": 1.3989,
      "theoretical_loss": 3.465592173675856,
      "tokens_seen": 1953628160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041209628886659984,
      "loss": 1.3836,
      "theoretical_loss": 3.465582515931608,
      "tokens_seen": 1953693696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004120762286860582,
      "loss": 1.3167,
      "theoretical_loss": 3.4655728586020267,
      "tokens_seen": 1953759232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041205616850551656,
      "loss": 1.3702,
      "theoretical_loss": 3.4655632016870817,
      "tokens_seen": 1953824768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412036108324975,
      "loss": 1.2654,
      "theoretical_loss": 3.46555354518674,
      "tokens_seen": 1953890304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041201604814443334,
      "loss": 1.3751,
      "theoretical_loss": 3.465543889100971,
      "tokens_seen": 1953955840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119959879638917,
      "loss": 1.4138,
      "theoretical_loss": 3.465534233429742,
      "tokens_seen": 1954021376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041197592778335006,
      "loss": 1.4064,
      "theoretical_loss": 3.4655245781730226,
      "tokens_seen": 1954086912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119558676028085,
      "loss": 1.3203,
      "theoretical_loss": 3.4655149233307796,
      "tokens_seen": 1954152448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041193580742226684,
      "loss": 1.3949,
      "theoretical_loss": 3.465505268902982,
      "tokens_seen": 1954217984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119157472417252,
      "loss": 1.4469,
      "theoretical_loss": 3.4654956148895986,
      "tokens_seen": 1954283520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041189568706118357,
      "loss": 1.3346,
      "theoretical_loss": 3.4654859612905966,
      "tokens_seen": 1954349056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041187562688064193,
      "loss": 1.4949,
      "theoretical_loss": 3.465476308105945,
      "tokens_seen": 1954414592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041185556670010035,
      "loss": 1.3463,
      "theoretical_loss": 3.4654666553356126,
      "tokens_seen": 1954480128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118355065195587,
      "loss": 1.3953,
      "theoretical_loss": 3.4654570029795666,
      "tokens_seen": 1954545664
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041181544633901707,
      "loss": 1.4025,
      "theoretical_loss": 3.465447351037776,
      "tokens_seen": 1954611200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041179538615847543,
      "loss": 1.354,
      "theoretical_loss": 3.4654376995102094,
      "tokens_seen": 1954676736
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 712850,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.313586711883545,
      "objective/train/theoretical_loss": 3.4654280483968347,
      "objective/train/tokens_used": 325201376,
      "theoretical_loss": 3.4654280483968347,
      "tokens_seen": 1954742272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041177532597793385,
      "loss": 1.3061,
      "theoretical_loss": 3.4654280483968347,
      "tokens_seen": 1954742272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117552657973922,
      "loss": 1.4814,
      "theoretical_loss": 3.4654183976976203,
      "tokens_seen": 1954807808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117352056168506,
      "loss": 1.5366,
      "theoretical_loss": 3.4654087474125346,
      "tokens_seen": 1954873344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041171514543630894,
      "loss": 1.3106,
      "theoretical_loss": 3.4653990975415456,
      "tokens_seen": 1954938880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116950852557673,
      "loss": 1.4516,
      "theoretical_loss": 3.4653894480846223,
      "tokens_seen": 1955004416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116750250752257,
      "loss": 1.4531,
      "theoretical_loss": 3.4653797990417328,
      "tokens_seen": 1955069952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116549648946841,
      "loss": 1.4647,
      "theoretical_loss": 3.465370150412845,
      "tokens_seen": 1955135488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041163490471414244,
      "loss": 1.3235,
      "theoretical_loss": 3.465360502197928,
      "tokens_seen": 1955201024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004116148445336008,
      "loss": 1.4262,
      "theoretical_loss": 3.4653508543969496,
      "tokens_seen": 1955266560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115947843530592,
      "loss": 1.3023,
      "theoretical_loss": 3.465341207009878,
      "tokens_seen": 1955332096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115747241725176,
      "loss": 1.4084,
      "theoretical_loss": 3.4653315600366827,
      "tokens_seen": 1955397632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041155466399197594,
      "loss": 1.3525,
      "theoretical_loss": 3.465321913477331,
      "tokens_seen": 1955463168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115346038114343,
      "loss": 1.3024,
      "theoretical_loss": 3.4653122673317913,
      "tokens_seen": 1955528704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041151454363089267,
      "loss": 1.3955,
      "theoretical_loss": 3.4653026216000327,
      "tokens_seen": 1955594240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114944834503511,
      "loss": 1.3814,
      "theoretical_loss": 3.465292976282023,
      "tokens_seen": 1955659776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041147442326980944,
      "loss": 1.4225,
      "theoretical_loss": 3.465283331377731,
      "tokens_seen": 1955725312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114543630892678,
      "loss": 1.3423,
      "theoretical_loss": 3.465273686887124,
      "tokens_seen": 1955790848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041143430290872617,
      "loss": 1.3667,
      "theoretical_loss": 3.4652640428101718,
      "tokens_seen": 1955856384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114142427281846,
      "loss": 1.3452,
      "theoretical_loss": 3.4652543991468425,
      "tokens_seen": 1955921920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041139418254764295,
      "loss": 1.3615,
      "theoretical_loss": 3.4652447558971042,
      "tokens_seen": 1955987456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004113741223671013,
      "loss": 1.3197,
      "theoretical_loss": 3.465235113060925,
      "tokens_seen": 1956052992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041135406218655967,
      "loss": 1.3394,
      "theoretical_loss": 3.4652254706382735,
      "tokens_seen": 1956118528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041133400200601803,
      "loss": 1.3216,
      "theoretical_loss": 3.4652158286291184,
      "tokens_seen": 1956184064
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041131394182547645,
      "loss": 1.2706,
      "theoretical_loss": 3.465206187033428,
      "tokens_seen": 1956249600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112938816449348,
      "loss": 1.3731,
      "theoretical_loss": 3.465196545851171,
      "tokens_seen": 1956315136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112738214643932,
      "loss": 1.3464,
      "theoretical_loss": 3.465186905082315,
      "tokens_seen": 1956380672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041125376128385154,
      "loss": 1.4162,
      "theoretical_loss": 3.4651772647268295,
      "tokens_seen": 1956446208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041123370110330995,
      "loss": 1.2998,
      "theoretical_loss": 3.465167624784682,
      "tokens_seen": 1956511744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112136409227683,
      "loss": 1.3612,
      "theoretical_loss": 3.4651579852558414,
      "tokens_seen": 1956577280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111935807422267,
      "loss": 1.4036,
      "theoretical_loss": 3.465148346140276,
      "tokens_seen": 1956642816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041117352056168504,
      "loss": 1.3542,
      "theoretical_loss": 3.4651387074379545,
      "tokens_seen": 1956708352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111534603811434,
      "loss": 1.3389,
      "theoretical_loss": 3.465129069148845,
      "tokens_seen": 1956773888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111334002006018,
      "loss": 1.3757,
      "theoretical_loss": 3.4651194312729157,
      "tokens_seen": 1956839424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111133400200602,
      "loss": 1.3944,
      "theoretical_loss": 3.465109793810136,
      "tokens_seen": 1956904960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041109327983951854,
      "loss": 1.4407,
      "theoretical_loss": 3.4651001567604736,
      "tokens_seen": 1956970496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110732196589769,
      "loss": 1.4016,
      "theoretical_loss": 3.465090520123897,
      "tokens_seen": 1957036032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110531594784353,
      "loss": 1.3521,
      "theoretical_loss": 3.465080883900375,
      "tokens_seen": 1957101568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110330992978937,
      "loss": 1.4703,
      "theoretical_loss": 3.4650712480898758,
      "tokens_seen": 1957167104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041101303911735205,
      "loss": 1.332,
      "theoretical_loss": 3.4650616126923675,
      "tokens_seen": 1957232640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109929789368104,
      "loss": 1.4206,
      "theoretical_loss": 3.4650519777078195,
      "tokens_seen": 1957298176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041097291875626877,
      "loss": 1.408,
      "theoretical_loss": 3.4650423431362,
      "tokens_seen": 1957363712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109528585757272,
      "loss": 1.4464,
      "theoretical_loss": 3.4650327089774766,
      "tokens_seen": 1957429248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041093279839518555,
      "loss": 1.4103,
      "theoretical_loss": 3.4650230752316187,
      "tokens_seen": 1957494784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109127382146439,
      "loss": 1.3697,
      "theoretical_loss": 3.465013441898595,
      "tokens_seen": 1957560320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108926780341023,
      "loss": 1.3378,
      "theoretical_loss": 3.4650038089783726,
      "tokens_seen": 1957625856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108726178535607,
      "loss": 1.365,
      "theoretical_loss": 3.4649941764709213,
      "tokens_seen": 1957691392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041085255767301905,
      "loss": 1.3233,
      "theoretical_loss": 3.4649845443762093,
      "tokens_seen": 1957756928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108324974924774,
      "loss": 1.3419,
      "theoretical_loss": 3.4649749126942053,
      "tokens_seen": 1957822464
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108124373119358,
      "loss": 1.4326,
      "theoretical_loss": 3.4649652814248775,
      "tokens_seen": 1957888000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107923771313942,
      "loss": 1.3894,
      "theoretical_loss": 3.4649556505681938,
      "tokens_seen": 1957953536
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 713999,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5213258266448975,
      "objective/train/theoretical_loss": 3.4649460201241236,
      "objective/train/tokens_used": 328478176,
      "theoretical_loss": 3.4649460201241236,
      "tokens_seen": 1958019072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041077231695085256,
      "loss": 1.4182,
      "theoretical_loss": 3.4649460201241236,
      "tokens_seen": 1958019072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107522567703109,
      "loss": 1.3096,
      "theoretical_loss": 3.4649363900926353,
      "tokens_seen": 1958084608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107321965897693,
      "loss": 1.4114,
      "theoretical_loss": 3.4649267604736966,
      "tokens_seen": 1958150144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041071213640922764,
      "loss": 1.4125,
      "theoretical_loss": 3.4649171312672777,
      "tokens_seen": 1958215680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041069207622868606,
      "loss": 1.4225,
      "theoretical_loss": 3.4649075024733453,
      "tokens_seen": 1958281216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106720160481444,
      "loss": 1.4486,
      "theoretical_loss": 3.464897874091869,
      "tokens_seen": 1958346752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041065195586760284,
      "loss": 1.4043,
      "theoretical_loss": 3.4648882461228174,
      "tokens_seen": 1958412288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106318956870612,
      "loss": 1.3946,
      "theoretical_loss": 3.464878618566158,
      "tokens_seen": 1958477824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106118355065196,
      "loss": 1.3537,
      "theoretical_loss": 3.4648689914218602,
      "tokens_seen": 1958543360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410591775325978,
      "loss": 1.4046,
      "theoretical_loss": 3.4648593646898926,
      "tokens_seen": 1958608896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041057171514543634,
      "loss": 1.253,
      "theoretical_loss": 3.4648497383702237,
      "tokens_seen": 1958674432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105516549648947,
      "loss": 1.4268,
      "theoretical_loss": 3.464840112462821,
      "tokens_seen": 1958739968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041053159478435307,
      "loss": 1.4509,
      "theoretical_loss": 3.4648304869676547,
      "tokens_seen": 1958805504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105115346038115,
      "loss": 1.4759,
      "theoretical_loss": 3.464820861884692,
      "tokens_seen": 1958871040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041049147442326985,
      "loss": 1.4996,
      "theoretical_loss": 3.464811237213903,
      "tokens_seen": 1958936576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004104714142427282,
      "loss": 1.3003,
      "theoretical_loss": 3.464801612955254,
      "tokens_seen": 1959002112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041045135406218657,
      "loss": 1.4419,
      "theoretical_loss": 3.4647919891087158,
      "tokens_seen": 1959067648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410431293881645,
      "loss": 1.4572,
      "theoretical_loss": 3.4647823656742553,
      "tokens_seen": 1959133184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041041123370110335,
      "loss": 1.3774,
      "theoretical_loss": 3.464772742651842,
      "tokens_seen": 1959198720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103911735205617,
      "loss": 1.386,
      "theoretical_loss": 3.464763120041445,
      "tokens_seen": 1959264256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041037111334002007,
      "loss": 1.4774,
      "theoretical_loss": 3.4647534978430308,
      "tokens_seen": 1959329792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041035105315947844,
      "loss": 1.4291,
      "theoretical_loss": 3.4647438760565703,
      "tokens_seen": 1959395328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041033099297893685,
      "loss": 1.3884,
      "theoretical_loss": 3.464734254682031,
      "tokens_seen": 1959460864
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103109327983952,
      "loss": 1.2816,
      "theoretical_loss": 3.4647246337193813,
      "tokens_seen": 1959526400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102908726178536,
      "loss": 1.4086,
      "theoretical_loss": 3.4647150131685898,
      "tokens_seen": 1959591936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041027081243731194,
      "loss": 1.4011,
      "theoretical_loss": 3.464705393029626,
      "tokens_seen": 1959657472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041025075225677035,
      "loss": 1.2854,
      "theoretical_loss": 3.4646957733024575,
      "tokens_seen": 1959723008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102306920762287,
      "loss": 1.419,
      "theoretical_loss": 3.4646861539870537,
      "tokens_seen": 1959788544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102106318956871,
      "loss": 1.4778,
      "theoretical_loss": 3.4646765350833824,
      "tokens_seen": 1959854080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041019057171514544,
      "loss": 1.3693,
      "theoretical_loss": 3.464666916591413,
      "tokens_seen": 1959919616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101705115346038,
      "loss": 1.351,
      "theoretical_loss": 3.4646572985111126,
      "tokens_seen": 1959985152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101504513540622,
      "loss": 1.2651,
      "theoretical_loss": 3.464647680842452,
      "tokens_seen": 1960050688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101303911735206,
      "loss": 1.3727,
      "theoretical_loss": 3.464638063585398,
      "tokens_seen": 1960116224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041011033099297894,
      "loss": 1.3827,
      "theoretical_loss": 3.464628446739921,
      "tokens_seen": 1960181760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100902708124373,
      "loss": 1.2286,
      "theoretical_loss": 3.464618830305988,
      "tokens_seen": 1960247296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100702106318957,
      "loss": 1.3899,
      "theoretical_loss": 3.464609214283568,
      "tokens_seen": 1960312832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100501504513541,
      "loss": 1.3641,
      "theoretical_loss": 3.4645995986726303,
      "tokens_seen": 1960378368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041003009027081245,
      "loss": 1.3661,
      "theoretical_loss": 3.464589983473143,
      "tokens_seen": 1960443904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100100300902708,
      "loss": 1.417,
      "theoretical_loss": 3.4645803686850747,
      "tokens_seen": 1960509440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040998996990972917,
      "loss": 1.4208,
      "theoretical_loss": 3.464570754308394,
      "tokens_seen": 1960574976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099699097291876,
      "loss": 1.3454,
      "theoretical_loss": 3.4645611403430703,
      "tokens_seen": 1960640512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040994984954864595,
      "loss": 1.3053,
      "theoretical_loss": 3.464551526789071,
      "tokens_seen": 1960706048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099297893681043,
      "loss": 1.4207,
      "theoretical_loss": 3.464541913646366,
      "tokens_seen": 1960771584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099097291875627,
      "loss": 1.4109,
      "theoretical_loss": 3.4645323009149234,
      "tokens_seen": 1960837120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098896690070211,
      "loss": 1.2338,
      "theoretical_loss": 3.4645226885947116,
      "tokens_seen": 1960902656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040986960882647945,
      "loss": 1.3504,
      "theoretical_loss": 3.4645130766856993,
      "tokens_seen": 1960968192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098495486459378,
      "loss": 1.5134,
      "theoretical_loss": 3.464503465187856,
      "tokens_seen": 1961033728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098294884653962,
      "loss": 1.3625,
      "theoretical_loss": 3.464493854101149,
      "tokens_seen": 1961099264
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098094282848546,
      "loss": 1.4069,
      "theoretical_loss": 3.464484243425548,
      "tokens_seen": 1961164800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040978936810431296,
      "loss": 1.3649,
      "theoretical_loss": 3.4644746331610214,
      "tokens_seen": 1961230336
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 715186,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2633056640625,
      "objective/train/theoretical_loss": 3.464465023307538,
      "objective/train/tokens_used": 331754976,
      "theoretical_loss": 3.464465023307538,
      "tokens_seen": 1961295872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097693079237713,
      "loss": 1.3257,
      "theoretical_loss": 3.464465023307538,
      "tokens_seen": 1961295872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097492477432297,
      "loss": 1.395,
      "theoretical_loss": 3.4644554138650667,
      "tokens_seen": 1961361408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040972918756268804,
      "loss": 1.3748,
      "theoretical_loss": 3.4644458048335753,
      "tokens_seen": 1961426944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040970912738214646,
      "loss": 1.4036,
      "theoretical_loss": 3.4644361962130334,
      "tokens_seen": 1961492480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096890672016048,
      "loss": 1.2868,
      "theoretical_loss": 3.464426588003409,
      "tokens_seen": 1961558016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096690070210632,
      "loss": 1.3478,
      "theoretical_loss": 3.4644169802046716,
      "tokens_seen": 1961623552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040964894684052155,
      "loss": 1.3996,
      "theoretical_loss": 3.4644073728167895,
      "tokens_seen": 1961689088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040962888665997996,
      "loss": 1.2375,
      "theoretical_loss": 3.4643977658397307,
      "tokens_seen": 1961754624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096088264794383,
      "loss": 1.4601,
      "theoretical_loss": 3.464388159273465,
      "tokens_seen": 1961820160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095887662988967,
      "loss": 1.367,
      "theoretical_loss": 3.4643785531179607,
      "tokens_seen": 1961885696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040956870611835505,
      "loss": 1.4649,
      "theoretical_loss": 3.4643689473731865,
      "tokens_seen": 1961951232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095486459378134,
      "loss": 1.3531,
      "theoretical_loss": 3.464359342039111,
      "tokens_seen": 1962016768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040952858575727183,
      "loss": 1.4511,
      "theoretical_loss": 3.464349737115703,
      "tokens_seen": 1962082304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095085255767302,
      "loss": 1.3885,
      "theoretical_loss": 3.4643401326029313,
      "tokens_seen": 1962147840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040948846539618855,
      "loss": 1.438,
      "theoretical_loss": 3.464330528500765,
      "tokens_seen": 1962213376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094684052156469,
      "loss": 1.3779,
      "theoretical_loss": 3.464320924809172,
      "tokens_seen": 1962278912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040944834503510533,
      "loss": 1.3406,
      "theoretical_loss": 3.464311321528122,
      "tokens_seen": 1962344448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094282848545637,
      "loss": 1.3559,
      "theoretical_loss": 3.464301718657582,
      "tokens_seen": 1962409984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040940822467402206,
      "loss": 1.3227,
      "theoretical_loss": 3.464292116197523,
      "tokens_seen": 1962475520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093881644934804,
      "loss": 1.4132,
      "theoretical_loss": 3.4642825141479126,
      "tokens_seen": 1962541056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093681043129388,
      "loss": 1.3032,
      "theoretical_loss": 3.4642729125087195,
      "tokens_seen": 1962606592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093480441323972,
      "loss": 1.2996,
      "theoretical_loss": 3.464263311279912,
      "tokens_seen": 1962672128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040932798395185556,
      "loss": 1.2976,
      "theoretical_loss": 3.4642537104614606,
      "tokens_seen": 1962737664
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093079237713139,
      "loss": 1.3485,
      "theoretical_loss": 3.464244110053332,
      "tokens_seen": 1962803200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092878635907723,
      "loss": 1.3694,
      "theoretical_loss": 3.4642345100554963,
      "tokens_seen": 1962868736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092678034102307,
      "loss": 1.3855,
      "theoretical_loss": 3.4642249104679217,
      "tokens_seen": 1962934272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040924774322968906,
      "loss": 1.3285,
      "theoretical_loss": 3.464215311290577,
      "tokens_seen": 1962999808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092276830491474,
      "loss": 1.3954,
      "theoretical_loss": 3.4642057125234316,
      "tokens_seen": 1963065344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092076228686058,
      "loss": 1.4268,
      "theoretical_loss": 3.4641961141664535,
      "tokens_seen": 1963130880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040918756268806415,
      "loss": 1.3513,
      "theoretical_loss": 3.4641865162196117,
      "tokens_seen": 1963196416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040916750250752257,
      "loss": 1.4266,
      "theoretical_loss": 3.464176918682875,
      "tokens_seen": 1963261952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409147442326981,
      "loss": 1.3756,
      "theoretical_loss": 3.464167321556212,
      "tokens_seen": 1963327488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040912738214643934,
      "loss": 1.4381,
      "theoretical_loss": 3.464157724839592,
      "tokens_seen": 1963393024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091073219658977,
      "loss": 1.319,
      "theoretical_loss": 3.4641481285329836,
      "tokens_seen": 1963458560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090872617853561,
      "loss": 1.4492,
      "theoretical_loss": 3.4641385326363556,
      "tokens_seen": 1963524096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090672016048145,
      "loss": 1.3999,
      "theoretical_loss": 3.4641289371496766,
      "tokens_seen": 1963589632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040904714142427285,
      "loss": 1.4133,
      "theoretical_loss": 3.464119342072915,
      "tokens_seen": 1963655168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090270812437312,
      "loss": 1.3892,
      "theoretical_loss": 3.4641097474060407,
      "tokens_seen": 1963720704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040900702106318957,
      "loss": 1.2092,
      "theoretical_loss": 3.4641001531490216,
      "tokens_seen": 1963786240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408986960882648,
      "loss": 1.4021,
      "theoretical_loss": 3.464090559301827,
      "tokens_seen": 1963851776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040896690070210635,
      "loss": 1.2607,
      "theoretical_loss": 3.464080965864426,
      "tokens_seen": 1963917312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089468405215647,
      "loss": 1.3859,
      "theoretical_loss": 3.4640713728367865,
      "tokens_seen": 1963982848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089267803410231,
      "loss": 1.4089,
      "theoretical_loss": 3.464061780218878,
      "tokens_seen": 1964048384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089067201604815,
      "loss": 1.3719,
      "theoretical_loss": 3.464052188010669,
      "tokens_seen": 1964113920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040888665997993985,
      "loss": 1.4005,
      "theoretical_loss": 3.4640425962121286,
      "tokens_seen": 1964179456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088665997993982,
      "loss": 1.3828,
      "theoretical_loss": 3.464033004823225,
      "tokens_seen": 1964244992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088465396188566,
      "loss": 1.3791,
      "theoretical_loss": 3.464023413843928,
      "tokens_seen": 1964310528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408826479438315,
      "loss": 1.4303,
      "theoretical_loss": 3.4640138232742066,
      "tokens_seen": 1964376064
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040880641925777336,
      "loss": 1.349,
      "theoretical_loss": 3.464004233114028,
      "tokens_seen": 1964441600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087863590772317,
      "loss": 1.419,
      "theoretical_loss": 3.4639946433633626,
      "tokens_seen": 1964507136
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 716326,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.427863359451294,
      "objective/train/theoretical_loss": 3.463985054022179,
      "objective/train/tokens_used": 335031776,
      "theoretical_loss": 3.463985054022179,
      "tokens_seen": 1964572672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087662988966901,
      "loss": 1.3491,
      "theoretical_loss": 3.463985054022179,
      "tokens_seen": 1964572672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040874623871614844,
      "loss": 1.3889,
      "theoretical_loss": 3.4639754650904453,
      "tokens_seen": 1964638208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040872617853560686,
      "loss": 1.3494,
      "theoretical_loss": 3.463965876568131,
      "tokens_seen": 1964703744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087061183550652,
      "loss": 1.3962,
      "theoretical_loss": 3.463956288455205,
      "tokens_seen": 1964769280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086860581745236,
      "loss": 1.3086,
      "theoretical_loss": 3.463946700751636,
      "tokens_seen": 1964834816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040866599799398195,
      "loss": 1.3982,
      "theoretical_loss": 3.4639371134573924,
      "tokens_seen": 1964900352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040864593781344036,
      "loss": 1.3411,
      "theoretical_loss": 3.463927526572444,
      "tokens_seen": 1964965888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086258776328987,
      "loss": 1.4402,
      "theoretical_loss": 3.463917940096759,
      "tokens_seen": 1965031424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086058174523571,
      "loss": 1.439,
      "theoretical_loss": 3.4639083540303064,
      "tokens_seen": 1965096960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040858575727181545,
      "loss": 1.3006,
      "theoretical_loss": 3.4638987683730553,
      "tokens_seen": 1965162496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085656970912738,
      "loss": 1.4645,
      "theoretical_loss": 3.4638891831249747,
      "tokens_seen": 1965228032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040854563691073223,
      "loss": 1.3335,
      "theoretical_loss": 3.463879598286033,
      "tokens_seen": 1965293568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085255767301906,
      "loss": 1.3908,
      "theoretical_loss": 3.4638700138561993,
      "tokens_seen": 1965359104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040850551654964895,
      "loss": 1.4118,
      "theoretical_loss": 3.4638604298354427,
      "tokens_seen": 1965424640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084854563691073,
      "loss": 1.4107,
      "theoretical_loss": 3.463850846223732,
      "tokens_seen": 1965490176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040846539618856573,
      "loss": 1.32,
      "theoretical_loss": 3.4638412630210356,
      "tokens_seen": 1965555712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084453360080241,
      "loss": 1.2415,
      "theoretical_loss": 3.4638316802273232,
      "tokens_seen": 1965621248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040842527582748246,
      "loss": 1.3294,
      "theoretical_loss": 3.463822097842564,
      "tokens_seen": 1965686784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084052156469408,
      "loss": 1.2951,
      "theoretical_loss": 3.4638125158667252,
      "tokens_seen": 1965752320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083851554663992,
      "loss": 1.3545,
      "theoretical_loss": 3.4638029342997774,
      "tokens_seen": 1965817856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083650952858576,
      "loss": 1.4138,
      "theoretical_loss": 3.4637933531416882,
      "tokens_seen": 1965883392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040834503510531596,
      "loss": 1.3182,
      "theoretical_loss": 3.4637837723924276,
      "tokens_seen": 1965948928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083249749247743,
      "loss": 1.4394,
      "theoretical_loss": 3.4637741920519645,
      "tokens_seen": 1966014464
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083049147442327,
      "loss": 1.3229,
      "theoretical_loss": 3.4637646121202676,
      "tokens_seen": 1966080000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082848545636911,
      "loss": 1.2749,
      "theoretical_loss": 3.463755032597305,
      "tokens_seen": 1966145536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040826479438314946,
      "loss": 1.3454,
      "theoretical_loss": 3.463745453483047,
      "tokens_seen": 1966211072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082447342026078,
      "loss": 1.4628,
      "theoretical_loss": 3.463735874777462,
      "tokens_seen": 1966276608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082246740220662,
      "loss": 1.3792,
      "theoretical_loss": 3.4637262964805187,
      "tokens_seen": 1966342144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040820461384152455,
      "loss": 1.3749,
      "theoretical_loss": 3.4637167185921856,
      "tokens_seen": 1966407680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040818455366098297,
      "loss": 1.4007,
      "theoretical_loss": 3.463707141112433,
      "tokens_seen": 1966473216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040816449348044133,
      "loss": 1.3654,
      "theoretical_loss": 3.4636975640412286,
      "tokens_seen": 1966538752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081444332998997,
      "loss": 1.2707,
      "theoretical_loss": 3.4636879873785427,
      "tokens_seen": 1966604288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040812437311935805,
      "loss": 1.3707,
      "theoretical_loss": 3.4636784111243424,
      "tokens_seen": 1966669824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040810431293881647,
      "loss": 1.343,
      "theoretical_loss": 3.4636688352785985,
      "tokens_seen": 1966735360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040808425275827483,
      "loss": 1.3446,
      "theoretical_loss": 3.463659259841279,
      "tokens_seen": 1966800896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004080641925777332,
      "loss": 1.4308,
      "theoretical_loss": 3.463649684812353,
      "tokens_seen": 1966866432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040804413239719156,
      "loss": 1.3536,
      "theoretical_loss": 3.463640110191789,
      "tokens_seen": 1966931968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004080240722166499,
      "loss": 1.4149,
      "theoretical_loss": 3.463630535979557,
      "tokens_seen": 1966997504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040800401203610834,
      "loss": 1.3658,
      "theoretical_loss": 3.463620962175625,
      "tokens_seen": 1967063040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079839518555667,
      "loss": 1.3509,
      "theoretical_loss": 3.4636113887799627,
      "tokens_seen": 1967128576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040796389167502506,
      "loss": 1.4298,
      "theoretical_loss": 3.463601815792539,
      "tokens_seen": 1967194112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079438314944834,
      "loss": 1.3599,
      "theoretical_loss": 3.4635922432133226,
      "tokens_seen": 1967259648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040792377131394184,
      "loss": 1.3437,
      "theoretical_loss": 3.463582671042283,
      "tokens_seen": 1967325184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079037111334002,
      "loss": 1.3339,
      "theoretical_loss": 3.4635730992793885,
      "tokens_seen": 1967390720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040788365095285856,
      "loss": 1.2887,
      "theoretical_loss": 3.463563527924608,
      "tokens_seen": 1967456256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078635907723169,
      "loss": 1.4638,
      "theoretical_loss": 3.4635539569779117,
      "tokens_seen": 1967521792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078435305917753,
      "loss": 1.4974,
      "theoretical_loss": 3.463544386439268,
      "tokens_seen": 1967587328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078234704112337,
      "loss": 1.3594,
      "theoretical_loss": 3.463534816308645,
      "tokens_seen": 1967652864
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040780341023069207,
      "loss": 1.4089,
      "theoretical_loss": 3.4635252465860127,
      "tokens_seen": 1967718400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040778335005015043,
      "loss": 1.3437,
      "theoretical_loss": 3.46351567727134,
      "tokens_seen": 1967783936
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 717661,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4225447177886963,
      "objective/train/theoretical_loss": 3.4635061083645957,
      "objective/train/tokens_used": 338308576,
      "theoretical_loss": 3.4635061083645957,
      "tokens_seen": 1967849472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077632898696088,
      "loss": 1.4754,
      "theoretical_loss": 3.4635061083645957,
      "tokens_seen": 1967849472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077432296890672,
      "loss": 1.3661,
      "theoretical_loss": 3.463496539865749,
      "tokens_seen": 1967915008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040772316950852557,
      "loss": 1.291,
      "theoretical_loss": 3.463486971774769,
      "tokens_seen": 1967980544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040770310932798393,
      "loss": 1.3921,
      "theoretical_loss": 3.4634774040916243,
      "tokens_seen": 1968046080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076830491474423,
      "loss": 1.4786,
      "theoretical_loss": 3.463467836816285,
      "tokens_seen": 1968111616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076629889669007,
      "loss": 1.4207,
      "theoretical_loss": 3.4634582699487186,
      "tokens_seen": 1968177152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076429287863591,
      "loss": 1.2721,
      "theoretical_loss": 3.463448703488895,
      "tokens_seen": 1968242688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076228686058175,
      "loss": 1.363,
      "theoretical_loss": 3.4634391374367834,
      "tokens_seen": 1968308224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040760280842527585,
      "loss": 1.3684,
      "theoretical_loss": 3.4634295717923527,
      "tokens_seen": 1968373760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075827482447342,
      "loss": 1.401,
      "theoretical_loss": 3.463420006555572,
      "tokens_seen": 1968439296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040756268806419263,
      "loss": 1.3656,
      "theoretical_loss": 3.4634104417264098,
      "tokens_seen": 1968504832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407542627883651,
      "loss": 1.3786,
      "theoretical_loss": 3.463400877304836,
      "tokens_seen": 1968570368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040752256770310935,
      "loss": 1.3912,
      "theoretical_loss": 3.4633913132908187,
      "tokens_seen": 1968635904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075025075225677,
      "loss": 1.3916,
      "theoretical_loss": 3.463381749684328,
      "tokens_seen": 1968701440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040748244734202613,
      "loss": 1.295,
      "theoretical_loss": 3.463372186485332,
      "tokens_seen": 1968766976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074623871614845,
      "loss": 1.407,
      "theoretical_loss": 3.4633626236938007,
      "tokens_seen": 1968832512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040744232698094286,
      "loss": 1.2804,
      "theoretical_loss": 3.4633530613097028,
      "tokens_seen": 1968898048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074222668004012,
      "loss": 1.3455,
      "theoretical_loss": 3.463343499333007,
      "tokens_seen": 1968963584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074022066198596,
      "loss": 1.2737,
      "theoretical_loss": 3.463333937763683,
      "tokens_seen": 1969029120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407382146439318,
      "loss": 1.3666,
      "theoretical_loss": 3.4633243766016992,
      "tokens_seen": 1969094656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040736208625877636,
      "loss": 1.3628,
      "theoretical_loss": 3.463314815847026,
      "tokens_seen": 1969160192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073420260782347,
      "loss": 1.3972,
      "theoretical_loss": 3.4633052554996304,
      "tokens_seen": 1969225728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073219658976931,
      "loss": 1.4204,
      "theoretical_loss": 3.4632956955594834,
      "tokens_seen": 1969291264
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073019057171515,
      "loss": 1.3792,
      "theoretical_loss": 3.4632861360265528,
      "tokens_seen": 1969356800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040728184553660986,
      "loss": 1.3825,
      "theoretical_loss": 3.463276576900809,
      "tokens_seen": 1969422336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072617853560682,
      "loss": 1.5021,
      "theoretical_loss": 3.4632670181822203,
      "tokens_seen": 1969487872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072417251755266,
      "loss": 1.3361,
      "theoretical_loss": 3.4632574598707553,
      "tokens_seen": 1969553408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040722166499498495,
      "loss": 1.3903,
      "theoretical_loss": 3.463247901966384,
      "tokens_seen": 1969618944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040720160481444337,
      "loss": 1.3482,
      "theoretical_loss": 3.463238344469075,
      "tokens_seen": 1969684480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040718154463390173,
      "loss": 1.3917,
      "theoretical_loss": 3.463228787378798,
      "tokens_seen": 1969750016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071614844533601,
      "loss": 1.3821,
      "theoretical_loss": 3.463219230695522,
      "tokens_seen": 1969815552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040714142427281845,
      "loss": 1.2996,
      "theoretical_loss": 3.4632096744192156,
      "tokens_seen": 1969881088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040712136409227687,
      "loss": 1.3301,
      "theoretical_loss": 3.4632001185498478,
      "tokens_seen": 1969946624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040710130391173523,
      "loss": 1.3884,
      "theoretical_loss": 3.4631905630873883,
      "tokens_seen": 1970012160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070812437311936,
      "loss": 1.4168,
      "theoretical_loss": 3.4631810080318064,
      "tokens_seen": 1970077696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040706118355065196,
      "loss": 1.4092,
      "theoretical_loss": 3.4631714533830706,
      "tokens_seen": 1970143232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070411233701103,
      "loss": 1.2583,
      "theoretical_loss": 3.4631618991411512,
      "tokens_seen": 1970208768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040702106318956874,
      "loss": 1.3538,
      "theoretical_loss": 3.463152345306016,
      "tokens_seen": 1970274304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070010030090271,
      "loss": 1.3519,
      "theoretical_loss": 3.463142791877634,
      "tokens_seen": 1970339840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040698094282848546,
      "loss": 1.2357,
      "theoretical_loss": 3.463133238855976,
      "tokens_seen": 1970405376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069608826479438,
      "loss": 1.3443,
      "theoretical_loss": 3.46312368624101,
      "tokens_seen": 1970470912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040694082246740224,
      "loss": 1.361,
      "theoretical_loss": 3.463114134032705,
      "tokens_seen": 1970536448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069207622868606,
      "loss": 1.3766,
      "theoretical_loss": 3.463104582231031,
      "tokens_seen": 1970601984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040690070210631896,
      "loss": 1.3534,
      "theoretical_loss": 3.4630950308359565,
      "tokens_seen": 1970667520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068806419257773,
      "loss": 1.2618,
      "theoretical_loss": 3.4630854798474506,
      "tokens_seen": 1970733056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068605817452357,
      "loss": 1.4153,
      "theoretical_loss": 3.463075929265483,
      "tokens_seen": 1970798592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068405215646941,
      "loss": 1.2814,
      "theoretical_loss": 3.4630663790900225,
      "tokens_seen": 1970864128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040682046138415247,
      "loss": 1.4098,
      "theoretical_loss": 3.4630568293210384,
      "tokens_seen": 1970929664
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040680040120361083,
      "loss": 1.3919,
      "theoretical_loss": 3.4630472799585,
      "tokens_seen": 1970995200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067803410230692,
      "loss": 1.3337,
      "theoretical_loss": 3.4630377310023763,
      "tokens_seen": 1971060736
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 718798,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.515406608581543,
      "objective/train/theoretical_loss": 3.4630281824526365,
      "objective/train/tokens_used": 341585376,
      "theoretical_loss": 3.4630281824526365,
      "tokens_seen": 1971126272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067602808425276,
      "loss": 1.4503,
      "theoretical_loss": 3.4630281824526365,
      "tokens_seen": 1971126272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040674022066198597,
      "loss": 1.5234,
      "theoretical_loss": 3.4630186343092495,
      "tokens_seen": 1971191808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040672016048144433,
      "loss": 1.3731,
      "theoretical_loss": 3.4630090865721854,
      "tokens_seen": 1971257344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067001003009027,
      "loss": 1.3886,
      "theoretical_loss": 3.462999539241413,
      "tokens_seen": 1971322880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066800401203611,
      "loss": 1.4353,
      "theoretical_loss": 3.462989992316901,
      "tokens_seen": 1971388416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040665997993981947,
      "loss": 1.3108,
      "theoretical_loss": 3.4629804457986184,
      "tokens_seen": 1971453952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040663991975927783,
      "loss": 1.3695,
      "theoretical_loss": 3.462970899686536,
      "tokens_seen": 1971519488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066198595787362,
      "loss": 1.4911,
      "theoretical_loss": 3.462961353980621,
      "tokens_seen": 1971585024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040659979939819456,
      "loss": 1.3678,
      "theoretical_loss": 3.4629518086808444,
      "tokens_seen": 1971650560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000406579739217653,
      "loss": 1.2559,
      "theoretical_loss": 3.4629422637871743,
      "tokens_seen": 1971716096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040655967903711134,
      "loss": 1.3756,
      "theoretical_loss": 3.4629327192995802,
      "tokens_seen": 1971781632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004065396188565697,
      "loss": 1.3373,
      "theoretical_loss": 3.4629231752180316,
      "tokens_seen": 1971847168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040651955867602806,
      "loss": 1.3767,
      "theoretical_loss": 3.4629136315424973,
      "tokens_seen": 1971912704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064994984954865,
      "loss": 1.3396,
      "theoretical_loss": 3.4629040882729467,
      "tokens_seen": 1971978240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040647943831494484,
      "loss": 1.4023,
      "theoretical_loss": 3.462894545409349,
      "tokens_seen": 1972043776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064593781344032,
      "loss": 1.3312,
      "theoretical_loss": 3.462885002951674,
      "tokens_seen": 1972109312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040643931795386157,
      "loss": 1.3648,
      "theoretical_loss": 3.46287546089989,
      "tokens_seen": 1972174848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040641925777331993,
      "loss": 1.3527,
      "theoretical_loss": 3.462865919253967,
      "tokens_seen": 1972240384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040639919759277834,
      "loss": 1.3348,
      "theoretical_loss": 3.462856378013874,
      "tokens_seen": 1972305920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063791374122367,
      "loss": 1.3202,
      "theoretical_loss": 3.4628468371795798,
      "tokens_seen": 1972371456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040635907723169507,
      "loss": 1.4713,
      "theoretical_loss": 3.4628372967510543,
      "tokens_seen": 1972436992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040633901705115343,
      "loss": 1.381,
      "theoretical_loss": 3.4628277567282666,
      "tokens_seen": 1972502528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040631895687061185,
      "loss": 1.4462,
      "theoretical_loss": 3.4628182171111863,
      "tokens_seen": 1972568064
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062988966900702,
      "loss": 1.3506,
      "theoretical_loss": 3.4628086778997815,
      "tokens_seen": 1972633600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040627883650952857,
      "loss": 1.2991,
      "theoretical_loss": 3.462799139094023,
      "tokens_seen": 1972699136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040625877632898693,
      "loss": 1.454,
      "theoretical_loss": 3.462789600693879,
      "tokens_seen": 1972764672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062387161484453,
      "loss": 1.3945,
      "theoretical_loss": 3.462780062699319,
      "tokens_seen": 1972830208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062186559679037,
      "loss": 1.3899,
      "theoretical_loss": 3.4627705251103125,
      "tokens_seen": 1972895744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061985957873621,
      "loss": 1.5648,
      "theoretical_loss": 3.4627609879268286,
      "tokens_seen": 1972961280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040617853560682044,
      "loss": 1.2879,
      "theoretical_loss": 3.4627514511488364,
      "tokens_seen": 1973026816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061584754262788,
      "loss": 1.4236,
      "theoretical_loss": 3.4627419147763057,
      "tokens_seen": 1973092352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040613841524573727,
      "loss": 1.3979,
      "theoretical_loss": 3.462732378809206,
      "tokens_seen": 1973157888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040611835506519563,
      "loss": 1.4721,
      "theoretical_loss": 3.4627228432475055,
      "tokens_seen": 1973223424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000406098294884654,
      "loss": 1.4256,
      "theoretical_loss": 3.4627133080911743,
      "tokens_seen": 1973288960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040607823470411236,
      "loss": 1.4593,
      "theoretical_loss": 3.4627037733401815,
      "tokens_seen": 1973354496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060581745235707,
      "loss": 1.2972,
      "theoretical_loss": 3.462694238994497,
      "tokens_seen": 1973420032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040603811434302914,
      "loss": 1.3536,
      "theoretical_loss": 3.462684705054089,
      "tokens_seen": 1973485568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060180541624875,
      "loss": 1.4176,
      "theoretical_loss": 3.4626751715189275,
      "tokens_seen": 1973551104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040599799398194586,
      "loss": 1.3094,
      "theoretical_loss": 3.4626656383889816,
      "tokens_seen": 1973616640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004059779338014042,
      "loss": 1.3243,
      "theoretical_loss": 3.462656105664221,
      "tokens_seen": 1973682176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040595787362086264,
      "loss": 1.3508,
      "theoretical_loss": 3.4626465733446143,
      "tokens_seen": 1973747712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405937813440321,
      "loss": 1.4177,
      "theoretical_loss": 3.4626370414301317,
      "tokens_seen": 1973813248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040591775325977936,
      "loss": 1.2631,
      "theoretical_loss": 3.462627509920742,
      "tokens_seen": 1973878784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058976930792377,
      "loss": 1.4504,
      "theoretical_loss": 3.4626179788164144,
      "tokens_seen": 1973944320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058776328986961,
      "loss": 1.4411,
      "theoretical_loss": 3.462608448117119,
      "tokens_seen": 1974009856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058575727181545,
      "loss": 1.3834,
      "theoretical_loss": 3.4625989178228243,
      "tokens_seen": 1974075392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040583751253761287,
      "loss": 1.4627,
      "theoretical_loss": 3.4625893879335,
      "tokens_seen": 1974140928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040581745235707123,
      "loss": 1.3651,
      "theoretical_loss": 3.462579858449115,
      "tokens_seen": 1974206464
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057973921765296,
      "loss": 1.3855,
      "theoretical_loss": 3.4625703293696395,
      "tokens_seen": 1974272000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405777331995988,
      "loss": 1.3035,
      "theoretical_loss": 3.4625608006950426,
      "tokens_seen": 1974337536
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 720061,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3045427799224854,
      "objective/train/theoretical_loss": 3.4625512724252934,
      "objective/train/tokens_used": 344862176,
      "theoretical_loss": 3.4625512724252934,
      "tokens_seen": 1974403072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040575727181544637,
      "loss": 1.2807,
      "theoretical_loss": 3.4625512724252934,
      "tokens_seen": 1974403072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040573721163490473,
      "loss": 1.458,
      "theoretical_loss": 3.4625417445603617,
      "tokens_seen": 1974468608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057171514543631,
      "loss": 1.3762,
      "theoretical_loss": 3.4625322171002155,
      "tokens_seen": 1974534144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040569709127382146,
      "loss": 1.4555,
      "theoretical_loss": 3.462522690044826,
      "tokens_seen": 1974599680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056770310932799,
      "loss": 1.522,
      "theoretical_loss": 3.4625131633941617,
      "tokens_seen": 1974665216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040565697091273824,
      "loss": 1.4294,
      "theoretical_loss": 3.462503637148192,
      "tokens_seen": 1974730752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056369107321966,
      "loss": 1.325,
      "theoretical_loss": 3.462494111306886,
      "tokens_seen": 1974796288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040561685055165496,
      "loss": 1.3833,
      "theoretical_loss": 3.4624845858702136,
      "tokens_seen": 1974861824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055967903711134,
      "loss": 1.3998,
      "theoretical_loss": 3.4624750608381434,
      "tokens_seen": 1974927360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040557673019057174,
      "loss": 1.3835,
      "theoretical_loss": 3.462465536210646,
      "tokens_seen": 1974992896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055566700100301,
      "loss": 1.347,
      "theoretical_loss": 3.4624560119876904,
      "tokens_seen": 1975058432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040553660982948846,
      "loss": 1.3459,
      "theoretical_loss": 3.462446488169245,
      "tokens_seen": 1975123968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055165496489469,
      "loss": 1.4374,
      "theoretical_loss": 3.4624369647552804,
      "tokens_seen": 1975189504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040549648946840524,
      "loss": 1.4228,
      "theoretical_loss": 3.462427441745765,
      "tokens_seen": 1975255040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054764292878636,
      "loss": 1.4275,
      "theoretical_loss": 3.4624179191406697,
      "tokens_seen": 1975320576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040545636910732197,
      "loss": 1.3858,
      "theoretical_loss": 3.4624083969399626,
      "tokens_seen": 1975386112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040543630892678033,
      "loss": 1.3259,
      "theoretical_loss": 3.462398875143613,
      "tokens_seen": 1975451648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040541624874623874,
      "loss": 1.4255,
      "theoretical_loss": 3.4623893537515915,
      "tokens_seen": 1975517184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053961885656971,
      "loss": 1.3874,
      "theoretical_loss": 3.462379832763866,
      "tokens_seen": 1975582720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040537612838515547,
      "loss": 1.3627,
      "theoretical_loss": 3.4623703121804077,
      "tokens_seen": 1975648256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040535606820461383,
      "loss": 1.4221,
      "theoretical_loss": 3.462360792001184,
      "tokens_seen": 1975713792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040533600802407225,
      "loss": 1.377,
      "theoretical_loss": 3.462351272226166,
      "tokens_seen": 1975779328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053159478435306,
      "loss": 1.4155,
      "theoretical_loss": 3.4623417528553224,
      "tokens_seen": 1975844864
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040529588766298897,
      "loss": 1.4925,
      "theoretical_loss": 3.462332233888623,
      "tokens_seen": 1975910400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040527582748244733,
      "loss": 1.3736,
      "theoretical_loss": 3.462322715326037,
      "tokens_seen": 1975975936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052557673019057,
      "loss": 1.2822,
      "theoretical_loss": 3.462313197167534,
      "tokens_seen": 1976041472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052357071213641,
      "loss": 1.4228,
      "theoretical_loss": 3.4623036794130826,
      "tokens_seen": 1976107008
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052156469408225,
      "loss": 1.3359,
      "theoretical_loss": 3.462294162062653,
      "tokens_seen": 1976172544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040519558676028084,
      "loss": 1.3545,
      "theoretical_loss": 3.462284645116215,
      "tokens_seen": 1976238080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051755265797392,
      "loss": 1.3037,
      "theoretical_loss": 3.4622751285737374,
      "tokens_seen": 1976303616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051554663991976,
      "loss": 1.3608,
      "theoretical_loss": 3.46226561243519,
      "tokens_seen": 1976369152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405135406218656,
      "loss": 1.387,
      "theoretical_loss": 3.462256096700542,
      "tokens_seen": 1976434688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040511534603811434,
      "loss": 1.3466,
      "theoretical_loss": 3.462246581369763,
      "tokens_seen": 1976500224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050952858575727,
      "loss": 1.2942,
      "theoretical_loss": 3.462237066442823,
      "tokens_seen": 1976565760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040507522567703107,
      "loss": 1.4503,
      "theoretical_loss": 3.46222755191969,
      "tokens_seen": 1976631296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050551654964895,
      "loss": 1.4201,
      "theoretical_loss": 3.4622180378003353,
      "tokens_seen": 1976696832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040503510531594784,
      "loss": 1.4125,
      "theoretical_loss": 3.462208524084727,
      "tokens_seen": 1976762368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050150451354062,
      "loss": 1.2858,
      "theoretical_loss": 3.4621990107728355,
      "tokens_seen": 1976827904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040499498495486457,
      "loss": 1.4234,
      "theoretical_loss": 3.46218949786463,
      "tokens_seen": 1976893440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404974924774323,
      "loss": 1.3062,
      "theoretical_loss": 3.462179985360079,
      "tokens_seen": 1976958976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040495486459378135,
      "loss": 1.3107,
      "theoretical_loss": 3.4621704732591536,
      "tokens_seen": 1977024512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004049348044132397,
      "loss": 1.3963,
      "theoretical_loss": 3.462160961561822,
      "tokens_seen": 1977090048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040491474423269807,
      "loss": 1.3372,
      "theoretical_loss": 3.4621514502680553,
      "tokens_seen": 1977155584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040489468405215643,
      "loss": 1.2865,
      "theoretical_loss": 3.4621419393778208,
      "tokens_seen": 1977221120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040487462387161485,
      "loss": 1.3857,
      "theoretical_loss": 3.4621324288910897,
      "tokens_seen": 1977286656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048545636910732,
      "loss": 1.2147,
      "theoretical_loss": 3.462122918807831,
      "tokens_seen": 1977352192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048345035105316,
      "loss": 1.3404,
      "theoretical_loss": 3.4621134091280137,
      "tokens_seen": 1977417728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040481444332998994,
      "loss": 1.3318,
      "theoretical_loss": 3.4621038998516083,
      "tokens_seen": 1977483264
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040479438314944835,
      "loss": 1.3173,
      "theoretical_loss": 3.4620943909785837,
      "tokens_seen": 1977548800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047743229689067,
      "loss": 1.3237,
      "theoretical_loss": 3.4620848825089094,
      "tokens_seen": 1977614336
    },
    {
      "debugging/Compilability": 0.9230769230769231,
      "debugging/distinct-1-grams": 0.7440830156991918,
      "debugging/entropy-1-grams": 5.4710865197057785,
      "debugging/length": 464.53846153846155,
      "debugging/num_segments": 13,
      "debugging/score": 0.008355915970037622,
      "debugging/score_std": 0.00655650099508135,
      "epoch": 0.2,
      "objective/train/docs_used": 721216,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3672314882278442,
      "objective/train/theoretical_loss": 3.4620753744425556,
      "objective/train/tokens_used": 348138976,
      "theoretical_loss": 3.4620753744425556,
      "tokens_seen": 1977679872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047542627883651,
      "loss": 1.4309,
      "theoretical_loss": 3.4620753744425556,
      "tokens_seen": 1977679872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040473420260782344,
      "loss": 1.4119,
      "theoretical_loss": 3.462065866779491,
      "tokens_seen": 1977745408
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047141424272818,
      "loss": 1.2317,
      "theoretical_loss": 3.4620563595196847,
      "tokens_seen": 1977810944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046940822467402,
      "loss": 1.3361,
      "theoretical_loss": 3.462046852663108,
      "tokens_seen": 1977876480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046740220661986,
      "loss": 1.3346,
      "theoretical_loss": 3.462037346209729,
      "tokens_seen": 1977942016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404653961885657,
      "loss": 1.2395,
      "theoretical_loss": 3.462027840159517,
      "tokens_seen": 1978007552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040463390170511536,
      "loss": 1.3338,
      "theoretical_loss": 3.462018334512443,
      "tokens_seen": 1978073088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046138415245738,
      "loss": 1.3785,
      "theoretical_loss": 3.462008829268475,
      "tokens_seen": 1978138624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040459378134403214,
      "loss": 1.338,
      "theoretical_loss": 3.461999324427584,
      "tokens_seen": 1978204160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045737211634905,
      "loss": 1.378,
      "theoretical_loss": 3.4619898199897388,
      "tokens_seen": 1978269696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040455366098294886,
      "loss": 1.3553,
      "theoretical_loss": 3.461980315954909,
      "tokens_seen": 1978335232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045336008024073,
      "loss": 1.3256,
      "theoretical_loss": 3.461970812323064,
      "tokens_seen": 1978400768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040451354062186564,
      "loss": 1.4078,
      "theoretical_loss": 3.461961309094173,
      "tokens_seen": 1978466304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404493480441324,
      "loss": 1.5034,
      "theoretical_loss": 3.4619518062682064,
      "tokens_seen": 1978531840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040447342026078237,
      "loss": 1.4358,
      "theoretical_loss": 3.4619423038451336,
      "tokens_seen": 1978597376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040445336008024073,
      "loss": 1.3451,
      "theoretical_loss": 3.461932801824924,
      "tokens_seen": 1978662912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040443329989969915,
      "loss": 1.3902,
      "theoretical_loss": 3.4619233002075473,
      "tokens_seen": 1978728448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004044132397191575,
      "loss": 1.3679,
      "theoretical_loss": 3.4619137989929727,
      "tokens_seen": 1978793984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040439317953861587,
      "loss": 1.3711,
      "theoretical_loss": 3.46190429818117,
      "tokens_seen": 1978859520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040437311935807423,
      "loss": 1.3006,
      "theoretical_loss": 3.4618947977721097,
      "tokens_seen": 1978925056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040435305917753265,
      "loss": 1.3252,
      "theoretical_loss": 3.4618852977657593,
      "tokens_seen": 1978990592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404332998996991,
      "loss": 1.4415,
      "theoretical_loss": 3.4618757981620907,
      "tokens_seen": 1979056128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040431293881644937,
      "loss": 1.1852,
      "theoretical_loss": 3.461866298961072,
      "tokens_seen": 1979121664
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040429287863590773,
      "loss": 1.379,
      "theoretical_loss": 3.461856800162673,
      "tokens_seen": 1979187200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042728184553661,
      "loss": 1.4844,
      "theoretical_loss": 3.461847301766864,
      "tokens_seen": 1979252736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042527582748245,
      "loss": 1.3181,
      "theoretical_loss": 3.461837803773614,
      "tokens_seen": 1979318272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042326980942829,
      "loss": 1.311,
      "theoretical_loss": 3.461828306182893,
      "tokens_seen": 1979383808
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040421263791374124,
      "loss": 1.365,
      "theoretical_loss": 3.4618188089946695,
      "tokens_seen": 1979449344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041925777331996,
      "loss": 1.3647,
      "theoretical_loss": 3.461809312208915,
      "tokens_seen": 1979514880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404172517552658,
      "loss": 1.3756,
      "theoretical_loss": 3.4617998158255974,
      "tokens_seen": 1979580416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041524573721164,
      "loss": 1.2608,
      "theoretical_loss": 3.4617903198446878,
      "tokens_seen": 1979645952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040413239719157474,
      "loss": 1.2831,
      "theoretical_loss": 3.4617808242661545,
      "tokens_seen": 1979711488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041123370110331,
      "loss": 1.298,
      "theoretical_loss": 3.461771329089968,
      "tokens_seen": 1979777024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040409227683049147,
      "loss": 1.4019,
      "theoretical_loss": 3.461761834316097,
      "tokens_seen": 1979842560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004040722166499499,
      "loss": 1.4078,
      "theoretical_loss": 3.4617523399445123,
      "tokens_seen": 1979908096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040405215646940824,
      "loss": 1.3002,
      "theoretical_loss": 3.4617428459751833,
      "tokens_seen": 1979973632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004040320962888666,
      "loss": 1.3515,
      "theoretical_loss": 3.4617333524080784,
      "tokens_seen": 1980039168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040401203610832497,
      "loss": 1.3781,
      "theoretical_loss": 3.461723859243169,
      "tokens_seen": 1980104704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004039919759277834,
      "loss": 1.3454,
      "theoretical_loss": 3.461714366480424,
      "tokens_seen": 1980170240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040397191574724175,
      "loss": 1.2799,
      "theoretical_loss": 3.461704874119813,
      "tokens_seen": 1980235776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004039518555667001,
      "loss": 1.4118,
      "theoretical_loss": 3.461695382161305,
      "tokens_seen": 1980301312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040393179538615847,
      "loss": 1.3863,
      "theoretical_loss": 3.461685890604871,
      "tokens_seen": 1980366848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040391173520561683,
      "loss": 1.2998,
      "theoretical_loss": 3.4616763994504796,
      "tokens_seen": 1980432384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040389167502507525,
      "loss": 1.3822,
      "theoretical_loss": 3.461666908698101,
      "tokens_seen": 1980497920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004038716148445336,
      "loss": 1.325,
      "theoretical_loss": 3.4616574183477047,
      "tokens_seen": 1980563456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403851554663992,
      "loss": 1.238,
      "theoretical_loss": 3.4616479283992607,
      "tokens_seen": 1980628992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040383149448345034,
      "loss": 1.4144,
      "theoretical_loss": 3.461638438852738,
      "tokens_seen": 1980694528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040381143430290875,
      "loss": 1.3013,
      "theoretical_loss": 3.461628949708106,
      "tokens_seen": 1980760064
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037913741223671,
      "loss": 1.3298,
      "theoretical_loss": 3.461619460965336,
      "tokens_seen": 1980825600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037713139418255,
      "loss": 1.4929,
      "theoretical_loss": 3.4616099726243963,
      "tokens_seen": 1980891136
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 722322,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3439686298370361,
      "objective/train/theoretical_loss": 3.4616004846852575,
      "objective/train/tokens_used": 351415776,
      "theoretical_loss": 3.4616004846852575,
      "tokens_seen": 1980956672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040375125376128384,
      "loss": 1.3363,
      "theoretical_loss": 3.4616004846852575,
      "tokens_seen": 1980956672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037311935807422,
      "loss": 1.3386,
      "theoretical_loss": 3.461590997147889,
      "tokens_seen": 1981022208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037111334002006,
      "loss": 1.4471,
      "theoretical_loss": 3.4615815100122593,
      "tokens_seen": 1981087744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403691073219659,
      "loss": 1.4177,
      "theoretical_loss": 3.4615720232783396,
      "tokens_seen": 1981153280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040367101303911734,
      "loss": 1.384,
      "theoretical_loss": 3.461562536946099,
      "tokens_seen": 1981218816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036509528585757,
      "loss": 1.3137,
      "theoretical_loss": 3.4615530510155077,
      "tokens_seen": 1981284352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036308926780341,
      "loss": 1.4487,
      "theoretical_loss": 3.4615435654865347,
      "tokens_seen": 1981349888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036108324974925,
      "loss": 1.2266,
      "theoretical_loss": 3.46153408035915,
      "tokens_seen": 1981415424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040359077231695085,
      "loss": 1.4123,
      "theoretical_loss": 3.4615245956333234,
      "tokens_seen": 1981480960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004035707121364092,
      "loss": 1.3577,
      "theoretical_loss": 3.4615151113090246,
      "tokens_seen": 1981546496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040355065195586757,
      "loss": 1.3255,
      "theoretical_loss": 3.4615056273862232,
      "tokens_seen": 1981612032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403530591775326,
      "loss": 1.3068,
      "theoretical_loss": 3.4614961438648892,
      "tokens_seen": 1981677568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040351053159478435,
      "loss": 1.3412,
      "theoretical_loss": 3.461486660744992,
      "tokens_seen": 1981743104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034904714142427,
      "loss": 1.3234,
      "theoretical_loss": 3.4614771780265015,
      "tokens_seen": 1981808640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034704112337011,
      "loss": 1.4,
      "theoretical_loss": 3.4614676957093873,
      "tokens_seen": 1981874176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034503510531595,
      "loss": 1.3813,
      "theoretical_loss": 3.4614582137936196,
      "tokens_seen": 1981939712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040343029087261785,
      "loss": 1.3875,
      "theoretical_loss": 3.4614487322791674,
      "tokens_seen": 1982005248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034102306920762,
      "loss": 1.3009,
      "theoretical_loss": 3.4614392511660013,
      "tokens_seen": 1982070784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033901705115346,
      "loss": 1.3277,
      "theoretical_loss": 3.46142977045409,
      "tokens_seen": 1982136320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403370110330993,
      "loss": 1.3453,
      "theoretical_loss": 3.461420290143404,
      "tokens_seen": 1982201856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040335005015045136,
      "loss": 1.2591,
      "theoretical_loss": 3.461410810233913,
      "tokens_seen": 1982267392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033299899699097,
      "loss": 1.2978,
      "theoretical_loss": 3.461401330725587,
      "tokens_seen": 1982332928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033099297893681,
      "loss": 1.4364,
      "theoretical_loss": 3.4613918516183952,
      "tokens_seen": 1982398464
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040328986960882644,
      "loss": 1.3573,
      "theoretical_loss": 3.4613823729123077,
      "tokens_seen": 1982464000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040326980942828486,
      "loss": 1.4078,
      "theoretical_loss": 3.461372894607294,
      "tokens_seen": 1982529536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032497492477432,
      "loss": 1.3428,
      "theoretical_loss": 3.461363416703324,
      "tokens_seen": 1982595072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032296890672016,
      "loss": 1.3255,
      "theoretical_loss": 3.461353939200367,
      "tokens_seen": 1982660608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040320962888665995,
      "loss": 1.3935,
      "theoretical_loss": 3.4613444620983937,
      "tokens_seen": 1982726144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040318956870611836,
      "loss": 1.3975,
      "theoretical_loss": 3.461334985397374,
      "tokens_seen": 1982791680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031695085255767,
      "loss": 1.3443,
      "theoretical_loss": 3.4613255090972768,
      "tokens_seen": 1982857216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040314944834503514,
      "loss": 1.3687,
      "theoretical_loss": 3.461316033198072,
      "tokens_seen": 1982922752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031293881644935,
      "loss": 1.3576,
      "theoretical_loss": 3.4613065576997295,
      "tokens_seen": 1982988288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040310932798395187,
      "loss": 1.3636,
      "theoretical_loss": 3.4612970826022194,
      "tokens_seen": 1983053824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004030892678034103,
      "loss": 1.2994,
      "theoretical_loss": 3.4612876079055113,
      "tokens_seen": 1983119360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040306920762286864,
      "loss": 1.3355,
      "theoretical_loss": 3.461278133609575,
      "tokens_seen": 1983184896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403049147442327,
      "loss": 1.2429,
      "theoretical_loss": 3.461268659714381,
      "tokens_seen": 1983250432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040302908726178537,
      "loss": 1.4229,
      "theoretical_loss": 3.4612591862198974,
      "tokens_seen": 1983315968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004030090270812438,
      "loss": 1.3125,
      "theoretical_loss": 3.4612497131260955,
      "tokens_seen": 1983381504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040298896690070215,
      "loss": 1.3685,
      "theoretical_loss": 3.4612402404329448,
      "tokens_seen": 1983447040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004029689067201605,
      "loss": 1.4441,
      "theoretical_loss": 3.461230768140415,
      "tokens_seen": 1983512576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040294884653961887,
      "loss": 1.3796,
      "theoretical_loss": 3.4612212962484756,
      "tokens_seen": 1983578112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040292878635907723,
      "loss": 1.2975,
      "theoretical_loss": 3.461211824757097,
      "tokens_seen": 1983643648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040290872617853565,
      "loss": 1.4448,
      "theoretical_loss": 3.4612023536662484,
      "tokens_seen": 1983709184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402888665997994,
      "loss": 1.3734,
      "theoretical_loss": 3.4611928829759004,
      "tokens_seen": 1983774720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028686058174524,
      "loss": 1.2953,
      "theoretical_loss": 3.4611834126860224,
      "tokens_seen": 1983840256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040284854563691074,
      "loss": 1.3181,
      "theoretical_loss": 3.4611739427965835,
      "tokens_seen": 1983905792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040282848545636915,
      "loss": 1.254,
      "theoretical_loss": 3.461164473307555,
      "tokens_seen": 1983971328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028084252758275,
      "loss": 1.3653,
      "theoretical_loss": 3.461155004218906,
      "tokens_seen": 1984036864
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027883650952859,
      "loss": 1.3605,
      "theoretical_loss": 3.4611455355306067,
      "tokens_seen": 1984102400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040276830491474424,
      "loss": 1.3062,
      "theoretical_loss": 3.461136067242626,
      "tokens_seen": 1984167936
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 723589,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2202953100204468,
      "objective/train/theoretical_loss": 3.4611265993549347,
      "objective/train/tokens_used": 354692576,
      "theoretical_loss": 3.4611265993549347,
      "tokens_seen": 1984233472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027482447342026,
      "loss": 1.2569,
      "theoretical_loss": 3.4611265993549347,
      "tokens_seen": 1984233472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402728184553661,
      "loss": 1.4007,
      "theoretical_loss": 3.461117131867502,
      "tokens_seen": 1984299008
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027081243731194,
      "loss": 1.4271,
      "theoretical_loss": 3.4611076647802985,
      "tokens_seen": 1984364544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040268806419257774,
      "loss": 1.3595,
      "theoretical_loss": 3.4610981980932936,
      "tokens_seen": 1984430080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026680040120361,
      "loss": 1.4322,
      "theoretical_loss": 3.461088731806457,
      "tokens_seen": 1984495616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026479438314945,
      "loss": 1.3715,
      "theoretical_loss": 3.461079265919759,
      "tokens_seen": 1984561152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026278836509529,
      "loss": 1.3563,
      "theoretical_loss": 3.4610698004331693,
      "tokens_seen": 1984626688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040260782347041125,
      "loss": 1.425,
      "theoretical_loss": 3.461060335346658,
      "tokens_seen": 1984692224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004025877632898696,
      "loss": 1.3892,
      "theoretical_loss": 3.461050870660194,
      "tokens_seen": 1984757760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040256770310932797,
      "loss": 1.3396,
      "theoretical_loss": 3.4610414063737487,
      "tokens_seen": 1984823296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004025476429287864,
      "loss": 1.4118,
      "theoretical_loss": 3.46103194248729,
      "tokens_seen": 1984888832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040252758274824475,
      "loss": 1.3685,
      "theoretical_loss": 3.4610224790007904,
      "tokens_seen": 1984954368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004025075225677031,
      "loss": 1.3364,
      "theoretical_loss": 3.461013015914218,
      "tokens_seen": 1985019904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024874623871615,
      "loss": 1.3812,
      "theoretical_loss": 3.4610035532275427,
      "tokens_seen": 1985085440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024674022066199,
      "loss": 1.5181,
      "theoretical_loss": 3.4609940909407353,
      "tokens_seen": 1985150976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040244734202607825,
      "loss": 1.4117,
      "theoretical_loss": 3.4609846290537645,
      "tokens_seen": 1985216512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024272818455366,
      "loss": 1.3317,
      "theoretical_loss": 3.4609751675666014,
      "tokens_seen": 1985282048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402407221664995,
      "loss": 1.31,
      "theoretical_loss": 3.4609657064792154,
      "tokens_seen": 1985347584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023871614844534,
      "loss": 1.3652,
      "theoretical_loss": 3.460956245791576,
      "tokens_seen": 1985413120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040236710130391176,
      "loss": 1.3886,
      "theoretical_loss": 3.460946785503654,
      "tokens_seen": 1985478656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023470411233701,
      "loss": 1.2739,
      "theoretical_loss": 3.4609373256154186,
      "tokens_seen": 1985544192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023269809428285,
      "loss": 1.3078,
      "theoretical_loss": 3.4609278661268403,
      "tokens_seen": 1985609728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040230692076228684,
      "loss": 1.3446,
      "theoretical_loss": 3.4609184070378882,
      "tokens_seen": 1985675264
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040228686058174526,
      "loss": 1.4043,
      "theoretical_loss": 3.460908948348533,
      "tokens_seen": 1985740800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004022668004012036,
      "loss": 1.4045,
      "theoretical_loss": 3.460899490058744,
      "tokens_seen": 1985806336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402246740220662,
      "loss": 1.2708,
      "theoretical_loss": 3.460890032168492,
      "tokens_seen": 1985871872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040222668004012035,
      "loss": 1.4782,
      "theoretical_loss": 3.4608805746777467,
      "tokens_seen": 1985937408
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040220661985957876,
      "loss": 1.426,
      "theoretical_loss": 3.460871117586477,
      "tokens_seen": 1986002944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021865596790371,
      "loss": 1.4417,
      "theoretical_loss": 3.460861660894654,
      "tokens_seen": 1986068480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021664994984955,
      "loss": 1.3367,
      "theoretical_loss": 3.460852204602247,
      "tokens_seen": 1986134016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040214643931795385,
      "loss": 1.4435,
      "theoretical_loss": 3.4608427487092266,
      "tokens_seen": 1986199552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021263791374122,
      "loss": 1.4076,
      "theoretical_loss": 3.460833293215562,
      "tokens_seen": 1986265088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040210631895687063,
      "loss": 1.2717,
      "theoretical_loss": 3.4608238381212235,
      "tokens_seen": 1986330624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402086258776329,
      "loss": 1.351,
      "theoretical_loss": 3.4608143834261815,
      "tokens_seen": 1986396160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040206619859578735,
      "loss": 1.4113,
      "theoretical_loss": 3.4608049291304055,
      "tokens_seen": 1986461696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020461384152457,
      "loss": 1.3995,
      "theoretical_loss": 3.4607954752338657,
      "tokens_seen": 1986527232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040202607823470413,
      "loss": 1.4118,
      "theoretical_loss": 3.4607860217365314,
      "tokens_seen": 1986592768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020060180541625,
      "loss": 1.2689,
      "theoretical_loss": 3.460776568638373,
      "tokens_seen": 1986658304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040198595787362086,
      "loss": 1.278,
      "theoretical_loss": 3.460767115939361,
      "tokens_seen": 1986723840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004019658976930792,
      "loss": 1.3531,
      "theoretical_loss": 3.4607576636394644,
      "tokens_seen": 1986789376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004019458375125376,
      "loss": 1.3583,
      "theoretical_loss": 3.460748211738654,
      "tokens_seen": 1986854912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000401925777331996,
      "loss": 1.2634,
      "theoretical_loss": 3.4607387602368997,
      "tokens_seen": 1986920448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040190571715145436,
      "loss": 1.4257,
      "theoretical_loss": 3.4607293091341713,
      "tokens_seen": 1986985984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018856569709127,
      "loss": 1.4302,
      "theoretical_loss": 3.4607198584304384,
      "tokens_seen": 1987051520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018655967903711,
      "loss": 1.3279,
      "theoretical_loss": 3.4607104081256717,
      "tokens_seen": 1987117056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018455366098295,
      "loss": 1.4233,
      "theoretical_loss": 3.4607009582198405,
      "tokens_seen": 1987182592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040182547642928786,
      "loss": 1.3579,
      "theoretical_loss": 3.4606915087129155,
      "tokens_seen": 1987248128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018054162487462,
      "loss": 1.4424,
      "theoretical_loss": 3.460682059604866,
      "tokens_seen": 1987313664
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017853560682046,
      "loss": 1.3945,
      "theoretical_loss": 3.4606726108956627,
      "tokens_seen": 1987379200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040176529588766295,
      "loss": 1.3562,
      "theoretical_loss": 3.4606631625852757,
      "tokens_seen": 1987444736
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 724791,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3534629344940186,
      "objective/train/theoretical_loss": 3.460653714673674,
      "objective/train/tokens_used": 357969376,
      "theoretical_loss": 3.460653714673674,
      "tokens_seen": 1987510272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040174523570712137,
      "loss": 1.32,
      "theoretical_loss": 3.460653714673674,
      "tokens_seen": 1987510272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040172517552657973,
      "loss": 1.2754,
      "theoretical_loss": 3.460644267160829,
      "tokens_seen": 1987575808
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017051153460381,
      "loss": 1.3013,
      "theoretical_loss": 3.4606348200467094,
      "tokens_seen": 1987641344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040168505516549645,
      "loss": 1.3393,
      "theoretical_loss": 3.4606253733312857,
      "tokens_seen": 1987706880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040166499498495487,
      "loss": 1.3689,
      "theoretical_loss": 3.460615927014528,
      "tokens_seen": 1987772416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004016449348044133,
      "loss": 1.2709,
      "theoretical_loss": 3.4606064810964066,
      "tokens_seen": 1987837952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040162487462387165,
      "loss": 1.3312,
      "theoretical_loss": 3.4605970355768916,
      "tokens_seen": 1987903488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040160481444333,
      "loss": 1.4216,
      "theoretical_loss": 3.460587590455952,
      "tokens_seen": 1987969024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040158475426278837,
      "loss": 1.4097,
      "theoretical_loss": 3.4605781457335594,
      "tokens_seen": 1988034560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004015646940822468,
      "loss": 1.3176,
      "theoretical_loss": 3.4605687014096826,
      "tokens_seen": 1988100096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040154463390170515,
      "loss": 1.3375,
      "theoretical_loss": 3.4605592574842916,
      "tokens_seen": 1988165632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004015245737211635,
      "loss": 1.4103,
      "theoretical_loss": 3.460549813957358,
      "tokens_seen": 1988231168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004015045135406219,
      "loss": 1.4437,
      "theoretical_loss": 3.46054037082885,
      "tokens_seen": 1988296704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014844533600803,
      "loss": 1.3659,
      "theoretical_loss": 3.4605309280987386,
      "tokens_seen": 1988362240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040146439317953865,
      "loss": 1.3338,
      "theoretical_loss": 3.4605214857669937,
      "tokens_seen": 1988427776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000401444332998997,
      "loss": 1.4236,
      "theoretical_loss": 3.4605120438335852,
      "tokens_seen": 1988493312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014242728184554,
      "loss": 1.3153,
      "theoretical_loss": 3.4605026022984835,
      "tokens_seen": 1988558848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014042126379138,
      "loss": 1.3989,
      "theoretical_loss": 3.4604931611616587,
      "tokens_seen": 1988624384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040138415245737216,
      "loss": 1.2986,
      "theoretical_loss": 3.4604837204230803,
      "tokens_seen": 1988689920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013640922768305,
      "loss": 1.3591,
      "theoretical_loss": 3.460474280082719,
      "tokens_seen": 1988755456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013440320962889,
      "loss": 1.3271,
      "theoretical_loss": 3.4604648401405447,
      "tokens_seen": 1988820992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040132397191574724,
      "loss": 1.3282,
      "theoretical_loss": 3.460455400596527,
      "tokens_seen": 1988886528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040130391173520566,
      "loss": 1.3491,
      "theoretical_loss": 3.460445961450637,
      "tokens_seen": 1988952064
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000401283851554664,
      "loss": 1.3012,
      "theoretical_loss": 3.4604365227028437,
      "tokens_seen": 1989017600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012637913741224,
      "loss": 1.3187,
      "theoretical_loss": 3.460427084353118,
      "tokens_seen": 1989083136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040124373119358075,
      "loss": 1.4753,
      "theoretical_loss": 3.4604176464014293,
      "tokens_seen": 1989148672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040122367101303916,
      "loss": 1.31,
      "theoretical_loss": 3.4604082088477486,
      "tokens_seen": 1989214208
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012036108324975,
      "loss": 1.3481,
      "theoretical_loss": 3.460398771692045,
      "tokens_seen": 1989279744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011835506519559,
      "loss": 1.3408,
      "theoretical_loss": 3.4603893349342894,
      "tokens_seen": 1989345280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040116349047141425,
      "loss": 1.3436,
      "theoretical_loss": 3.460379898574452,
      "tokens_seen": 1989410816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011434302908726,
      "loss": 1.4249,
      "theoretical_loss": 3.4603704626125023,
      "tokens_seen": 1989476352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040112337011033103,
      "loss": 1.4224,
      "theoretical_loss": 3.4603610270484104,
      "tokens_seen": 1989541888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011033099297894,
      "loss": 1.3064,
      "theoretical_loss": 3.4603515918821466,
      "tokens_seen": 1989607424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040108324974924775,
      "loss": 1.3215,
      "theoretical_loss": 3.460342157113681,
      "tokens_seen": 1989672960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004010631895687061,
      "loss": 1.3517,
      "theoretical_loss": 3.4603327227429843,
      "tokens_seen": 1989738496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040104312938816453,
      "loss": 1.3034,
      "theoretical_loss": 3.460323288770026,
      "tokens_seen": 1989804032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004010230692076229,
      "loss": 1.3197,
      "theoretical_loss": 3.460313855194776,
      "tokens_seen": 1989869568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040100300902708126,
      "loss": 1.334,
      "theoretical_loss": 3.460304422017205,
      "tokens_seen": 1989935104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009829488465396,
      "loss": 1.3495,
      "theoretical_loss": 3.4602949892372834,
      "tokens_seen": 1990000640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400962888665998,
      "loss": 1.3496,
      "theoretical_loss": 3.4602855568549806,
      "tokens_seen": 1990066176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009428284854564,
      "loss": 1.3916,
      "theoretical_loss": 3.4602761248702665,
      "tokens_seen": 1990131712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040092276830491476,
      "loss": 1.3857,
      "theoretical_loss": 3.4602666932831125,
      "tokens_seen": 1990197248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009027081243731,
      "loss": 1.4262,
      "theoretical_loss": 3.4602572620934877,
      "tokens_seen": 1990262784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008826479438315,
      "loss": 1.3856,
      "theoretical_loss": 3.4602478313013627,
      "tokens_seen": 1990328320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008625877632899,
      "loss": 1.392,
      "theoretical_loss": 3.4602384009067073,
      "tokens_seen": 1990393856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040084252758274826,
      "loss": 1.3777,
      "theoretical_loss": 3.460228970909492,
      "tokens_seen": 1990459392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008224674022066,
      "loss": 1.3827,
      "theoretical_loss": 3.4602195413096877,
      "tokens_seen": 1990524928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400802407221665,
      "loss": 1.3606,
      "theoretical_loss": 3.460210112107263,
      "tokens_seen": 1990590464
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040078234704112335,
      "loss": 1.5201,
      "theoretical_loss": 3.4602006833021886,
      "tokens_seen": 1990656000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040076228686058177,
      "loss": 1.4248,
      "theoretical_loss": 3.4601912548944354,
      "tokens_seen": 1990721536
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 726036,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3863608837127686,
      "objective/train/theoretical_loss": 3.4601818268839732,
      "objective/train/tokens_used": 361246176,
      "theoretical_loss": 3.4601818268839732,
      "tokens_seen": 1990787072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040074222668004013,
      "loss": 1.3912,
      "theoretical_loss": 3.4601818268839732,
      "tokens_seen": 1990787072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007221664994985,
      "loss": 1.3423,
      "theoretical_loss": 3.4601723992707716,
      "tokens_seen": 1990852608
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040070210631895685,
      "loss": 1.3434,
      "theoretical_loss": 3.4601629720548015,
      "tokens_seen": 1990918144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040068204613841527,
      "loss": 1.3627,
      "theoretical_loss": 3.460153545236033,
      "tokens_seen": 1990983680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040066198595787363,
      "loss": 1.5416,
      "theoretical_loss": 3.4601441188144353,
      "tokens_seen": 1991049216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400641925777332,
      "loss": 1.4456,
      "theoretical_loss": 3.46013469278998,
      "tokens_seen": 1991114752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040062186559679036,
      "loss": 1.3619,
      "theoretical_loss": 3.4601252671626366,
      "tokens_seen": 1991180288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004006018054162487,
      "loss": 1.3675,
      "theoretical_loss": 3.460115841932376,
      "tokens_seen": 1991245824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040058174523570713,
      "loss": 1.2165,
      "theoretical_loss": 3.4601064170991673,
      "tokens_seen": 1991311360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005616850551655,
      "loss": 1.5023,
      "theoretical_loss": 3.460096992662981,
      "tokens_seen": 1991376896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040054162487462386,
      "loss": 1.4434,
      "theoretical_loss": 3.4600875686237877,
      "tokens_seen": 1991442432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005215646940822,
      "loss": 1.3902,
      "theoretical_loss": 3.4600781449815576,
      "tokens_seen": 1991507968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040050150451354064,
      "loss": 1.3706,
      "theoretical_loss": 3.4600687217362607,
      "tokens_seen": 1991573504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400481444332999,
      "loss": 1.4068,
      "theoretical_loss": 3.4600592988878676,
      "tokens_seen": 1991639040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040046138415245736,
      "loss": 1.3891,
      "theoretical_loss": 3.4600498764363476,
      "tokens_seen": 1991704576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004413239719157,
      "loss": 1.3449,
      "theoretical_loss": 3.460040454381672,
      "tokens_seen": 1991770112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004212637913741,
      "loss": 1.3519,
      "theoretical_loss": 3.46003103272381,
      "tokens_seen": 1991835648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004012036108325,
      "loss": 1.4163,
      "theoretical_loss": 3.460021611462733,
      "tokens_seen": 1991901184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040038114343029087,
      "loss": 1.4043,
      "theoretical_loss": 3.4600121905984103,
      "tokens_seen": 1991966720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040036108324974923,
      "loss": 1.4246,
      "theoretical_loss": 3.4600027701308127,
      "tokens_seen": 1992032256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004003410230692076,
      "loss": 1.2987,
      "theoretical_loss": 3.45999335005991,
      "tokens_seen": 1992097792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400320962888666,
      "loss": 1.3392,
      "theoretical_loss": 3.4599839303856728,
      "tokens_seen": 1992163328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040030090270812437,
      "loss": 1.4051,
      "theoretical_loss": 3.4599745111080713,
      "tokens_seen": 1992228864
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040028084252758273,
      "loss": 1.3026,
      "theoretical_loss": 3.4599650922270753,
      "tokens_seen": 1992294400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004002607823470411,
      "loss": 1.4301,
      "theoretical_loss": 3.4599556737426553,
      "tokens_seen": 1992359936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040024072216649946,
      "loss": 1.3771,
      "theoretical_loss": 3.459946255654782,
      "tokens_seen": 1992425472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040022066198595787,
      "loss": 1.364,
      "theoretical_loss": 3.4599368379634257,
      "tokens_seen": 1992491008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040020060180541623,
      "loss": 1.4313,
      "theoretical_loss": 3.4599274206685555,
      "tokens_seen": 1992556544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001805416248746,
      "loss": 1.3656,
      "theoretical_loss": 3.4599180037701434,
      "tokens_seen": 1992622080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040016048144433296,
      "loss": 1.3189,
      "theoretical_loss": 3.459908587268158,
      "tokens_seen": 1992687616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040014042126379143,
      "loss": 1.4828,
      "theoretical_loss": 3.4598991711625704,
      "tokens_seen": 1992753152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001203610832498,
      "loss": 1.43,
      "theoretical_loss": 3.459889755453351,
      "tokens_seen": 1992818688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040010030090270815,
      "loss": 1.3663,
      "theoretical_loss": 3.45988034014047,
      "tokens_seen": 1992884224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004000802407221665,
      "loss": 1.3134,
      "theoretical_loss": 3.459870925223897,
      "tokens_seen": 1992949760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040006018054162493,
      "loss": 1.4239,
      "theoretical_loss": 3.4598615107036035,
      "tokens_seen": 1993015296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004000401203610833,
      "loss": 1.378,
      "theoretical_loss": 3.4598520965795587,
      "tokens_seen": 1993080832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040002006018054166,
      "loss": 1.3894,
      "theoretical_loss": 3.4598426828517335,
      "tokens_seen": 1993146368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004,
      "loss": 1.2587,
      "theoretical_loss": 3.4598332695200984,
      "tokens_seen": 1993211904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999799398194584,
      "loss": 1.483,
      "theoretical_loss": 3.459823856584623,
      "tokens_seen": 1993277440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999598796389168,
      "loss": 1.3825,
      "theoretical_loss": 3.4598144440452776,
      "tokens_seen": 1993342976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039993981945837516,
      "loss": 1.356,
      "theoretical_loss": 3.459805031902033,
      "tokens_seen": 1993408512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999197592778335,
      "loss": 1.3301,
      "theoretical_loss": 3.459795620154859,
      "tokens_seen": 1993474048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998996990972919,
      "loss": 1.4109,
      "theoretical_loss": 3.4597862088037266,
      "tokens_seen": 1993539584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998796389167503,
      "loss": 1.3448,
      "theoretical_loss": 3.4597767978486065,
      "tokens_seen": 1993605120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039985957873620866,
      "loss": 1.3179,
      "theoretical_loss": 3.4597673872894674,
      "tokens_seen": 1993670656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399839518555667,
      "loss": 1.4707,
      "theoretical_loss": 3.459757977126281,
      "tokens_seen": 1993736192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998194583751254,
      "loss": 1.3044,
      "theoretical_loss": 3.4597485673590165,
      "tokens_seen": 1993801728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039979939819458375,
      "loss": 1.3809,
      "theoretical_loss": 3.459739157987645,
      "tokens_seen": 1993867264
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039977933801404217,
      "loss": 1.3681,
      "theoretical_loss": 3.459729749012137,
      "tokens_seen": 1993932800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039975927783350053,
      "loss": 1.318,
      "theoretical_loss": 3.4597203404324626,
      "tokens_seen": 1993998336
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 727257,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3149657249450684,
      "objective/train/theoretical_loss": 3.459710932248592,
      "objective/train/tokens_used": 364522976,
      "theoretical_loss": 3.459710932248592,
      "tokens_seen": 1994063872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003997392176529589,
      "loss": 1.2551,
      "theoretical_loss": 3.459710932248592,
      "tokens_seen": 1994063872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039971915747241725,
      "loss": 1.3166,
      "theoretical_loss": 3.4597015244604954,
      "tokens_seen": 1994129408
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039969909729187567,
      "loss": 1.2919,
      "theoretical_loss": 3.459692117068144,
      "tokens_seen": 1994194944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039967903711133403,
      "loss": 1.3725,
      "theoretical_loss": 3.459682710071507,
      "tokens_seen": 1994260480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996589769307924,
      "loss": 1.4415,
      "theoretical_loss": 3.459673303470555,
      "tokens_seen": 1994326016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039963891675025076,
      "loss": 1.3836,
      "theoretical_loss": 3.4596638972652594,
      "tokens_seen": 1994391552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996188565697091,
      "loss": 1.3605,
      "theoretical_loss": 3.4596544914555887,
      "tokens_seen": 1994457088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039959879638916754,
      "loss": 1.3941,
      "theoretical_loss": 3.4596450860415153,
      "tokens_seen": 1994522624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003995787362086259,
      "loss": 1.4052,
      "theoretical_loss": 3.4596356810230082,
      "tokens_seen": 1994588160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039955867602808426,
      "loss": 1.3498,
      "theoretical_loss": 3.459626276400038,
      "tokens_seen": 1994653696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003995386158475426,
      "loss": 1.4425,
      "theoretical_loss": 3.4596168721725755,
      "tokens_seen": 1994719232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039951855566700104,
      "loss": 1.2963,
      "theoretical_loss": 3.459607468340591,
      "tokens_seen": 1994784768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994984954864594,
      "loss": 1.2997,
      "theoretical_loss": 3.4595980649040543,
      "tokens_seen": 1994850304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039947843530591776,
      "loss": 1.34,
      "theoretical_loss": 3.459588661862936,
      "tokens_seen": 1994915840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994583751253761,
      "loss": 1.4009,
      "theoretical_loss": 3.459579259217207,
      "tokens_seen": 1994981376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994383149448345,
      "loss": 1.4091,
      "theoretical_loss": 3.459569856966837,
      "tokens_seen": 1995046912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994182547642929,
      "loss": 1.3064,
      "theoretical_loss": 3.4595604551117973,
      "tokens_seen": 1995112448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039939819458375127,
      "loss": 1.2864,
      "theoretical_loss": 3.4595510536520573,
      "tokens_seen": 1995177984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039937813440320963,
      "loss": 1.3486,
      "theoretical_loss": 3.459541652587588,
      "tokens_seen": 1995243520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399358074222668,
      "loss": 1.4531,
      "theoretical_loss": 3.4595322519183593,
      "tokens_seen": 1995309056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993380140421264,
      "loss": 1.4836,
      "theoretical_loss": 3.4595228516443424,
      "tokens_seen": 1995374592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039931795386158477,
      "loss": 1.3949,
      "theoretical_loss": 3.459513451765507,
      "tokens_seen": 1995440128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039929789368104313,
      "loss": 1.3453,
      "theoretical_loss": 3.4595040522818232,
      "tokens_seen": 1995505664
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003992778335005015,
      "loss": 1.2817,
      "theoretical_loss": 3.4594946531932624,
      "tokens_seen": 1995571200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039925777331995986,
      "loss": 1.3728,
      "theoretical_loss": 3.459485254499795,
      "tokens_seen": 1995636736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039923771313941827,
      "loss": 1.3255,
      "theoretical_loss": 3.4594758562013905,
      "tokens_seen": 1995702272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039921765295887663,
      "loss": 1.3223,
      "theoretical_loss": 3.45946645829802,
      "tokens_seen": 1995767808
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399197592778335,
      "loss": 1.3109,
      "theoretical_loss": 3.4594570607896538,
      "tokens_seen": 1995833344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039917753259779336,
      "loss": 1.2799,
      "theoretical_loss": 3.4594476636762614,
      "tokens_seen": 1995898880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003991574724172518,
      "loss": 1.3226,
      "theoretical_loss": 3.459438266957815,
      "tokens_seen": 1995964416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039913741223671014,
      "loss": 1.2797,
      "theoretical_loss": 3.4594288706342837,
      "tokens_seen": 1996029952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003991173520561685,
      "loss": 1.2252,
      "theoretical_loss": 3.459419474705639,
      "tokens_seen": 1996095488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039909729187562686,
      "loss": 1.2257,
      "theoretical_loss": 3.4594100791718496,
      "tokens_seen": 1996161024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003990772316950853,
      "loss": 1.4258,
      "theoretical_loss": 3.4594006840328877,
      "tokens_seen": 1996226560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039905717151454364,
      "loss": 1.3857,
      "theoretical_loss": 3.459391289288723,
      "tokens_seen": 1996292096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399037111334002,
      "loss": 1.3814,
      "theoretical_loss": 3.459381894939326,
      "tokens_seen": 1996357632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039901705115346037,
      "loss": 1.3452,
      "theoretical_loss": 3.459372500984667,
      "tokens_seen": 1996423168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039899699097291873,
      "loss": 1.2958,
      "theoretical_loss": 3.459363107424717,
      "tokens_seen": 1996488704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039897693079237714,
      "loss": 1.4375,
      "theoretical_loss": 3.459353714259446,
      "tokens_seen": 1996554240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003989568706118355,
      "loss": 1.4296,
      "theoretical_loss": 3.4593443214888246,
      "tokens_seen": 1996619776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039893681043129387,
      "loss": 1.3543,
      "theoretical_loss": 3.4593349291128233,
      "tokens_seen": 1996685312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039891675025075223,
      "loss": 1.3687,
      "theoretical_loss": 3.459325537131412,
      "tokens_seen": 1996750848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039889669007021065,
      "loss": 1.3407,
      "theoretical_loss": 3.459316145544562,
      "tokens_seen": 1996816384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398876629889669,
      "loss": 1.4782,
      "theoretical_loss": 3.4593067543522436,
      "tokens_seen": 1996881920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039885656970912737,
      "loss": 1.3607,
      "theoretical_loss": 3.459297363554427,
      "tokens_seen": 1996947456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039883650952858573,
      "loss": 1.3666,
      "theoretical_loss": 3.4592879731510826,
      "tokens_seen": 1997012992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003988164493480441,
      "loss": 1.3864,
      "theoretical_loss": 3.459278583142181,
      "tokens_seen": 1997078528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987963891675025,
      "loss": 1.4695,
      "theoretical_loss": 3.4592691935276934,
      "tokens_seen": 1997144064
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987763289869609,
      "loss": 1.3558,
      "theoretical_loss": 3.4592598043075893,
      "tokens_seen": 1997209600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039875626880641924,
      "loss": 1.367,
      "theoretical_loss": 3.4592504154818395,
      "tokens_seen": 1997275136
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 728467,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3009185791015625,
      "objective/train/theoretical_loss": 3.459241027050415,
      "objective/train/tokens_used": 367799776,
      "theoretical_loss": 3.459241027050415,
      "tokens_seen": 1997340672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987362086258776,
      "loss": 1.3176,
      "theoretical_loss": 3.459241027050415,
      "tokens_seen": 1997340672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398716148445336,
      "loss": 1.3363,
      "theoretical_loss": 3.4592316390132853,
      "tokens_seen": 1997406208
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986960882647944,
      "loss": 1.3445,
      "theoretical_loss": 3.459222251370422,
      "tokens_seen": 1997471744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039867602808425274,
      "loss": 1.4535,
      "theoretical_loss": 3.4592128641217945,
      "tokens_seen": 1997537280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986559679037111,
      "loss": 1.3713,
      "theoretical_loss": 3.459203477267374,
      "tokens_seen": 1997602816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986359077231695,
      "loss": 1.4328,
      "theoretical_loss": 3.459194090807131,
      "tokens_seen": 1997668352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039861584754262794,
      "loss": 1.382,
      "theoretical_loss": 3.4591847047410362,
      "tokens_seen": 1997733888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985957873620863,
      "loss": 1.3076,
      "theoretical_loss": 3.4591753190690593,
      "tokens_seen": 1997799424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039857572718154466,
      "loss": 1.2875,
      "theoretical_loss": 3.459165933791172,
      "tokens_seen": 1997864960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398555667001003,
      "loss": 1.4044,
      "theoretical_loss": 3.4591565489073437,
      "tokens_seen": 1997930496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039853560682046144,
      "loss": 1.2351,
      "theoretical_loss": 3.4591471644175456,
      "tokens_seen": 1997996032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985155466399198,
      "loss": 1.2581,
      "theoretical_loss": 3.4591377803217482,
      "tokens_seen": 1998061568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039849548645937816,
      "loss": 1.3277,
      "theoretical_loss": 3.4591283966199216,
      "tokens_seen": 1998127104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984754262788365,
      "loss": 1.3737,
      "theoretical_loss": 3.4591190133120366,
      "tokens_seen": 1998192640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984553660982949,
      "loss": 1.2965,
      "theoretical_loss": 3.459109630398064,
      "tokens_seen": 1998258176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984353059177533,
      "loss": 1.4011,
      "theoretical_loss": 3.4591002478779744,
      "tokens_seen": 1998323712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039841524573721167,
      "loss": 1.3383,
      "theoretical_loss": 3.4590908657517376,
      "tokens_seen": 1998389248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039839518555667003,
      "loss": 1.4244,
      "theoretical_loss": 3.4590814840193245,
      "tokens_seen": 1998454784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983751253761284,
      "loss": 1.3942,
      "theoretical_loss": 3.459072102680706,
      "tokens_seen": 1998520320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983550651955868,
      "loss": 1.3323,
      "theoretical_loss": 3.4590627217358523,
      "tokens_seen": 1998585856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039833500501504517,
      "loss": 1.4813,
      "theoretical_loss": 3.4590533411847346,
      "tokens_seen": 1998651392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039831494483450353,
      "loss": 1.2905,
      "theoretical_loss": 3.459043961027322,
      "tokens_seen": 1998716928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982948846539619,
      "loss": 1.3933,
      "theoretical_loss": 3.4590345812635865,
      "tokens_seen": 1998782464
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039827482447342026,
      "loss": 1.2936,
      "theoretical_loss": 3.459025201893499,
      "tokens_seen": 1998848000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039825476429287867,
      "loss": 1.3192,
      "theoretical_loss": 3.459015822917028,
      "tokens_seen": 1998913536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039823470411233703,
      "loss": 1.4075,
      "theoretical_loss": 3.459006444334146,
      "tokens_seen": 1998979072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982146439317954,
      "loss": 1.3472,
      "theoretical_loss": 3.4589970661448226,
      "tokens_seen": 1999044608
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039819458375125376,
      "loss": 1.414,
      "theoretical_loss": 3.458987688349029,
      "tokens_seen": 1999110144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003981745235707122,
      "loss": 1.2937,
      "theoretical_loss": 3.458978310946735,
      "tokens_seen": 1999175680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039815446339017054,
      "loss": 1.3359,
      "theoretical_loss": 3.4589689339379124,
      "tokens_seen": 1999241216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003981344032096289,
      "loss": 1.3613,
      "theoretical_loss": 3.4589595573225305,
      "tokens_seen": 1999306752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039811434302908726,
      "loss": 1.3896,
      "theoretical_loss": 3.458950181100561,
      "tokens_seen": 1999372288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003980942828485457,
      "loss": 1.4136,
      "theoretical_loss": 3.4589408052719737,
      "tokens_seen": 1999437824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039807422266800404,
      "loss": 1.2264,
      "theoretical_loss": 3.4589314298367393,
      "tokens_seen": 1999503360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003980541624874624,
      "loss": 1.3188,
      "theoretical_loss": 3.4589220547948285,
      "tokens_seen": 1999568896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039803410230692077,
      "loss": 1.4333,
      "theoretical_loss": 3.4589126801462124,
      "tokens_seen": 1999634432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039801404212637913,
      "loss": 1.2825,
      "theoretical_loss": 3.458903305890861,
      "tokens_seen": 1999699968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039799398194583754,
      "loss": 1.5012,
      "theoretical_loss": 3.4588939320287455,
      "tokens_seen": 1999765504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003979739217652959,
      "loss": 1.4465,
      "theoretical_loss": 3.458884558559835,
      "tokens_seen": 1999831040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039795386158475427,
      "loss": 1.4439,
      "theoretical_loss": 3.4588751854841027,
      "tokens_seen": 1999896576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039793380140421263,
      "loss": 1.426,
      "theoretical_loss": 3.4588658128015166,
      "tokens_seen": 1999962112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039791374122367105,
      "loss": 1.241,
      "theoretical_loss": 3.458856440512049,
      "tokens_seen": 2000027648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978936810431294,
      "loss": 1.3287,
      "theoretical_loss": 3.45884706861567,
      "tokens_seen": 2000093184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039787362086258777,
      "loss": 1.2514,
      "theoretical_loss": 3.4588376971123505,
      "tokens_seen": 2000158720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039785356068204613,
      "loss": 1.4046,
      "theoretical_loss": 3.4588283260020605,
      "tokens_seen": 2000224256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978335005015045,
      "loss": 1.3508,
      "theoretical_loss": 3.4588189552847712,
      "tokens_seen": 2000289792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978134403209629,
      "loss": 1.3497,
      "theoretical_loss": 3.458809584960453,
      "tokens_seen": 2000355328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977933801404213,
      "loss": 1.3653,
      "theoretical_loss": 3.458800215029077,
      "tokens_seen": 2000420864
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039777331995987964,
      "loss": 1.45,
      "theoretical_loss": 3.458790845490613,
      "tokens_seen": 2000486400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397753259779338,
      "loss": 1.3146,
      "theoretical_loss": 3.458781476345033,
      "tokens_seen": 2000551936
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 729542,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.47444748878479,
      "objective/train/theoretical_loss": 3.4587721075923055,
      "objective/train/tokens_used": 371076576,
      "theoretical_loss": 3.4587721075923055,
      "tokens_seen": 2000617472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977331995987964,
      "loss": 1.3538,
      "theoretical_loss": 3.4587721075923055,
      "tokens_seen": 2000617472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977131394182548,
      "loss": 1.4346,
      "theoretical_loss": 3.4587627392324034,
      "tokens_seen": 2000683008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039769307923771314,
      "loss": 1.3217,
      "theoretical_loss": 3.458753371265296,
      "tokens_seen": 2000748544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976730190571715,
      "loss": 1.4032,
      "theoretical_loss": 3.4587440036909545,
      "tokens_seen": 2000814080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039765295887662987,
      "loss": 1.2932,
      "theoretical_loss": 3.4587346365093494,
      "tokens_seen": 2000879616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976328986960883,
      "loss": 1.3539,
      "theoretical_loss": 3.4587252697204516,
      "tokens_seen": 2000945152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039761283851554664,
      "loss": 1.3534,
      "theoretical_loss": 3.458715903324231,
      "tokens_seen": 2001010688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397592778335005,
      "loss": 1.3157,
      "theoretical_loss": 3.45870653732066,
      "tokens_seen": 2001076224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039757271815446337,
      "loss": 1.2682,
      "theoretical_loss": 3.4586971717097077,
      "tokens_seen": 2001141760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975526579739218,
      "loss": 1.4524,
      "theoretical_loss": 3.4586878064913447,
      "tokens_seen": 2001207296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039753259779338015,
      "loss": 1.2751,
      "theoretical_loss": 3.4586784416655423,
      "tokens_seen": 2001272832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975125376128385,
      "loss": 1.4064,
      "theoretical_loss": 3.4586690772322717,
      "tokens_seen": 2001338368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039749247743229687,
      "loss": 1.4282,
      "theoretical_loss": 3.4586597131915027,
      "tokens_seen": 2001403904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039747241725175523,
      "loss": 1.4682,
      "theoretical_loss": 3.4586503495432064,
      "tokens_seen": 2001469440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039745235707121365,
      "loss": 1.3423,
      "theoretical_loss": 3.458640986287353,
      "tokens_seen": 2001534976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397432296890672,
      "loss": 1.3657,
      "theoretical_loss": 3.4586316234239143,
      "tokens_seen": 2001600512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003974122367101304,
      "loss": 1.4197,
      "theoretical_loss": 3.45862226095286,
      "tokens_seen": 2001666048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039739217652958874,
      "loss": 1.2928,
      "theoretical_loss": 3.458612898874161,
      "tokens_seen": 2001731584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039737211634904715,
      "loss": 1.4058,
      "theoretical_loss": 3.458603537187788,
      "tokens_seen": 2001797120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973520561685055,
      "loss": 1.4261,
      "theoretical_loss": 3.458594175893712,
      "tokens_seen": 2001862656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973319959879639,
      "loss": 1.3361,
      "theoretical_loss": 3.458584814991904,
      "tokens_seen": 2001928192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039731193580742224,
      "loss": 1.325,
      "theoretical_loss": 3.4585754544823337,
      "tokens_seen": 2001993728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972918756268806,
      "loss": 1.3129,
      "theoretical_loss": 3.458566094364973,
      "tokens_seen": 2002059264
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397271815446339,
      "loss": 1.3976,
      "theoretical_loss": 3.4585567346397914,
      "tokens_seen": 2002124800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972517552657974,
      "loss": 1.2419,
      "theoretical_loss": 3.458547375306761,
      "tokens_seen": 2002190336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039723169508525574,
      "loss": 1.2932,
      "theoretical_loss": 3.4585380163658512,
      "tokens_seen": 2002255872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972116349047141,
      "loss": 1.4214,
      "theoretical_loss": 3.4585286578170336,
      "tokens_seen": 2002321408
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971915747241725,
      "loss": 1.3653,
      "theoretical_loss": 3.4585192996602787,
      "tokens_seen": 2002386944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971715145436309,
      "loss": 1.3961,
      "theoretical_loss": 3.4585099418955574,
      "tokens_seen": 2002452480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039715145436308925,
      "loss": 1.3751,
      "theoretical_loss": 3.45850058452284,
      "tokens_seen": 2002518016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039713139418254766,
      "loss": 1.3466,
      "theoretical_loss": 3.4584912275420976,
      "tokens_seen": 2002583552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971113340020061,
      "loss": 1.319,
      "theoretical_loss": 3.458481870953301,
      "tokens_seen": 2002649088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039709127382146444,
      "loss": 1.4854,
      "theoretical_loss": 3.4584725147564206,
      "tokens_seen": 2002714624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003970712136409228,
      "loss": 1.4173,
      "theoretical_loss": 3.4584631589514276,
      "tokens_seen": 2002780160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039705115346038117,
      "loss": 1.3369,
      "theoretical_loss": 3.4584538035382923,
      "tokens_seen": 2002845696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039703109327983953,
      "loss": 1.3493,
      "theoretical_loss": 3.4584444485169863,
      "tokens_seen": 2002911232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039701103309929794,
      "loss": 1.2424,
      "theoretical_loss": 3.4584350938874797,
      "tokens_seen": 2002976768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003969909729187563,
      "loss": 1.4301,
      "theoretical_loss": 3.4584257396497433,
      "tokens_seen": 2003042304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039697091273821467,
      "loss": 1.3657,
      "theoretical_loss": 3.4584163858037478,
      "tokens_seen": 2003107840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039695085255767303,
      "loss": 1.3571,
      "theoretical_loss": 3.458407032349464,
      "tokens_seen": 2003173376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039693079237713145,
      "loss": 1.3425,
      "theoretical_loss": 3.458397679286863,
      "tokens_seen": 2003238912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003969107321965898,
      "loss": 1.4178,
      "theoretical_loss": 3.4583883266159154,
      "tokens_seen": 2003304448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039689067201604817,
      "loss": 1.3632,
      "theoretical_loss": 3.458378974336592,
      "tokens_seen": 2003369984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039687061183550653,
      "loss": 1.4131,
      "theoretical_loss": 3.4583696224488634,
      "tokens_seen": 2003435520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968505516549649,
      "loss": 1.4024,
      "theoretical_loss": 3.458360270952701,
      "tokens_seen": 2003501056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968304914744233,
      "loss": 1.3462,
      "theoretical_loss": 3.4583509198480744,
      "tokens_seen": 2003566592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968104312938817,
      "loss": 1.3712,
      "theoretical_loss": 3.458341569134956,
      "tokens_seen": 2003632128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039679037111334004,
      "loss": 1.3217,
      "theoretical_loss": 3.4583322188133154,
      "tokens_seen": 2003697664
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967703109327984,
      "loss": 1.3755,
      "theoretical_loss": 3.458322868883124,
      "tokens_seen": 2003763200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967502507522568,
      "loss": 1.3514,
      "theoretical_loss": 3.458313519344352,
      "tokens_seen": 2003828736
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 730811,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5580706596374512,
      "objective/train/theoretical_loss": 3.4583041701969712,
      "objective/train/tokens_used": 374353376,
      "theoretical_loss": 3.4583041701969712,
      "tokens_seen": 2003894272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967301905717152,
      "loss": 1.4292,
      "theoretical_loss": 3.4583041701969712,
      "tokens_seen": 2003894272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039671013039117354,
      "loss": 1.5256,
      "theoretical_loss": 3.4582948214409512,
      "tokens_seen": 2003959808
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966900702106319,
      "loss": 1.3629,
      "theoretical_loss": 3.4582854730762635,
      "tokens_seen": 2004025344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039667001003009027,
      "loss": 1.3278,
      "theoretical_loss": 3.4582761251028793,
      "tokens_seen": 2004090880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966499498495487,
      "loss": 1.3935,
      "theoretical_loss": 3.4582667775207687,
      "tokens_seen": 2004156416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039662988966900704,
      "loss": 1.3684,
      "theoretical_loss": 3.458257430329903,
      "tokens_seen": 2004221952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966098294884654,
      "loss": 1.3354,
      "theoretical_loss": 3.4582480835302527,
      "tokens_seen": 2004287488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039658976930792377,
      "loss": 1.3821,
      "theoretical_loss": 3.458238737121789,
      "tokens_seen": 2004353024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965697091273822,
      "loss": 1.4094,
      "theoretical_loss": 3.4582293911044824,
      "tokens_seen": 2004418560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039654964894684055,
      "loss": 1.3391,
      "theoretical_loss": 3.4582200454783036,
      "tokens_seen": 2004484096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965295887662989,
      "loss": 1.3748,
      "theoretical_loss": 3.4582107002432236,
      "tokens_seen": 2004549632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039650952858575727,
      "loss": 1.3999,
      "theoretical_loss": 3.458201355399214,
      "tokens_seen": 2004615168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039648946840521563,
      "loss": 1.3592,
      "theoretical_loss": 3.4581920109462443,
      "tokens_seen": 2004680704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039646940822467405,
      "loss": 1.3163,
      "theoretical_loss": 3.4581826668842863,
      "tokens_seen": 2004746240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003964493480441324,
      "loss": 1.3373,
      "theoretical_loss": 3.4581733232133107,
      "tokens_seen": 2004811776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003964292878635908,
      "loss": 1.4029,
      "theoretical_loss": 3.4581639799332886,
      "tokens_seen": 2004877312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039640922768304914,
      "loss": 1.3991,
      "theoretical_loss": 3.4581546370441902,
      "tokens_seen": 2004942848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039638916750250755,
      "loss": 1.3212,
      "theoretical_loss": 3.4581452945459867,
      "tokens_seen": 2005008384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963691073219659,
      "loss": 1.3337,
      "theoretical_loss": 3.4581359524386492,
      "tokens_seen": 2005073920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963490471414243,
      "loss": 1.4318,
      "theoretical_loss": 3.4581266107221484,
      "tokens_seen": 2005139456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039632898696088264,
      "loss": 1.4976,
      "theoretical_loss": 3.458117269396455,
      "tokens_seen": 2005204992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000396308926780341,
      "loss": 1.3582,
      "theoretical_loss": 3.4581079284615397,
      "tokens_seen": 2005270528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962888665997994,
      "loss": 1.2979,
      "theoretical_loss": 3.458098587917374,
      "tokens_seen": 2005336064
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962688064192578,
      "loss": 1.3085,
      "theoretical_loss": 3.4580892477639282,
      "tokens_seen": 2005401600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039624874623871614,
      "loss": 1.4137,
      "theoretical_loss": 3.4580799080011735,
      "tokens_seen": 2005467136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962286860581745,
      "loss": 1.3146,
      "theoretical_loss": 3.458070568629081,
      "tokens_seen": 2005532672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962086258776329,
      "loss": 1.3594,
      "theoretical_loss": 3.4580612296476207,
      "tokens_seen": 2005598208
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961885656970913,
      "loss": 1.3144,
      "theoretical_loss": 3.458051891056765,
      "tokens_seen": 2005663744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039616850551654965,
      "loss": 1.4711,
      "theoretical_loss": 3.458042552856484,
      "tokens_seen": 2005729280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000396148445336008,
      "loss": 1.4839,
      "theoretical_loss": 3.458033215046748,
      "tokens_seen": 2005794816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039612838515546637,
      "loss": 1.1741,
      "theoretical_loss": 3.4580238776275283,
      "tokens_seen": 2005860352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961083249749248,
      "loss": 1.3387,
      "theoretical_loss": 3.4580145405987963,
      "tokens_seen": 2005925888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039608826479438315,
      "loss": 1.4007,
      "theoretical_loss": 3.4580052039605222,
      "tokens_seen": 2005991424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003960682046138415,
      "loss": 1.4019,
      "theoretical_loss": 3.457995867712678,
      "tokens_seen": 2006056960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003960481444332999,
      "loss": 1.3817,
      "theoretical_loss": 3.4579865318552336,
      "tokens_seen": 2006122496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003960280842527583,
      "loss": 1.3628,
      "theoretical_loss": 3.45797719638816,
      "tokens_seen": 2006188032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039600802407221665,
      "loss": 1.2832,
      "theoretical_loss": 3.4579678613114284,
      "tokens_seen": 2006253568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000395987963891675,
      "loss": 1.3931,
      "theoretical_loss": 3.45795852662501,
      "tokens_seen": 2006319104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003959679037111334,
      "loss": 1.2503,
      "theoretical_loss": 3.457949192328875,
      "tokens_seen": 2006384640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003959478435305918,
      "loss": 1.3403,
      "theoretical_loss": 3.4579398584229954,
      "tokens_seen": 2006450176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039592778335005016,
      "loss": 1.4214,
      "theoretical_loss": 3.457930524907341,
      "tokens_seen": 2006515712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003959077231695085,
      "loss": 1.3907,
      "theoretical_loss": 3.457921191781883,
      "tokens_seen": 2006581248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958876629889669,
      "loss": 1.3735,
      "theoretical_loss": 3.457911859046593,
      "tokens_seen": 2006646784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039586760280842524,
      "loss": 1.3382,
      "theoretical_loss": 3.4579025267014414,
      "tokens_seen": 2006712320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039584754262788366,
      "loss": 1.4067,
      "theoretical_loss": 3.4578931947464,
      "tokens_seen": 2006777856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000395827482447342,
      "loss": 1.3581,
      "theoretical_loss": 3.4578838631814377,
      "tokens_seen": 2006843392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958074222668004,
      "loss": 1.1904,
      "theoretical_loss": 3.457874532006527,
      "tokens_seen": 2006908928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039578736208625875,
      "loss": 1.3163,
      "theoretical_loss": 3.4578652012216393,
      "tokens_seen": 2006974464
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039576730190571716,
      "loss": 1.3017,
      "theoretical_loss": 3.457855870826745,
      "tokens_seen": 2007040000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957472417251755,
      "loss": 1.3272,
      "theoretical_loss": 3.4578465408218144,
      "tokens_seen": 2007105536
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 732034,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3788890838623047,
      "objective/train/theoretical_loss": 3.4578372112068196,
      "objective/train/tokens_used": 377630176,
      "theoretical_loss": 3.4578372112068196,
      "tokens_seen": 2007171072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957271815446339,
      "loss": 1.4508,
      "theoretical_loss": 3.4578372112068196,
      "tokens_seen": 2007171072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039570712136409225,
      "loss": 1.2763,
      "theoretical_loss": 3.4578278819817303,
      "tokens_seen": 2007236608
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956870611835506,
      "loss": 1.5464,
      "theoretical_loss": 3.4578185531465184,
      "tokens_seen": 2007302144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039566700100300903,
      "loss": 1.3669,
      "theoretical_loss": 3.457809224701155,
      "tokens_seen": 2007367680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956469408224674,
      "loss": 1.3768,
      "theoretical_loss": 3.4577998966456107,
      "tokens_seen": 2007433216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956268806419258,
      "loss": 1.2839,
      "theoretical_loss": 3.4577905689798563,
      "tokens_seen": 2007498752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039560682046138417,
      "loss": 1.4591,
      "theoretical_loss": 3.4577812417038634,
      "tokens_seen": 2007564288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955867602808426,
      "loss": 1.461,
      "theoretical_loss": 3.4577719148176023,
      "tokens_seen": 2007629824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039556670010030095,
      "loss": 1.2743,
      "theoretical_loss": 3.457762588321044,
      "tokens_seen": 2007695360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955466399197593,
      "loss": 1.4023,
      "theoretical_loss": 3.4577532622141605,
      "tokens_seen": 2007760896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039552657973921767,
      "loss": 1.2974,
      "theoretical_loss": 3.4577439364969216,
      "tokens_seen": 2007826432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039550651955867603,
      "loss": 1.3581,
      "theoretical_loss": 3.4577346111692995,
      "tokens_seen": 2007891968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039548645937813445,
      "loss": 1.2333,
      "theoretical_loss": 3.4577252862312635,
      "tokens_seen": 2007957504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954663991975928,
      "loss": 1.3381,
      "theoretical_loss": 3.457715961682786,
      "tokens_seen": 2008023040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954463390170512,
      "loss": 1.1589,
      "theoretical_loss": 3.457706637523838,
      "tokens_seen": 2008088576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039542627883650954,
      "loss": 1.4452,
      "theoretical_loss": 3.45769731375439,
      "tokens_seen": 2008154112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039540621865596795,
      "loss": 1.375,
      "theoretical_loss": 3.457687990374413,
      "tokens_seen": 2008219648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953861584754263,
      "loss": 1.2629,
      "theoretical_loss": 3.4576786673838784,
      "tokens_seen": 2008285184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953660982948847,
      "loss": 1.342,
      "theoretical_loss": 3.457669344782757,
      "tokens_seen": 2008350720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039534603811434304,
      "loss": 1.4123,
      "theoretical_loss": 3.4576600225710195,
      "tokens_seen": 2008416256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953259779338014,
      "loss": 1.3774,
      "theoretical_loss": 3.457650700748638,
      "tokens_seen": 2008481792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953059177532598,
      "loss": 1.2621,
      "theoretical_loss": 3.4576413793155822,
      "tokens_seen": 2008547328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952858575727182,
      "loss": 1.4589,
      "theoretical_loss": 3.457632058271824,
      "tokens_seen": 2008612864
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039526579739217654,
      "loss": 1.3776,
      "theoretical_loss": 3.457622737617334,
      "tokens_seen": 2008678400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952457372116349,
      "loss": 1.3602,
      "theoretical_loss": 3.4576134173520834,
      "tokens_seen": 2008743936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952256770310933,
      "loss": 1.301,
      "theoretical_loss": 3.457604097476043,
      "tokens_seen": 2008809472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952056168505517,
      "loss": 1.3629,
      "theoretical_loss": 3.457594777989185,
      "tokens_seen": 2008875008
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039518555667001005,
      "loss": 1.3474,
      "theoretical_loss": 3.457585458891479,
      "tokens_seen": 2008940544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951654964894684,
      "loss": 1.386,
      "theoretical_loss": 3.457576140182897,
      "tokens_seen": 2009006080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039514543630892677,
      "loss": 1.3161,
      "theoretical_loss": 3.4575668218634092,
      "tokens_seen": 2009071616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951253761283852,
      "loss": 1.4063,
      "theoretical_loss": 3.4575575039329873,
      "tokens_seen": 2009137152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039510531594784355,
      "loss": 1.2974,
      "theoretical_loss": 3.457548186391602,
      "tokens_seen": 2009202688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950852557673019,
      "loss": 1.4151,
      "theoretical_loss": 3.457538869239225,
      "tokens_seen": 2009268224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950651955867603,
      "loss": 1.32,
      "theoretical_loss": 3.457529552475827,
      "tokens_seen": 2009333760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950451354062187,
      "loss": 1.4202,
      "theoretical_loss": 3.457520236101379,
      "tokens_seen": 2009399296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039502507522567705,
      "loss": 1.4361,
      "theoretical_loss": 3.457510920115852,
      "tokens_seen": 2009464832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950050150451354,
      "loss": 1.4331,
      "theoretical_loss": 3.4575016045192175,
      "tokens_seen": 2009530368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949849548645938,
      "loss": 1.3594,
      "theoretical_loss": 3.457492289311446,
      "tokens_seen": 2009595904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949648946840522,
      "loss": 1.4339,
      "theoretical_loss": 3.4574829744925086,
      "tokens_seen": 2009661440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039494483450351056,
      "loss": 1.3886,
      "theoretical_loss": 3.4574736600623774,
      "tokens_seen": 2009726976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949247743229689,
      "loss": 1.4655,
      "theoretical_loss": 3.457464346021022,
      "tokens_seen": 2009792512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949047141424273,
      "loss": 1.3693,
      "theoretical_loss": 3.4574550323684146,
      "tokens_seen": 2009858048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039488465396188564,
      "loss": 1.4258,
      "theoretical_loss": 3.457445719104526,
      "tokens_seen": 2009923584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039486459378134406,
      "loss": 1.353,
      "theoretical_loss": 3.457436406229327,
      "tokens_seen": 2009989120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003948445336008024,
      "loss": 1.3093,
      "theoretical_loss": 3.457427093742789,
      "tokens_seen": 2010054656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003948244734202608,
      "loss": 1.3077,
      "theoretical_loss": 3.4574177816448834,
      "tokens_seen": 2010120192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039480441323971915,
      "loss": 1.3563,
      "theoretical_loss": 3.4574084699355807,
      "tokens_seen": 2010185728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039478435305917756,
      "loss": 1.3726,
      "theoretical_loss": 3.4573991586148525,
      "tokens_seen": 2010251264
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947642928786359,
      "loss": 1.4283,
      "theoretical_loss": 3.4573898476826694,
      "tokens_seen": 2010316800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947442326980943,
      "loss": 1.3414,
      "theoretical_loss": 3.457380537139003,
      "tokens_seen": 2010382336
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 733267,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.300271987915039,
      "objective/train/theoretical_loss": 3.457371226983824,
      "objective/train/tokens_used": 380906976,
      "theoretical_loss": 3.457371226983824,
      "tokens_seen": 2010447872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039472417251755265,
      "loss": 1.3223,
      "theoretical_loss": 3.457371226983824,
      "tokens_seen": 2010447872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394704112337011,
      "loss": 1.3571,
      "theoretical_loss": 3.457361917217104,
      "tokens_seen": 2010513408
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039468405215646943,
      "loss": 1.3317,
      "theoretical_loss": 3.457352607838814,
      "tokens_seen": 2010578944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946639919759278,
      "loss": 1.3115,
      "theoretical_loss": 3.457343298848925,
      "tokens_seen": 2010644480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039464393179538615,
      "loss": 1.3378,
      "theoretical_loss": 3.4573339902474083,
      "tokens_seen": 2010710016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946238716148445,
      "loss": 1.263,
      "theoretical_loss": 3.4573246820342347,
      "tokens_seen": 2010775552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039460381143430293,
      "loss": 1.3437,
      "theoretical_loss": 3.457315374209376,
      "tokens_seen": 2010841088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945837512537613,
      "loss": 1.3889,
      "theoretical_loss": 3.4573060667728024,
      "tokens_seen": 2010906624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039456369107321966,
      "loss": 1.3939,
      "theoretical_loss": 3.4572967597244855,
      "tokens_seen": 2010972160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394543630892678,
      "loss": 1.3829,
      "theoretical_loss": 3.4572874530643967,
      "tokens_seen": 2011037696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945235707121364,
      "loss": 1.4263,
      "theoretical_loss": 3.4572781467925067,
      "tokens_seen": 2011103232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945035105315948,
      "loss": 1.3027,
      "theoretical_loss": 3.4572688409087875,
      "tokens_seen": 2011168768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039448345035105316,
      "loss": 1.3938,
      "theoretical_loss": 3.4572595354132094,
      "tokens_seen": 2011234304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944633901705115,
      "loss": 1.4077,
      "theoretical_loss": 3.4572502303057435,
      "tokens_seen": 2011299840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944433299899699,
      "loss": 1.3055,
      "theoretical_loss": 3.457240925586362,
      "tokens_seen": 2011365376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944232698094283,
      "loss": 1.2892,
      "theoretical_loss": 3.457231621255035,
      "tokens_seen": 2011430912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039440320962888666,
      "loss": 1.3718,
      "theoretical_loss": 3.457222317311734,
      "tokens_seen": 2011496448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394383149448345,
      "loss": 1.3397,
      "theoretical_loss": 3.45721301375643,
      "tokens_seen": 2011561984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943630892678034,
      "loss": 1.4151,
      "theoretical_loss": 3.457203710589095,
      "tokens_seen": 2011627520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039434302908726175,
      "loss": 1.2817,
      "theoretical_loss": 3.457194407809699,
      "tokens_seen": 2011693056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039432296890672017,
      "loss": 1.2498,
      "theoretical_loss": 3.4571851054182146,
      "tokens_seen": 2011758592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039430290872617853,
      "loss": 1.3338,
      "theoretical_loss": 3.4571758034146116,
      "tokens_seen": 2011824128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942828485456369,
      "loss": 1.3969,
      "theoretical_loss": 3.457166501798862,
      "tokens_seen": 2011889664
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039426278836509525,
      "loss": 1.3241,
      "theoretical_loss": 3.4571572005709363,
      "tokens_seen": 2011955200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039424272818455367,
      "loss": 1.377,
      "theoretical_loss": 3.4571478997308063,
      "tokens_seen": 2012020736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039422266800401203,
      "loss": 1.3231,
      "theoretical_loss": 3.4571385992784434,
      "tokens_seen": 2012086272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942026078234704,
      "loss": 1.4517,
      "theoretical_loss": 3.4571292992138183,
      "tokens_seen": 2012151808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039418254764292876,
      "loss": 1.2457,
      "theoretical_loss": 3.457119999536902,
      "tokens_seen": 2012217344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941624874623871,
      "loss": 1.3541,
      "theoretical_loss": 3.4571107002476666,
      "tokens_seen": 2012282880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941424272818456,
      "loss": 1.3976,
      "theoretical_loss": 3.457101401346083,
      "tokens_seen": 2012348416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039412236710130395,
      "loss": 1.3696,
      "theoretical_loss": 3.4570921028321218,
      "tokens_seen": 2012413952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941023069207623,
      "loss": 1.3038,
      "theoretical_loss": 3.4570828047057542,
      "tokens_seen": 2012479488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940822467402207,
      "loss": 1.2962,
      "theoretical_loss": 3.4570735069669523,
      "tokens_seen": 2012545024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940621865596791,
      "loss": 1.4073,
      "theoretical_loss": 3.457064209615687,
      "tokens_seen": 2012610560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039404212637913745,
      "loss": 1.3706,
      "theoretical_loss": 3.4570549126519294,
      "tokens_seen": 2012676096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940220661985958,
      "loss": 1.3941,
      "theoretical_loss": 3.457045616075651,
      "tokens_seen": 2012741632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940020060180542,
      "loss": 1.3407,
      "theoretical_loss": 3.457036319886822,
      "tokens_seen": 2012807168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939819458375126,
      "loss": 1.3197,
      "theoretical_loss": 3.4570270240854146,
      "tokens_seen": 2012872704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039396188565697096,
      "loss": 1.3686,
      "theoretical_loss": 3.4570177286714,
      "tokens_seen": 2012938240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939418254764293,
      "loss": 1.3787,
      "theoretical_loss": 3.4570084336447495,
      "tokens_seen": 2013003776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939217652958877,
      "loss": 1.2409,
      "theoretical_loss": 3.4569991390054335,
      "tokens_seen": 2013069312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039390170511534604,
      "loss": 1.2971,
      "theoretical_loss": 3.4569898447534246,
      "tokens_seen": 2013134848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039388164493480446,
      "loss": 1.3715,
      "theoretical_loss": 3.4569805508886935,
      "tokens_seen": 2013200384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003938615847542628,
      "loss": 1.3844,
      "theoretical_loss": 3.456971257411211,
      "tokens_seen": 2013265920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003938415245737212,
      "loss": 1.3091,
      "theoretical_loss": 3.456961964320948,
      "tokens_seen": 2013331456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039382146439317955,
      "loss": 1.3863,
      "theoretical_loss": 3.456952671617877,
      "tokens_seen": 2013396992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039380140421263796,
      "loss": 1.4279,
      "theoretical_loss": 3.4569433793019693,
      "tokens_seen": 2013462528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937813440320963,
      "loss": 1.2822,
      "theoretical_loss": 3.4569340873731944,
      "tokens_seen": 2013528064
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937612838515547,
      "loss": 1.2731,
      "theoretical_loss": 3.4569247958315255,
      "tokens_seen": 2013593600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039374122367101305,
      "loss": 1.385,
      "theoretical_loss": 3.4569155046769326,
      "tokens_seen": 2013659136
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 734578,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.427262544631958,
      "objective/train/theoretical_loss": 3.456906213909388,
      "objective/train/tokens_used": 384183776,
      "theoretical_loss": 3.456906213909388,
      "tokens_seen": 2013724672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937211634904714,
      "loss": 1.4218,
      "theoretical_loss": 3.456906213909388,
      "tokens_seen": 2013724672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039370110330992983,
      "loss": 1.4594,
      "theoretical_loss": 3.4568969235288627,
      "tokens_seen": 2013790208
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936810431293882,
      "loss": 1.2872,
      "theoretical_loss": 3.456887633535327,
      "tokens_seen": 2013855744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039366098294884655,
      "loss": 1.2838,
      "theoretical_loss": 3.456878343928753,
      "tokens_seen": 2013921280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936409227683049,
      "loss": 1.3751,
      "theoretical_loss": 3.456869054709112,
      "tokens_seen": 2013986816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039362086258776333,
      "loss": 1.3387,
      "theoretical_loss": 3.4568597658763753,
      "tokens_seen": 2014052352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936008024072217,
      "loss": 1.3155,
      "theoretical_loss": 3.4568504774305144,
      "tokens_seen": 2014117888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039358074222668006,
      "loss": 1.3978,
      "theoretical_loss": 3.4568411893715,
      "tokens_seen": 2014183424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935606820461384,
      "loss": 1.4352,
      "theoretical_loss": 3.456831901699304,
      "tokens_seen": 2014248960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935406218655968,
      "loss": 1.3223,
      "theoretical_loss": 3.4568226144138974,
      "tokens_seen": 2014314496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935205616850552,
      "loss": 1.2445,
      "theoretical_loss": 3.4568133275152513,
      "tokens_seen": 2014380032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039350050150451356,
      "loss": 1.4241,
      "theoretical_loss": 3.456804041003337,
      "tokens_seen": 2014445568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934804413239719,
      "loss": 1.286,
      "theoretical_loss": 3.4567947548781266,
      "tokens_seen": 2014511104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934603811434303,
      "loss": 1.4323,
      "theoretical_loss": 3.4567854691395903,
      "tokens_seen": 2014576640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934403209628887,
      "loss": 1.3695,
      "theoretical_loss": 3.4567761837877002,
      "tokens_seen": 2014642176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039342026078234706,
      "loss": 1.3333,
      "theoretical_loss": 3.456766898822428,
      "tokens_seen": 2014707712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934002006018054,
      "loss": 1.3333,
      "theoretical_loss": 3.456757614243744,
      "tokens_seen": 2014773248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003933801404212638,
      "loss": 1.2941,
      "theoretical_loss": 3.45674833005162,
      "tokens_seen": 2014838784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039336008024072215,
      "loss": 1.3604,
      "theoretical_loss": 3.4567390462460272,
      "tokens_seen": 2014904320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039334002006018057,
      "loss": 1.2572,
      "theoretical_loss": 3.456729762826937,
      "tokens_seen": 2014969856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039331995987963893,
      "loss": 1.2464,
      "theoretical_loss": 3.456720479794321,
      "tokens_seen": 2015035392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932998996990973,
      "loss": 1.3362,
      "theoretical_loss": 3.45671119714815,
      "tokens_seen": 2015100928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039327983951855565,
      "loss": 1.4009,
      "theoretical_loss": 3.456701914888396,
      "tokens_seen": 2015166464
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039325977933801407,
      "loss": 1.3729,
      "theoretical_loss": 3.45669263301503,
      "tokens_seen": 2015232000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039323971915747243,
      "loss": 1.2614,
      "theoretical_loss": 3.4566833515280235,
      "tokens_seen": 2015297536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932196589769308,
      "loss": 1.3329,
      "theoretical_loss": 3.456674070427347,
      "tokens_seen": 2015363072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039319959879638916,
      "loss": 1.3501,
      "theoretical_loss": 3.456664789712973,
      "tokens_seen": 2015428608
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931795386158475,
      "loss": 1.4081,
      "theoretical_loss": 3.4566555093848725,
      "tokens_seen": 2015494144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039315947843530593,
      "loss": 1.2733,
      "theoretical_loss": 3.456646229443017,
      "tokens_seen": 2015559680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931394182547643,
      "loss": 1.3006,
      "theoretical_loss": 3.4566369498873777,
      "tokens_seen": 2015625216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039311935807422266,
      "loss": 1.368,
      "theoretical_loss": 3.4566276707179258,
      "tokens_seen": 2015690752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000393099297893681,
      "loss": 1.3525,
      "theoretical_loss": 3.456618391934633,
      "tokens_seen": 2015756288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039307923771313944,
      "loss": 1.295,
      "theoretical_loss": 3.45660911353747,
      "tokens_seen": 2015821824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003930591775325978,
      "loss": 1.4187,
      "theoretical_loss": 3.4565998355264087,
      "tokens_seen": 2015887360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039303911735205616,
      "loss": 1.3145,
      "theoretical_loss": 3.456590557901421,
      "tokens_seen": 2015952896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003930190571715145,
      "loss": 1.4118,
      "theoretical_loss": 3.456581280662477,
      "tokens_seen": 2016018432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929989969909729,
      "loss": 1.3591,
      "theoretical_loss": 3.4565720038095495,
      "tokens_seen": 2016083968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929789368104313,
      "loss": 1.3329,
      "theoretical_loss": 3.456562727342609,
      "tokens_seen": 2016149504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039295887662988967,
      "loss": 1.3288,
      "theoretical_loss": 3.4565534512616267,
      "tokens_seen": 2016215040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039293881644934803,
      "loss": 1.3979,
      "theoretical_loss": 3.4565441755665747,
      "tokens_seen": 2016280576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929187562688064,
      "loss": 1.3076,
      "theoretical_loss": 3.4565349002574246,
      "tokens_seen": 2016346112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003928986960882648,
      "loss": 1.2742,
      "theoretical_loss": 3.456525625334147,
      "tokens_seen": 2016411648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039287863590772317,
      "loss": 1.3801,
      "theoretical_loss": 3.456516350796713,
      "tokens_seen": 2016477184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039285857572718153,
      "loss": 1.3603,
      "theoretical_loss": 3.4565070766450954,
      "tokens_seen": 2016542720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003928385155466399,
      "loss": 1.3577,
      "theoretical_loss": 3.4564978028792646,
      "tokens_seen": 2016608256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039281845536609825,
      "loss": 1.3676,
      "theoretical_loss": 3.4564885294991923,
      "tokens_seen": 2016673792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039279839518555667,
      "loss": 1.3992,
      "theoretical_loss": 3.4564792565048497,
      "tokens_seen": 2016739328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039277833500501503,
      "loss": 1.4252,
      "theoretical_loss": 3.4564699838962083,
      "tokens_seen": 2016804864
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003927582748244734,
      "loss": 1.35,
      "theoretical_loss": 3.45646071167324,
      "tokens_seen": 2016870400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039273821464393176,
      "loss": 1.4746,
      "theoretical_loss": 3.4564514398359156,
      "tokens_seen": 2016935936
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 735776,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.506068468093872,
      "objective/train/theoretical_loss": 3.456442168384207,
      "objective/train/tokens_used": 387460576,
      "theoretical_loss": 3.456442168384207,
      "tokens_seen": 2017001472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003927181544633902,
      "loss": 1.4847,
      "theoretical_loss": 3.456442168384207,
      "tokens_seen": 2017001472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039269809428284854,
      "loss": 1.2308,
      "theoretical_loss": 3.456432897318085,
      "tokens_seen": 2017067008
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926780341023069,
      "loss": 1.3872,
      "theoretical_loss": 3.456423626637522,
      "tokens_seen": 2017132544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039265797392176526,
      "loss": 1.4539,
      "theoretical_loss": 3.4564143563424885,
      "tokens_seen": 2017198080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039263791374122373,
      "loss": 1.3045,
      "theoretical_loss": 3.456405086432956,
      "tokens_seen": 2017263616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926178535606821,
      "loss": 1.3124,
      "theoretical_loss": 3.456395816908897,
      "tokens_seen": 2017329152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039259779338014046,
      "loss": 1.389,
      "theoretical_loss": 3.4563865477702818,
      "tokens_seen": 2017394688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925777331995988,
      "loss": 1.4732,
      "theoretical_loss": 3.4563772790170826,
      "tokens_seen": 2017460224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925576730190572,
      "loss": 1.3558,
      "theoretical_loss": 3.4563680106492702,
      "tokens_seen": 2017525760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925376128385156,
      "loss": 1.2866,
      "theoretical_loss": 3.4563587426668168,
      "tokens_seen": 2017591296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039251755265797396,
      "loss": 1.2783,
      "theoretical_loss": 3.456349475069693,
      "tokens_seen": 2017656832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924974924774323,
      "loss": 1.3588,
      "theoretical_loss": 3.4563402078578704,
      "tokens_seen": 2017722368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924774322968907,
      "loss": 1.3484,
      "theoretical_loss": 3.4563309410313217,
      "tokens_seen": 2017787904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924573721163491,
      "loss": 1.3495,
      "theoretical_loss": 3.4563216745900167,
      "tokens_seen": 2017853440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039243731193580746,
      "loss": 1.2857,
      "theoretical_loss": 3.456312408533928,
      "tokens_seen": 2017918976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924172517552658,
      "loss": 1.4029,
      "theoretical_loss": 3.456303142863027,
      "tokens_seen": 2017984512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923971915747242,
      "loss": 1.2003,
      "theoretical_loss": 3.4562938775772842,
      "tokens_seen": 2018050048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039237713139418255,
      "loss": 1.4004,
      "theoretical_loss": 3.456284612676672,
      "tokens_seen": 2018115584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039235707121364097,
      "loss": 1.2417,
      "theoretical_loss": 3.456275348161162,
      "tokens_seen": 2018181120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039233701103309933,
      "loss": 1.4594,
      "theoretical_loss": 3.456266084030725,
      "tokens_seen": 2018246656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923169508525577,
      "loss": 1.2536,
      "theoretical_loss": 3.4562568202853328,
      "tokens_seen": 2018312192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039229689067201605,
      "loss": 1.4737,
      "theoretical_loss": 3.456247556924957,
      "tokens_seen": 2018377728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039227683049147447,
      "loss": 1.3414,
      "theoretical_loss": 3.456238293949569,
      "tokens_seen": 2018443264
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039225677031093283,
      "loss": 1.3431,
      "theoretical_loss": 3.4562290313591406,
      "tokens_seen": 2018508800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003922367101303912,
      "loss": 1.4163,
      "theoretical_loss": 3.4562197691536425,
      "tokens_seen": 2018574336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039221664994984956,
      "loss": 1.3747,
      "theoretical_loss": 3.456210507333047,
      "tokens_seen": 2018639872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921965897693079,
      "loss": 1.4036,
      "theoretical_loss": 3.4562012458973257,
      "tokens_seen": 2018705408
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039217652958876633,
      "loss": 1.3448,
      "theoretical_loss": 3.4561919848464493,
      "tokens_seen": 2018770944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921564694082247,
      "loss": 1.2858,
      "theoretical_loss": 3.4561827241803895,
      "tokens_seen": 2018836480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039213640922768306,
      "loss": 1.3955,
      "theoretical_loss": 3.456173463899119,
      "tokens_seen": 2018902016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921163490471414,
      "loss": 1.3994,
      "theoretical_loss": 3.4561642040026075,
      "tokens_seen": 2018967552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039209628886659984,
      "loss": 1.3622,
      "theoretical_loss": 3.4561549444908275,
      "tokens_seen": 2019033088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003920762286860582,
      "loss": 1.3141,
      "theoretical_loss": 3.456145685363751,
      "tokens_seen": 2019098624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039205616850551656,
      "loss": 1.2607,
      "theoretical_loss": 3.456136426621349,
      "tokens_seen": 2019164160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003920361083249749,
      "loss": 1.3596,
      "theoretical_loss": 3.456127168263593,
      "tokens_seen": 2019229696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003920160481444333,
      "loss": 1.4247,
      "theoretical_loss": 3.4561179102904536,
      "tokens_seen": 2019295232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003919959879638917,
      "loss": 1.331,
      "theoretical_loss": 3.4561086527019045,
      "tokens_seen": 2019360768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039197592778335007,
      "loss": 1.4164,
      "theoretical_loss": 3.4560993954979153,
      "tokens_seen": 2019426304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039195586760280843,
      "loss": 1.3072,
      "theoretical_loss": 3.4560901386784586,
      "tokens_seen": 2019491840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003919358074222668,
      "loss": 1.4174,
      "theoretical_loss": 3.456080882243506,
      "tokens_seen": 2019557376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003919157472417252,
      "loss": 1.2632,
      "theoretical_loss": 3.456071626193028,
      "tokens_seen": 2019622912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039189568706118357,
      "loss": 1.4112,
      "theoretical_loss": 3.4560623705269973,
      "tokens_seen": 2019688448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039187562688064193,
      "loss": 1.3727,
      "theoretical_loss": 3.4560531152453846,
      "tokens_seen": 2019753984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003918555667001003,
      "loss": 1.3334,
      "theoretical_loss": 3.4560438603481622,
      "tokens_seen": 2019819520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039183550651955866,
      "loss": 1.3373,
      "theoretical_loss": 3.4560346058353018,
      "tokens_seen": 2019885056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039181544633901707,
      "loss": 1.3559,
      "theoretical_loss": 3.4560253517067734,
      "tokens_seen": 2019950592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039179538615847543,
      "loss": 1.3793,
      "theoretical_loss": 3.4560160979625505,
      "tokens_seen": 2020016128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003917753259779338,
      "loss": 1.3522,
      "theoretical_loss": 3.4560068446026033,
      "tokens_seen": 2020081664
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039175526579739216,
      "loss": 1.3314,
      "theoretical_loss": 3.455997591626905,
      "tokens_seen": 2020147200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003917352056168506,
      "loss": 1.3064,
      "theoretical_loss": 3.455988339035425,
      "tokens_seen": 2020212736
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 736878,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3108620643615723,
      "objective/train/theoretical_loss": 3.455979086828136,
      "objective/train/tokens_used": 390737376,
      "theoretical_loss": 3.455979086828136,
      "tokens_seen": 2020278272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039171514543630894,
      "loss": 1.3245,
      "theoretical_loss": 3.455979086828136,
      "tokens_seen": 2020278272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916950852557673,
      "loss": 1.2786,
      "theoretical_loss": 3.45596983500501,
      "tokens_seen": 2020343808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039167502507522566,
      "loss": 1.2895,
      "theoretical_loss": 3.455960583566018,
      "tokens_seen": 2020409344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916549648946841,
      "loss": 1.2725,
      "theoretical_loss": 3.4559513325111317,
      "tokens_seen": 2020474880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039163490471414244,
      "loss": 1.4011,
      "theoretical_loss": 3.4559420818403233,
      "tokens_seen": 2020540416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916148445336008,
      "loss": 1.3641,
      "theoretical_loss": 3.455932831553563,
      "tokens_seen": 2020605952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039159478435305916,
      "loss": 1.3793,
      "theoretical_loss": 3.455923581650824,
      "tokens_seen": 2020671488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915747241725175,
      "loss": 1.4393,
      "theoretical_loss": 3.4559143321320764,
      "tokens_seen": 2020737024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039155466399197594,
      "loss": 1.3359,
      "theoretical_loss": 3.455905082997293,
      "tokens_seen": 2020802560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915346038114343,
      "loss": 1.3857,
      "theoretical_loss": 3.4558958342464448,
      "tokens_seen": 2020868096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039151454363089267,
      "loss": 1.4345,
      "theoretical_loss": 3.4558865858795036,
      "tokens_seen": 2020933632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039149448345035103,
      "loss": 1.3459,
      "theoretical_loss": 3.4558773378964407,
      "tokens_seen": 2020999168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039147442326980945,
      "loss": 1.4263,
      "theoretical_loss": 3.4558680902972285,
      "tokens_seen": 2021064704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003914543630892678,
      "loss": 1.358,
      "theoretical_loss": 3.455858843081838,
      "tokens_seen": 2021130240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039143430290872617,
      "loss": 1.3971,
      "theoretical_loss": 3.455849596250241,
      "tokens_seen": 2021195776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039141424272818453,
      "loss": 1.3009,
      "theoretical_loss": 3.455840349802409,
      "tokens_seen": 2021261312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913941825476429,
      "loss": 1.3189,
      "theoretical_loss": 3.455831103738314,
      "tokens_seen": 2021326848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913741223671013,
      "loss": 1.3372,
      "theoretical_loss": 3.4558218580579263,
      "tokens_seen": 2021392384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913540621865597,
      "loss": 1.3047,
      "theoretical_loss": 3.4558126127612194,
      "tokens_seen": 2021457920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039133400200601804,
      "loss": 1.4054,
      "theoretical_loss": 3.4558033678481643,
      "tokens_seen": 2021523456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913139418254764,
      "loss": 1.3177,
      "theoretical_loss": 3.455794123318732,
      "tokens_seen": 2021588992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912938816449348,
      "loss": 1.3129,
      "theoretical_loss": 3.455784879172895,
      "tokens_seen": 2021654528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912738214643932,
      "loss": 1.3085,
      "theoretical_loss": 3.4557756354106246,
      "tokens_seen": 2021720064
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039125376128385154,
      "loss": 1.3978,
      "theoretical_loss": 3.455766392031892,
      "tokens_seen": 2021785600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912337011033099,
      "loss": 1.3599,
      "theoretical_loss": 3.4557571490366694,
      "tokens_seen": 2021851136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039121364092276826,
      "loss": 1.3401,
      "theoretical_loss": 3.4557479064249286,
      "tokens_seen": 2021916672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911935807422267,
      "loss": 1.4223,
      "theoretical_loss": 3.45573866419664,
      "tokens_seen": 2021982208
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039117352056168504,
      "loss": 1.4132,
      "theoretical_loss": 3.4557294223517774,
      "tokens_seen": 2022047744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911534603811434,
      "loss": 1.2706,
      "theoretical_loss": 3.455720180890311,
      "tokens_seen": 2022113280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911334002006018,
      "loss": 1.4531,
      "theoretical_loss": 3.455710939812213,
      "tokens_seen": 2022178816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039111334002006024,
      "loss": 1.4664,
      "theoretical_loss": 3.455701699117454,
      "tokens_seen": 2022244352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910932798395186,
      "loss": 1.3775,
      "theoretical_loss": 3.4556924588060074,
      "tokens_seen": 2022309888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039107321965897696,
      "loss": 1.3495,
      "theoretical_loss": 3.4556832188778435,
      "tokens_seen": 2022375424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910531594784353,
      "loss": 1.4289,
      "theoretical_loss": 3.4556739793329347,
      "tokens_seen": 2022440960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910330992978937,
      "loss": 1.3323,
      "theoretical_loss": 3.455664740171253,
      "tokens_seen": 2022506496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910130391173521,
      "loss": 1.3608,
      "theoretical_loss": 3.4556555013927683,
      "tokens_seen": 2022572032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039099297893681047,
      "loss": 1.3927,
      "theoretical_loss": 3.4556462629974547,
      "tokens_seen": 2022637568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039097291875626883,
      "loss": 1.3557,
      "theoretical_loss": 3.455637024985282,
      "tokens_seen": 2022703104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003909528585757272,
      "loss": 1.2985,
      "theoretical_loss": 3.455627787356223,
      "tokens_seen": 2022768640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003909327983951856,
      "loss": 1.3314,
      "theoretical_loss": 3.455618550110249,
      "tokens_seen": 2022834176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039091273821464397,
      "loss": 1.4317,
      "theoretical_loss": 3.455609313247332,
      "tokens_seen": 2022899712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039089267803410233,
      "loss": 1.3545,
      "theoretical_loss": 3.4556000767674435,
      "tokens_seen": 2022965248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003908726178535607,
      "loss": 1.3999,
      "theoretical_loss": 3.4555908406705544,
      "tokens_seen": 2023030784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039085255767301906,
      "loss": 1.3433,
      "theoretical_loss": 3.455581604956638,
      "tokens_seen": 2023096320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039083249749247747,
      "loss": 1.2721,
      "theoretical_loss": 3.455572369625665,
      "tokens_seen": 2023161856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039081243731193583,
      "loss": 1.3659,
      "theoretical_loss": 3.455563134677607,
      "tokens_seen": 2023227392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907923771313942,
      "loss": 1.4036,
      "theoretical_loss": 3.4555539001124362,
      "tokens_seen": 2023292928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039077231695085256,
      "loss": 1.241,
      "theoretical_loss": 3.455544665930124,
      "tokens_seen": 2023358464
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390752256770311,
      "loss": 1.3087,
      "theoretical_loss": 3.4555354321306426,
      "tokens_seen": 2023424000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039073219658976934,
      "loss": 1.1727,
      "theoretical_loss": 3.4555261987139634,
      "tokens_seen": 2023489536
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 738179,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.421309232711792,
      "objective/train/theoretical_loss": 3.455516965680058,
      "objective/train/tokens_used": 394014176,
      "theoretical_loss": 3.455516965680058,
      "tokens_seen": 2023555072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907121364092277,
      "loss": 1.3643,
      "theoretical_loss": 3.455516965680058,
      "tokens_seen": 2023555072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039069207622868606,
      "loss": 1.3152,
      "theoretical_loss": 3.4555077330288984,
      "tokens_seen": 2023620608
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906720160481445,
      "loss": 1.3485,
      "theoretical_loss": 3.455498500760456,
      "tokens_seen": 2023686144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039065195586760284,
      "loss": 1.3794,
      "theoretical_loss": 3.455489268874703,
      "tokens_seen": 2023751680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906318956870612,
      "loss": 1.3256,
      "theoretical_loss": 3.455480037371611,
      "tokens_seen": 2023817216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039061183550651957,
      "loss": 1.3138,
      "theoretical_loss": 3.455470806251151,
      "tokens_seen": 2023882752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039059177532597793,
      "loss": 1.3413,
      "theoretical_loss": 3.4554615755132962,
      "tokens_seen": 2023948288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039057171514543634,
      "loss": 1.364,
      "theoretical_loss": 3.455452345158017,
      "tokens_seen": 2024013824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003905516549648947,
      "loss": 1.4109,
      "theoretical_loss": 3.455443115185286,
      "tokens_seen": 2024079360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039053159478435307,
      "loss": 1.3155,
      "theoretical_loss": 3.455433885595075,
      "tokens_seen": 2024144896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039051153460381143,
      "loss": 1.3693,
      "theoretical_loss": 3.455424656387355,
      "tokens_seen": 2024210432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039049147442326985,
      "loss": 1.3418,
      "theoretical_loss": 3.455415427562098,
      "tokens_seen": 2024275968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003904714142427282,
      "loss": 1.3078,
      "theoretical_loss": 3.4554061991192766,
      "tokens_seen": 2024341504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039045135406218657,
      "loss": 1.3368,
      "theoretical_loss": 3.4553969710588612,
      "tokens_seen": 2024407040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039043129388164493,
      "loss": 1.2256,
      "theoretical_loss": 3.4553877433808253,
      "tokens_seen": 2024472576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003904112337011033,
      "loss": 1.2956,
      "theoretical_loss": 3.4553785160851387,
      "tokens_seen": 2024538112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903911735205617,
      "loss": 1.3184,
      "theoretical_loss": 3.4553692891717747,
      "tokens_seen": 2024603648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903711133400201,
      "loss": 1.3839,
      "theoretical_loss": 3.455360062640705,
      "tokens_seen": 2024669184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039035105315947844,
      "loss": 1.2909,
      "theoretical_loss": 3.4553508364919,
      "tokens_seen": 2024734720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903309929789368,
      "loss": 1.5578,
      "theoretical_loss": 3.4553416107253327,
      "tokens_seen": 2024800256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903109327983952,
      "loss": 1.3584,
      "theoretical_loss": 3.455332385340975,
      "tokens_seen": 2024865792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902908726178536,
      "loss": 1.296,
      "theoretical_loss": 3.455323160338798,
      "tokens_seen": 2024931328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039027081243731194,
      "loss": 1.4688,
      "theoretical_loss": 3.455313935718774,
      "tokens_seen": 2024996864
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902507522567703,
      "loss": 1.3915,
      "theoretical_loss": 3.4553047114808746,
      "tokens_seen": 2025062400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039023069207622866,
      "loss": 1.3522,
      "theoretical_loss": 3.455295487625072,
      "tokens_seen": 2025127936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902106318956871,
      "loss": 1.4055,
      "theoretical_loss": 3.4552862641513373,
      "tokens_seen": 2025193472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039019057171514544,
      "loss": 1.3314,
      "theoretical_loss": 3.4552770410596425,
      "tokens_seen": 2025259008
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901705115346038,
      "loss": 1.3877,
      "theoretical_loss": 3.4552678183499594,
      "tokens_seen": 2025324544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039015045135406217,
      "loss": 1.3633,
      "theoretical_loss": 3.4552585960222606,
      "tokens_seen": 2025390080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901303911735206,
      "loss": 1.3911,
      "theoretical_loss": 3.455249374076517,
      "tokens_seen": 2025455616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039011033099297895,
      "loss": 1.3176,
      "theoretical_loss": 3.455240152512701,
      "tokens_seen": 2025521152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003900902708124373,
      "loss": 1.3319,
      "theoretical_loss": 3.455230931330784,
      "tokens_seen": 2025586688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039007021063189567,
      "loss": 1.3708,
      "theoretical_loss": 3.4552217105307377,
      "tokens_seen": 2025652224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039005015045135403,
      "loss": 1.3965,
      "theoretical_loss": 3.4552124901125345,
      "tokens_seen": 2025717760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039003009027081245,
      "loss": 1.3143,
      "theoretical_loss": 3.4552032700761455,
      "tokens_seen": 2025783296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003900100300902708,
      "loss": 1.3551,
      "theoretical_loss": 3.455194050421544,
      "tokens_seen": 2025848832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003899899699097292,
      "loss": 1.2757,
      "theoretical_loss": 3.4551848311487,
      "tokens_seen": 2025914368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038996990972918754,
      "loss": 1.4894,
      "theoretical_loss": 3.4551756122575865,
      "tokens_seen": 2025979904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038994984954864595,
      "loss": 1.3522,
      "theoretical_loss": 3.4551663937481747,
      "tokens_seen": 2026045440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003899297893681043,
      "loss": 1.304,
      "theoretical_loss": 3.455157175620437,
      "tokens_seen": 2026110976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003899097291875627,
      "loss": 1.2434,
      "theoretical_loss": 3.455147957874345,
      "tokens_seen": 2026176512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038988966900702104,
      "loss": 1.3624,
      "theoretical_loss": 3.45513874050987,
      "tokens_seen": 2026242048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898696088264794,
      "loss": 1.4139,
      "theoretical_loss": 3.4551295235269857,
      "tokens_seen": 2026307584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898495486459378,
      "loss": 1.4757,
      "theoretical_loss": 3.4551203069256617,
      "tokens_seen": 2026373120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898294884653962,
      "loss": 1.3204,
      "theoretical_loss": 3.4551110907058713,
      "tokens_seen": 2026438656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038980942828485454,
      "loss": 1.3421,
      "theoretical_loss": 3.455101874867586,
      "tokens_seen": 2026504192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897893681043129,
      "loss": 1.3783,
      "theoretical_loss": 3.455092659410777,
      "tokens_seen": 2026569728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897693079237713,
      "loss": 1.4032,
      "theoretical_loss": 3.4550834443354175,
      "tokens_seen": 2026635264
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897492477432297,
      "loss": 1.2936,
      "theoretical_loss": 3.4550742296414785,
      "tokens_seen": 2026700800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038972918756268805,
      "loss": 1.2721,
      "theoretical_loss": 3.455065015328932,
      "tokens_seen": 2026766336
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 739363,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3435156345367432,
      "objective/train/theoretical_loss": 3.4550558013977497,
      "objective/train/tokens_used": 397290976,
      "theoretical_loss": 3.4550558013977497,
      "tokens_seen": 2026831872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897091273821464,
      "loss": 1.3073,
      "theoretical_loss": 3.4550558013977497,
      "tokens_seen": 2026831872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038968906720160477,
      "loss": 1.4158,
      "theoretical_loss": 3.455046587847904,
      "tokens_seen": 2026897408
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003896690070210632,
      "loss": 1.393,
      "theoretical_loss": 3.455037374679366,
      "tokens_seen": 2026962944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038964894684052155,
      "loss": 1.3133,
      "theoretical_loss": 3.455028161892108,
      "tokens_seen": 2027028480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038962888665997997,
      "loss": 1.4214,
      "theoretical_loss": 3.455018949486103,
      "tokens_seen": 2027094016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038960882647943833,
      "loss": 1.2748,
      "theoretical_loss": 3.455009737461321,
      "tokens_seen": 2027159552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038958876629889674,
      "loss": 1.2643,
      "theoretical_loss": 3.455000525817735,
      "tokens_seen": 2027225088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003895687061183551,
      "loss": 1.3237,
      "theoretical_loss": 3.4549913145553166,
      "tokens_seen": 2027290624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038954864593781347,
      "loss": 1.2969,
      "theoretical_loss": 3.4549821036740385,
      "tokens_seen": 2027356160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038952858575727183,
      "loss": 1.3471,
      "theoretical_loss": 3.454972893173871,
      "tokens_seen": 2027421696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038950852557673025,
      "loss": 1.3409,
      "theoretical_loss": 3.454963683054787,
      "tokens_seen": 2027487232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894884653961886,
      "loss": 1.3143,
      "theoretical_loss": 3.4549544733167585,
      "tokens_seen": 2027552768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038946840521564697,
      "loss": 1.3953,
      "theoretical_loss": 3.4549452639597575,
      "tokens_seen": 2027618304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038944834503510533,
      "loss": 1.3597,
      "theoretical_loss": 3.4549360549837553,
      "tokens_seen": 2027683840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894282848545637,
      "loss": 1.3207,
      "theoretical_loss": 3.4549268463887244,
      "tokens_seen": 2027749376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894082246740221,
      "loss": 1.3519,
      "theoretical_loss": 3.4549176381746367,
      "tokens_seen": 2027814912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003893881644934805,
      "loss": 1.4071,
      "theoretical_loss": 3.4549084303414634,
      "tokens_seen": 2027880448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038936810431293884,
      "loss": 1.3912,
      "theoretical_loss": 3.4548992228891775,
      "tokens_seen": 2027945984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003893480441323972,
      "loss": 1.3927,
      "theoretical_loss": 3.45489001581775,
      "tokens_seen": 2028011520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003893279839518556,
      "loss": 1.4808,
      "theoretical_loss": 3.4548808091271535,
      "tokens_seen": 2028077056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389307923771314,
      "loss": 1.4265,
      "theoretical_loss": 3.4548716028173594,
      "tokens_seen": 2028142592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038928786359077234,
      "loss": 1.4355,
      "theoretical_loss": 3.4548623968883403,
      "tokens_seen": 2028208128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892678034102307,
      "loss": 1.3491,
      "theoretical_loss": 3.4548531913400677,
      "tokens_seen": 2028273664
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038924774322968906,
      "loss": 1.3962,
      "theoretical_loss": 3.4548439861725138,
      "tokens_seen": 2028339200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892276830491475,
      "loss": 1.3312,
      "theoretical_loss": 3.45483478138565,
      "tokens_seen": 2028404736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038920762286860584,
      "loss": 1.2968,
      "theoretical_loss": 3.4548255769794487,
      "tokens_seen": 2028470272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891875626880642,
      "loss": 1.4201,
      "theoretical_loss": 3.454816372953882,
      "tokens_seen": 2028535808
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038916750250752257,
      "loss": 1.2663,
      "theoretical_loss": 3.4548071693089217,
      "tokens_seen": 2028601344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389147442326981,
      "loss": 1.4309,
      "theoretical_loss": 3.4547979660445396,
      "tokens_seen": 2028666880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038912738214643935,
      "loss": 1.3663,
      "theoretical_loss": 3.4547887631607077,
      "tokens_seen": 2028732416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891073219658977,
      "loss": 1.4298,
      "theoretical_loss": 3.4547795606573986,
      "tokens_seen": 2028797952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038908726178535607,
      "loss": 1.3635,
      "theoretical_loss": 3.4547703585345833,
      "tokens_seen": 2028863488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038906720160481443,
      "loss": 1.3353,
      "theoretical_loss": 3.4547611567922347,
      "tokens_seen": 2028929024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038904714142427285,
      "loss": 1.3469,
      "theoretical_loss": 3.4547519554303237,
      "tokens_seen": 2028994560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890270812437312,
      "loss": 1.3301,
      "theoretical_loss": 3.454742754448823,
      "tokens_seen": 2029060096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890070210631896,
      "loss": 1.4622,
      "theoretical_loss": 3.4547335538477046,
      "tokens_seen": 2029125632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038898696088264794,
      "loss": 1.4134,
      "theoretical_loss": 3.4547243536269407,
      "tokens_seen": 2029191168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038896690070210635,
      "loss": 1.4085,
      "theoretical_loss": 3.454715153786503,
      "tokens_seen": 2029256704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003889468405215647,
      "loss": 1.3614,
      "theoretical_loss": 3.454705954326363,
      "tokens_seen": 2029322240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003889267803410231,
      "loss": 1.3452,
      "theoretical_loss": 3.454696755246493,
      "tokens_seen": 2029387776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038890672016048144,
      "loss": 1.3949,
      "theoretical_loss": 3.4546875565468658,
      "tokens_seen": 2029453312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888866599799398,
      "loss": 1.3621,
      "theoretical_loss": 3.4546783582274525,
      "tokens_seen": 2029518848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888665997993982,
      "loss": 1.4129,
      "theoretical_loss": 3.4546691602882253,
      "tokens_seen": 2029584384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888465396188566,
      "loss": 1.2995,
      "theoretical_loss": 3.4546599627291563,
      "tokens_seen": 2029649920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038882647943831494,
      "loss": 1.3142,
      "theoretical_loss": 3.454650765550218,
      "tokens_seen": 2029715456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888064192577733,
      "loss": 1.3571,
      "theoretical_loss": 3.4546415687513807,
      "tokens_seen": 2029780992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003887863590772317,
      "loss": 1.2966,
      "theoretical_loss": 3.4546323723326187,
      "tokens_seen": 2029846528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003887662988966901,
      "loss": 1.383,
      "theoretical_loss": 3.454623176293903,
      "tokens_seen": 2029912064
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038874623871614845,
      "loss": 1.3505,
      "theoretical_loss": 3.4546139806352048,
      "tokens_seen": 2029977600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003887261785356068,
      "loss": 1.4078,
      "theoretical_loss": 3.4546047853564974,
      "tokens_seen": 2030043136
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 740596,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2953369617462158,
      "objective/train/theoretical_loss": 3.4545955904577523,
      "objective/train/tokens_used": 400567776,
      "theoretical_loss": 3.4545955904577523,
      "tokens_seen": 2030108672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038870611835506517,
      "loss": 1.3055,
      "theoretical_loss": 3.4545955904577523,
      "tokens_seen": 2030108672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886860581745236,
      "loss": 1.3938,
      "theoretical_loss": 3.4545863959389416,
      "tokens_seen": 2030174208
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038866599799398195,
      "loss": 1.1979,
      "theoretical_loss": 3.4545772018000376,
      "tokens_seen": 2030239744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886459378134403,
      "loss": 1.3141,
      "theoretical_loss": 3.4545680080410115,
      "tokens_seen": 2030305280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886258776328987,
      "loss": 1.4452,
      "theoretical_loss": 3.4545588146618362,
      "tokens_seen": 2030370816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886058174523571,
      "loss": 1.4127,
      "theoretical_loss": 3.4545496216624834,
      "tokens_seen": 2030436352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038858575727181545,
      "loss": 1.3573,
      "theoretical_loss": 3.454540429042925,
      "tokens_seen": 2030501888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885656970912738,
      "loss": 1.2665,
      "theoretical_loss": 3.454531236803134,
      "tokens_seen": 2030567424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885456369107322,
      "loss": 1.2752,
      "theoretical_loss": 3.4545220449430807,
      "tokens_seen": 2030632960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885255767301906,
      "loss": 1.2991,
      "theoretical_loss": 3.4545128534627385,
      "tokens_seen": 2030698496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038850551654964896,
      "loss": 1.3686,
      "theoretical_loss": 3.4545036623620793,
      "tokens_seen": 2030764032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003884854563691073,
      "loss": 1.2501,
      "theoretical_loss": 3.454494471641075,
      "tokens_seen": 2030829568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003884653961885657,
      "loss": 1.4353,
      "theoretical_loss": 3.4544852812996973,
      "tokens_seen": 2030895104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038844533600802404,
      "loss": 1.2883,
      "theoretical_loss": 3.454476091337919,
      "tokens_seen": 2030960640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038842527582748246,
      "loss": 1.3937,
      "theoretical_loss": 3.4544669017557115,
      "tokens_seen": 2031026176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003884052156469408,
      "loss": 1.3496,
      "theoretical_loss": 3.4544577125530473,
      "tokens_seen": 2031091712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883851554663992,
      "loss": 1.3512,
      "theoretical_loss": 3.4544485237298987,
      "tokens_seen": 2031157248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038836509528585755,
      "loss": 1.2724,
      "theoretical_loss": 3.454439335286237,
      "tokens_seen": 2031222784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038834503510531596,
      "loss": 1.3478,
      "theoretical_loss": 3.454430147222035,
      "tokens_seen": 2031288320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883249749247743,
      "loss": 1.349,
      "theoretical_loss": 3.4544209595372637,
      "tokens_seen": 2031353856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883049147442327,
      "loss": 1.4137,
      "theoretical_loss": 3.4544117722318965,
      "tokens_seen": 2031419392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038828485456369105,
      "loss": 1.4058,
      "theoretical_loss": 3.4544025853059055,
      "tokens_seen": 2031484928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882647943831494,
      "loss": 1.3437,
      "theoretical_loss": 3.4543933987592617,
      "tokens_seen": 2031550464
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038824473420260783,
      "loss": 1.3349,
      "theoretical_loss": 3.454384212591938,
      "tokens_seen": 2031616000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882246740220662,
      "loss": 1.4231,
      "theoretical_loss": 3.4543750268039064,
      "tokens_seen": 2031681536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038820461384152455,
      "loss": 1.3624,
      "theoretical_loss": 3.4543658413951386,
      "tokens_seen": 2031747072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881845536609829,
      "loss": 1.3414,
      "theoretical_loss": 3.454356656365607,
      "tokens_seen": 2031812608
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038816449348044133,
      "loss": 1.3464,
      "theoretical_loss": 3.454347471715284,
      "tokens_seen": 2031878144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881444332998997,
      "loss": 1.3932,
      "theoretical_loss": 3.4543382874441413,
      "tokens_seen": 2031943680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881243731193581,
      "loss": 1.3818,
      "theoretical_loss": 3.454329103552151,
      "tokens_seen": 2032009216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038810431293881647,
      "loss": 1.2806,
      "theoretical_loss": 3.454319920039285,
      "tokens_seen": 2032074752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038808425275827483,
      "loss": 1.482,
      "theoretical_loss": 3.4543107369055166,
      "tokens_seen": 2032140288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038806419257773325,
      "loss": 1.4247,
      "theoretical_loss": 3.4543015541508164,
      "tokens_seen": 2032205824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003880441323971916,
      "loss": 1.3067,
      "theoretical_loss": 3.454292371775158,
      "tokens_seen": 2032271360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038802407221665,
      "loss": 1.4338,
      "theoretical_loss": 3.4542831897785122,
      "tokens_seen": 2032336896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038800401203610834,
      "loss": 1.2584,
      "theoretical_loss": 3.4542740081608514,
      "tokens_seen": 2032402432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038798395185556675,
      "loss": 1.4509,
      "theoretical_loss": 3.4542648269221488,
      "tokens_seen": 2032467968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003879638916750251,
      "loss": 1.2899,
      "theoretical_loss": 3.454255646062375,
      "tokens_seen": 2032533504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003879438314944835,
      "loss": 1.4176,
      "theoretical_loss": 3.4542464655815035,
      "tokens_seen": 2032599040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038792377131394184,
      "loss": 1.4586,
      "theoretical_loss": 3.454237285479506,
      "tokens_seen": 2032664576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003879037111334002,
      "loss": 1.3193,
      "theoretical_loss": 3.454228105756354,
      "tokens_seen": 2032730112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878836509528586,
      "loss": 1.2534,
      "theoretical_loss": 3.4542189264120204,
      "tokens_seen": 2032795648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387863590772317,
      "loss": 1.3168,
      "theoretical_loss": 3.4542097474464772,
      "tokens_seen": 2032861184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038784353059177534,
      "loss": 1.4072,
      "theoretical_loss": 3.4542005688596964,
      "tokens_seen": 2032926720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878234704112337,
      "loss": 1.3353,
      "theoretical_loss": 3.4541913906516504,
      "tokens_seen": 2032992256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878034102306921,
      "loss": 1.3053,
      "theoretical_loss": 3.454182212822311,
      "tokens_seen": 2033057792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877833500501505,
      "loss": 1.4108,
      "theoretical_loss": 3.4541730353716504,
      "tokens_seen": 2033123328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038776328986960885,
      "loss": 1.4252,
      "theoretical_loss": 3.4541638582996415,
      "tokens_seen": 2033188864
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877432296890672,
      "loss": 1.3502,
      "theoretical_loss": 3.454154681606256,
      "tokens_seen": 2033254400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038772316950852557,
      "loss": 1.4077,
      "theoretical_loss": 3.454145505291465,
      "tokens_seen": 2033319936
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 741743,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2943028211593628,
      "objective/train/theoretical_loss": 3.4541363293552427,
      "objective/train/tokens_used": 403844576,
      "theoretical_loss": 3.4541363293552427,
      "tokens_seen": 2033385472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387703109327984,
      "loss": 1.2981,
      "theoretical_loss": 3.4541363293552427,
      "tokens_seen": 2033385472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038768304914744235,
      "loss": 1.4162,
      "theoretical_loss": 3.4541271537975593,
      "tokens_seen": 2033451008
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876629889669007,
      "loss": 1.3723,
      "theoretical_loss": 3.4541179786183887,
      "tokens_seen": 2033516544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876429287863591,
      "loss": 1.3366,
      "theoretical_loss": 3.454108803817702,
      "tokens_seen": 2033582080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876228686058175,
      "loss": 1.3827,
      "theoretical_loss": 3.454099629395472,
      "tokens_seen": 2033647616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038760280842527585,
      "loss": 1.301,
      "theoretical_loss": 3.45409045535167,
      "tokens_seen": 2033713152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875827482447342,
      "loss": 1.3125,
      "theoretical_loss": 3.4540812816862694,
      "tokens_seen": 2033778688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875626880641926,
      "loss": 1.3894,
      "theoretical_loss": 3.4540721083992416,
      "tokens_seen": 2033844224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387542627883651,
      "loss": 1.3829,
      "theoretical_loss": 3.454062935490559,
      "tokens_seen": 2033909760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038752256770310936,
      "loss": 1.2911,
      "theoretical_loss": 3.454053762960194,
      "tokens_seen": 2033975296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875025075225677,
      "loss": 1.3809,
      "theoretical_loss": 3.4540445908081185,
      "tokens_seen": 2034040832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874824473420261,
      "loss": 1.2672,
      "theoretical_loss": 3.454035419034305,
      "tokens_seen": 2034106368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038746238716148444,
      "loss": 1.3431,
      "theoretical_loss": 3.4540262476387253,
      "tokens_seen": 2034171904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038744232698094286,
      "loss": 1.3566,
      "theoretical_loss": 3.4540170766213523,
      "tokens_seen": 2034237440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874222668004012,
      "loss": 1.4055,
      "theoretical_loss": 3.4540079059821576,
      "tokens_seen": 2034302976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874022066198596,
      "loss": 1.3575,
      "theoretical_loss": 3.453998735721113,
      "tokens_seen": 2034368512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038738214643931795,
      "loss": 1.2981,
      "theoretical_loss": 3.453989565838192,
      "tokens_seen": 2034434048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038736208625877636,
      "loss": 1.3379,
      "theoretical_loss": 3.4539803963333657,
      "tokens_seen": 2034499584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873420260782347,
      "loss": 1.3378,
      "theoretical_loss": 3.4539712272066074,
      "tokens_seen": 2034565120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873219658976931,
      "loss": 1.3163,
      "theoretical_loss": 3.453962058457888,
      "tokens_seen": 2034630656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038730190571715145,
      "loss": 1.3456,
      "theoretical_loss": 3.453952890087181,
      "tokens_seen": 2034696192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872818455366098,
      "loss": 1.356,
      "theoretical_loss": 3.4539437220944578,
      "tokens_seen": 2034761728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038726178535606823,
      "loss": 1.2547,
      "theoretical_loss": 3.453934554479691,
      "tokens_seen": 2034827264
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872417251755266,
      "loss": 1.3025,
      "theoretical_loss": 3.453925387242853,
      "tokens_seen": 2034892800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038722166499498495,
      "loss": 1.2914,
      "theoretical_loss": 3.453916220383916,
      "tokens_seen": 2034958336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872016048144433,
      "loss": 1.4451,
      "theoretical_loss": 3.4539070539028516,
      "tokens_seen": 2035023872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038718154463390173,
      "loss": 1.3704,
      "theoretical_loss": 3.4538978877996325,
      "tokens_seen": 2035089408
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871614844533601,
      "loss": 1.3373,
      "theoretical_loss": 3.4538887220742316,
      "tokens_seen": 2035154944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038714142427281846,
      "loss": 1.2474,
      "theoretical_loss": 3.45387955672662,
      "tokens_seen": 2035220480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871213640922768,
      "loss": 1.3338,
      "theoretical_loss": 3.4538703917567712,
      "tokens_seen": 2035286016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871013039117352,
      "loss": 1.3395,
      "theoretical_loss": 3.4538612271646563,
      "tokens_seen": 2035351552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870812437311936,
      "loss": 1.3663,
      "theoretical_loss": 3.453852062950248,
      "tokens_seen": 2035417088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038706118355065196,
      "loss": 1.2895,
      "theoretical_loss": 3.453842899113519,
      "tokens_seen": 2035482624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870411233701103,
      "loss": 1.3297,
      "theoretical_loss": 3.453833735654441,
      "tokens_seen": 2035548160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870210631895687,
      "loss": 1.2956,
      "theoretical_loss": 3.453824572572987,
      "tokens_seen": 2035613696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870010030090271,
      "loss": 1.3892,
      "theoretical_loss": 3.453815409869128,
      "tokens_seen": 2035679232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038698094282848546,
      "loss": 1.3754,
      "theoretical_loss": 3.4538062475428375,
      "tokens_seen": 2035744768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003869608826479438,
      "loss": 1.3424,
      "theoretical_loss": 3.453797085594087,
      "tokens_seen": 2035810304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003869408224674022,
      "loss": 1.4326,
      "theoretical_loss": 3.4537879240228495,
      "tokens_seen": 2035875840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038692076228686055,
      "loss": 1.3881,
      "theoretical_loss": 3.453778762829097,
      "tokens_seen": 2035941376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038690070210631896,
      "loss": 1.4186,
      "theoretical_loss": 3.4537696020128017,
      "tokens_seen": 2036006912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038688064192577733,
      "loss": 1.3043,
      "theoretical_loss": 3.4537604415739356,
      "tokens_seen": 2036072448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003868605817452357,
      "loss": 1.2976,
      "theoretical_loss": 3.4537512815124716,
      "tokens_seen": 2036137984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038684052156469405,
      "loss": 1.3379,
      "theoretical_loss": 3.4537421218283817,
      "tokens_seen": 2036203520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038682046138415247,
      "loss": 1.3641,
      "theoretical_loss": 3.4537329625216384,
      "tokens_seen": 2036269056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038680040120361083,
      "loss": 1.3004,
      "theoretical_loss": 3.4537238035922133,
      "tokens_seen": 2036334592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867803410230692,
      "loss": 1.4283,
      "theoretical_loss": 3.4537146450400797,
      "tokens_seen": 2036400128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038676028084252755,
      "loss": 1.3937,
      "theoretical_loss": 3.4537054868652097,
      "tokens_seen": 2036465664
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867402206619859,
      "loss": 1.3937,
      "theoretical_loss": 3.4536963290675753,
      "tokens_seen": 2036531200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038672016048144433,
      "loss": 1.3632,
      "theoretical_loss": 3.453687171647149,
      "tokens_seen": 2036596736
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 742794,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5007350444793701,
      "objective/train/theoretical_loss": 3.4536780146039026,
      "objective/train/tokens_used": 407121376,
      "theoretical_loss": 3.4536780146039026,
      "tokens_seen": 2036662272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867001003009027,
      "loss": 1.4683,
      "theoretical_loss": 3.4536780146039026,
      "tokens_seen": 2036662272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038668004012036106,
      "loss": 1.5579,
      "theoretical_loss": 3.453668857937809,
      "tokens_seen": 2036727808
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003866599799398194,
      "loss": 1.3813,
      "theoretical_loss": 3.4536597016488404,
      "tokens_seen": 2036793344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038663991975927784,
      "loss": 1.326,
      "theoretical_loss": 3.4536505457369695,
      "tokens_seen": 2036858880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038661985957873625,
      "loss": 1.3433,
      "theoretical_loss": 3.453641390202168,
      "tokens_seen": 2036924416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865997993981946,
      "loss": 1.4184,
      "theoretical_loss": 3.4536322350444086,
      "tokens_seen": 2036989952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000386579739217653,
      "loss": 1.2716,
      "theoretical_loss": 3.4536230802636636,
      "tokens_seen": 2037055488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865596790371114,
      "loss": 1.3808,
      "theoretical_loss": 3.4536139258599055,
      "tokens_seen": 2037121024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038653961885656976,
      "loss": 1.3594,
      "theoretical_loss": 3.453604771833106,
      "tokens_seen": 2037186560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865195586760281,
      "loss": 1.211,
      "theoretical_loss": 3.4535956181832383,
      "tokens_seen": 2037252096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864994984954865,
      "loss": 1.3527,
      "theoretical_loss": 3.4535864649102743,
      "tokens_seen": 2037317632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038647943831494484,
      "loss": 1.3579,
      "theoretical_loss": 3.453577312014186,
      "tokens_seen": 2037383168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038645937813440326,
      "loss": 1.404,
      "theoretical_loss": 3.4535681594949468,
      "tokens_seen": 2037448704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864393179538616,
      "loss": 1.3965,
      "theoretical_loss": 3.4535590073525286,
      "tokens_seen": 2037514240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038641925777332,
      "loss": 1.3719,
      "theoretical_loss": 3.4535498555869033,
      "tokens_seen": 2037579776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038639919759277835,
      "loss": 1.4499,
      "theoretical_loss": 3.4535407041980433,
      "tokens_seen": 2037645312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038637913741223676,
      "loss": 1.3947,
      "theoretical_loss": 3.453531553185922,
      "tokens_seen": 2037710848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003863590772316951,
      "loss": 1.2748,
      "theoretical_loss": 3.45352240255051,
      "tokens_seen": 2037776384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003863390170511535,
      "loss": 1.2704,
      "theoretical_loss": 3.4535132522917813,
      "tokens_seen": 2037841920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038631895687061185,
      "loss": 1.4036,
      "theoretical_loss": 3.4535041024097075,
      "tokens_seen": 2037907456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862988966900702,
      "loss": 1.343,
      "theoretical_loss": 3.4534949529042613,
      "tokens_seen": 2037972992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038627883650952863,
      "loss": 1.285,
      "theoretical_loss": 3.4534858037754153,
      "tokens_seen": 2038038528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000386258776328987,
      "loss": 1.2852,
      "theoretical_loss": 3.453476655023141,
      "tokens_seen": 2038104064
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038623871614844535,
      "loss": 1.4242,
      "theoretical_loss": 3.453467506647412,
      "tokens_seen": 2038169600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862186559679037,
      "loss": 1.2899,
      "theoretical_loss": 3.4534583586481995,
      "tokens_seen": 2038235136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038619859578736213,
      "loss": 1.2833,
      "theoretical_loss": 3.453449211025476,
      "tokens_seen": 2038300672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861785356068205,
      "loss": 1.333,
      "theoretical_loss": 3.4534400637792153,
      "tokens_seen": 2038366208
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038615847542627886,
      "loss": 1.2974,
      "theoretical_loss": 3.4534309169093884,
      "tokens_seen": 2038431744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861384152457372,
      "loss": 1.3435,
      "theoretical_loss": 3.453421770415968,
      "tokens_seen": 2038497280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861183550651956,
      "loss": 1.4038,
      "theoretical_loss": 3.453412624298927,
      "tokens_seen": 2038562816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000386098294884654,
      "loss": 1.3596,
      "theoretical_loss": 3.4534034785582373,
      "tokens_seen": 2038628352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038607823470411236,
      "loss": 1.2759,
      "theoretical_loss": 3.4533943331938715,
      "tokens_seen": 2038693888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860581745235707,
      "loss": 1.3478,
      "theoretical_loss": 3.453385188205802,
      "tokens_seen": 2038759424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860381143430291,
      "loss": 1.3647,
      "theoretical_loss": 3.4533760435940013,
      "tokens_seen": 2038824960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860180541624875,
      "loss": 1.4712,
      "theoretical_loss": 3.453366899358442,
      "tokens_seen": 2038890496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038599799398194586,
      "loss": 1.4224,
      "theoretical_loss": 3.4533577554990957,
      "tokens_seen": 2038956032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003859779338014042,
      "loss": 1.2756,
      "theoretical_loss": 3.453348612015936,
      "tokens_seen": 2039021568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003859578736208626,
      "loss": 1.3375,
      "theoretical_loss": 3.453339468908934,
      "tokens_seen": 2039087104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038593781344032095,
      "loss": 1.3705,
      "theoretical_loss": 3.4533303261780635,
      "tokens_seen": 2039152640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038591775325977937,
      "loss": 1.4297,
      "theoretical_loss": 3.453321183823296,
      "tokens_seen": 2039218176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038589769307923773,
      "loss": 1.3233,
      "theoretical_loss": 3.4533120418446046,
      "tokens_seen": 2039283712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858776328986961,
      "loss": 1.5151,
      "theoretical_loss": 3.453302900241961,
      "tokens_seen": 2039349248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038585757271815445,
      "loss": 1.2903,
      "theoretical_loss": 3.453293759015338,
      "tokens_seen": 2039414784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038583751253761287,
      "loss": 1.3717,
      "theoretical_loss": 3.4532846181647088,
      "tokens_seen": 2039480320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038581745235707123,
      "loss": 1.4175,
      "theoretical_loss": 3.4532754776900445,
      "tokens_seen": 2039545856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857973921765296,
      "loss": 1.3789,
      "theoretical_loss": 3.4532663375913186,
      "tokens_seen": 2039611392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038577733199598796,
      "loss": 1.3848,
      "theoretical_loss": 3.4532571978685027,
      "tokens_seen": 2039676928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857572718154463,
      "loss": 1.4254,
      "theoretical_loss": 3.45324805852157,
      "tokens_seen": 2039742464
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038573721163490473,
      "loss": 1.2984,
      "theoretical_loss": 3.4532389195504933,
      "tokens_seen": 2039808000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857171514543631,
      "loss": 1.3998,
      "theoretical_loss": 3.453229780955244,
      "tokens_seen": 2039873536
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 744036,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4734983444213867,
      "objective/train/theoretical_loss": 3.4532206427357948,
      "objective/train/tokens_used": 410398176,
      "theoretical_loss": 3.4532206427357948,
      "tokens_seen": 2039939072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038569709127382146,
      "loss": 1.4085,
      "theoretical_loss": 3.4532206427357948,
      "tokens_seen": 2039939072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856770310932798,
      "loss": 1.3001,
      "theoretical_loss": 3.4532115048921184,
      "tokens_seen": 2040004608
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038565697091273824,
      "loss": 1.4096,
      "theoretical_loss": 3.4532023674241876,
      "tokens_seen": 2040070144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856369107321966,
      "loss": 1.3953,
      "theoretical_loss": 3.4531932303319746,
      "tokens_seen": 2040135680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038561685055165496,
      "loss": 1.3066,
      "theoretical_loss": 3.4531840936154516,
      "tokens_seen": 2040201216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855967903711133,
      "loss": 1.4007,
      "theoretical_loss": 3.4531749572745913,
      "tokens_seen": 2040266752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855767301905717,
      "loss": 1.3585,
      "theoretical_loss": 3.4531658213093666,
      "tokens_seen": 2040332288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855566700100301,
      "loss": 1.3067,
      "theoretical_loss": 3.4531566857197493,
      "tokens_seen": 2040397824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038553660982948846,
      "loss": 1.2366,
      "theoretical_loss": 3.4531475505057125,
      "tokens_seen": 2040463360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855165496489468,
      "loss": 1.4365,
      "theoretical_loss": 3.453138415667228,
      "tokens_seen": 2040528896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854964894684052,
      "loss": 1.3368,
      "theoretical_loss": 3.453129281204269,
      "tokens_seen": 2040594432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854764292878636,
      "loss": 1.4471,
      "theoretical_loss": 3.4531201471168074,
      "tokens_seen": 2040659968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038545636910732197,
      "loss": 1.3365,
      "theoretical_loss": 3.4531110134048166,
      "tokens_seen": 2040725504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038543630892678033,
      "loss": 1.4399,
      "theoretical_loss": 3.453101880068268,
      "tokens_seen": 2040791040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854162487462387,
      "loss": 1.3237,
      "theoretical_loss": 3.4530927471071347,
      "tokens_seen": 2040856576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038539618856569705,
      "loss": 1.3338,
      "theoretical_loss": 3.4530836145213892,
      "tokens_seen": 2040922112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038537612838515547,
      "loss": 1.4383,
      "theoretical_loss": 3.453074482311004,
      "tokens_seen": 2040987648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038535606820461383,
      "loss": 1.2489,
      "theoretical_loss": 3.453065350475952,
      "tokens_seen": 2041053184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003853360080240722,
      "loss": 1.327,
      "theoretical_loss": 3.4530562190162044,
      "tokens_seen": 2041118720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038531594784353056,
      "loss": 1.3274,
      "theoretical_loss": 3.4530470879317354,
      "tokens_seen": 2041184256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000385295887662989,
      "loss": 1.2954,
      "theoretical_loss": 3.4530379572225165,
      "tokens_seen": 2041249792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038527582748244734,
      "loss": 1.3721,
      "theoretical_loss": 3.4530288268885205,
      "tokens_seen": 2041315328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852557673019057,
      "loss": 1.334,
      "theoretical_loss": 3.45301969692972,
      "tokens_seen": 2041380864
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038523570712136406,
      "loss": 1.3752,
      "theoretical_loss": 3.4530105673460874,
      "tokens_seen": 2041446400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852156469408225,
      "loss": 1.3378,
      "theoretical_loss": 3.4530014381375955,
      "tokens_seen": 2041511936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038519558676028084,
      "loss": 1.4225,
      "theoretical_loss": 3.4529923093042165,
      "tokens_seen": 2041577472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851755265797392,
      "loss": 1.4502,
      "theoretical_loss": 3.452983180845923,
      "tokens_seen": 2041643008
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038515546639919756,
      "loss": 1.4751,
      "theoretical_loss": 3.4529740527626878,
      "tokens_seen": 2041708544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851354062186559,
      "loss": 1.3939,
      "theoretical_loss": 3.4529649250544834,
      "tokens_seen": 2041774080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851153460381144,
      "loss": 1.2775,
      "theoretical_loss": 3.452955797721282,
      "tokens_seen": 2041839616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038509528585757276,
      "loss": 1.4563,
      "theoretical_loss": 3.4529466707630565,
      "tokens_seen": 2041905152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850752256770311,
      "loss": 1.4465,
      "theoretical_loss": 3.45293754417978,
      "tokens_seen": 2041970688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850551654964895,
      "loss": 1.2924,
      "theoretical_loss": 3.4529284179714237,
      "tokens_seen": 2042036224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850351053159479,
      "loss": 1.3668,
      "theoretical_loss": 3.4529192921379606,
      "tokens_seen": 2042101760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038501504513540626,
      "loss": 1.3468,
      "theoretical_loss": 3.4529101666793642,
      "tokens_seen": 2042167296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003849949849548646,
      "loss": 1.2854,
      "theoretical_loss": 3.4529010415956063,
      "tokens_seen": 2042232832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384974924774323,
      "loss": 1.3504,
      "theoretical_loss": 3.4528919168866596,
      "tokens_seen": 2042298368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038495486459378135,
      "loss": 1.3947,
      "theoretical_loss": 3.4528827925524967,
      "tokens_seen": 2042363904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038493480441323977,
      "loss": 1.3875,
      "theoretical_loss": 3.4528736685930905,
      "tokens_seen": 2042429440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038491474423269813,
      "loss": 1.4093,
      "theoretical_loss": 3.4528645450084126,
      "tokens_seen": 2042494976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003848946840521565,
      "loss": 1.4187,
      "theoretical_loss": 3.4528554217984366,
      "tokens_seen": 2042560512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038487462387161485,
      "loss": 1.3936,
      "theoretical_loss": 3.4528462989631348,
      "tokens_seen": 2042626048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038485456369107327,
      "loss": 1.3232,
      "theoretical_loss": 3.45283717650248,
      "tokens_seen": 2042691584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038483450351053163,
      "loss": 1.3368,
      "theoretical_loss": 3.452828054416444,
      "tokens_seen": 2042757120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038481444332999,
      "loss": 1.4154,
      "theoretical_loss": 3.452818932705,
      "tokens_seen": 2042822656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038479438314944836,
      "loss": 1.3316,
      "theoretical_loss": 3.452809811368121,
      "tokens_seen": 2042888192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003847743229689067,
      "loss": 1.418,
      "theoretical_loss": 3.4528006904057786,
      "tokens_seen": 2042953728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038475426278836513,
      "loss": 1.4383,
      "theoretical_loss": 3.452791569817946,
      "tokens_seen": 2043019264
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003847342026078235,
      "loss": 1.353,
      "theoretical_loss": 3.4527824496045962,
      "tokens_seen": 2043084800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038471414242728186,
      "loss": 1.2809,
      "theoretical_loss": 3.452773329765701,
      "tokens_seen": 2043150336
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 745253,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.362492561340332,
      "objective/train/theoretical_loss": 3.4527642103012335,
      "objective/train/tokens_used": 413674976,
      "theoretical_loss": 3.4527642103012335,
      "tokens_seen": 2043215872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846940822467402,
      "loss": 1.3692,
      "theoretical_loss": 3.4527642103012335,
      "tokens_seen": 2043215872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038467402206619864,
      "loss": 1.2931,
      "theoretical_loss": 3.452755091211166,
      "tokens_seen": 2043281408
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384653961885657,
      "loss": 1.2935,
      "theoretical_loss": 3.452745972495472,
      "tokens_seen": 2043346944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038463390170511536,
      "loss": 1.2623,
      "theoretical_loss": 3.4527368541541223,
      "tokens_seen": 2043412480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846138415245737,
      "loss": 1.3338,
      "theoretical_loss": 3.452727736187091,
      "tokens_seen": 2043478016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845937813440321,
      "loss": 1.3515,
      "theoretical_loss": 3.452718618594351,
      "tokens_seen": 2043543552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845737211634905,
      "loss": 1.3427,
      "theoretical_loss": 3.4527095013758737,
      "tokens_seen": 2043609088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038455366098294886,
      "loss": 1.2894,
      "theoretical_loss": 3.452700384531633,
      "tokens_seen": 2043674624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038453360080240723,
      "loss": 1.2934,
      "theoretical_loss": 3.4526912680616006,
      "tokens_seen": 2043740160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845135406218656,
      "loss": 1.2907,
      "theoretical_loss": 3.4526821519657496,
      "tokens_seen": 2043805696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384493480441324,
      "loss": 1.297,
      "theoretical_loss": 3.452673036244052,
      "tokens_seen": 2043871232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038447342026078237,
      "loss": 1.305,
      "theoretical_loss": 3.452663920896481,
      "tokens_seen": 2043936768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038445336008024073,
      "loss": 1.3429,
      "theoretical_loss": 3.45265480592301,
      "tokens_seen": 2044002304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003844332998996991,
      "loss": 1.3216,
      "theoretical_loss": 3.4526456913236103,
      "tokens_seen": 2044067840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038441323971915745,
      "loss": 1.3142,
      "theoretical_loss": 3.4526365770982546,
      "tokens_seen": 2044133376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038439317953861587,
      "loss": 1.3304,
      "theoretical_loss": 3.452627463246917,
      "tokens_seen": 2044198912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038437311935807423,
      "loss": 1.3744,
      "theoretical_loss": 3.452618349769568,
      "tokens_seen": 2044264448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843530591775326,
      "loss": 1.3371,
      "theoretical_loss": 3.4526092366661825,
      "tokens_seen": 2044329984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038433299899699096,
      "loss": 1.2915,
      "theoretical_loss": 3.4526001239367323,
      "tokens_seen": 2044395520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843129388164494,
      "loss": 1.3101,
      "theoretical_loss": 3.4525910115811893,
      "tokens_seen": 2044461056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038429287863590774,
      "loss": 1.2463,
      "theoretical_loss": 3.4525818995995268,
      "tokens_seen": 2044526592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842728184553661,
      "loss": 1.3502,
      "theoretical_loss": 3.452572787991718,
      "tokens_seen": 2044592128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038425275827482446,
      "loss": 1.505,
      "theoretical_loss": 3.452563676757735,
      "tokens_seen": 2044657664
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842326980942829,
      "loss": 1.3697,
      "theoretical_loss": 3.4525545658975494,
      "tokens_seen": 2044723200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038421263791374124,
      "loss": 1.3999,
      "theoretical_loss": 3.452545455411136,
      "tokens_seen": 2044788736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841925777331996,
      "loss": 1.3617,
      "theoretical_loss": 3.4525363452984665,
      "tokens_seen": 2044854272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038417251755265796,
      "loss": 1.2997,
      "theoretical_loss": 3.452527235559513,
      "tokens_seen": 2044919808
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841524573721163,
      "loss": 1.1854,
      "theoretical_loss": 3.4525181261942492,
      "tokens_seen": 2044985344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038413239719157474,
      "loss": 1.3809,
      "theoretical_loss": 3.452509017202648,
      "tokens_seen": 2045050880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841123370110331,
      "loss": 1.3263,
      "theoretical_loss": 3.452499908584681,
      "tokens_seen": 2045116416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038409227683049147,
      "loss": 1.3863,
      "theoretical_loss": 3.4524908003403207,
      "tokens_seen": 2045181952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038407221664994983,
      "loss": 1.3317,
      "theoretical_loss": 3.452481692469541,
      "tokens_seen": 2045247488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038405215646940825,
      "loss": 1.3132,
      "theoretical_loss": 3.4524725849723144,
      "tokens_seen": 2045313024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003840320962888666,
      "loss": 1.2879,
      "theoretical_loss": 3.452463477848613,
      "tokens_seen": 2045378560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038401203610832497,
      "loss": 1.3786,
      "theoretical_loss": 3.4524543710984097,
      "tokens_seen": 2045444096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038399197592778333,
      "loss": 1.2848,
      "theoretical_loss": 3.4524452647216775,
      "tokens_seen": 2045509632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003839719157472417,
      "loss": 1.3435,
      "theoretical_loss": 3.452436158718389,
      "tokens_seen": 2045575168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003839518555667001,
      "loss": 1.2941,
      "theoretical_loss": 3.452427053088517,
      "tokens_seen": 2045640704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003839317953861585,
      "loss": 1.2937,
      "theoretical_loss": 3.452417947832034,
      "tokens_seen": 2045706240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038391173520561684,
      "loss": 1.304,
      "theoretical_loss": 3.452408842948912,
      "tokens_seen": 2045771776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003838916750250752,
      "loss": 1.3078,
      "theoretical_loss": 3.4523997384391256,
      "tokens_seen": 2045837312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003838716148445336,
      "loss": 1.2936,
      "theoretical_loss": 3.4523906343026463,
      "tokens_seen": 2045902848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383851554663992,
      "loss": 1.3888,
      "theoretical_loss": 3.4523815305394465,
      "tokens_seen": 2045968384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038383149448345034,
      "loss": 1.372,
      "theoretical_loss": 3.4523724271495,
      "tokens_seen": 2046033920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003838114343029087,
      "loss": 1.466,
      "theoretical_loss": 3.452363324132779,
      "tokens_seen": 2046099456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038379137412236706,
      "loss": 1.3502,
      "theoretical_loss": 3.452354221489256,
      "tokens_seen": 2046164992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837713139418255,
      "loss": 1.4677,
      "theoretical_loss": 3.452345119218904,
      "tokens_seen": 2046230528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038375125376128384,
      "loss": 1.419,
      "theoretical_loss": 3.452336017321696,
      "tokens_seen": 2046296064
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837311935807422,
      "loss": 1.3453,
      "theoretical_loss": 3.452326915797604,
      "tokens_seen": 2046361600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038371113340020057,
      "loss": 1.322,
      "theoretical_loss": 3.4523178146466016,
      "tokens_seen": 2046427136
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 746480,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2623852491378784,
      "objective/train/theoretical_loss": 3.4523087138686614,
      "objective/train/tokens_used": 416951776,
      "theoretical_loss": 3.4523087138686614,
      "tokens_seen": 2046492672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383691073219659,
      "loss": 1.2568,
      "theoretical_loss": 3.4523087138686614,
      "tokens_seen": 2046492672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038367101303911735,
      "loss": 1.3009,
      "theoretical_loss": 3.452299613463756,
      "tokens_seen": 2046558208
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836509528585757,
      "loss": 1.3756,
      "theoretical_loss": 3.4522905134318576,
      "tokens_seen": 2046623744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836308926780341,
      "loss": 1.389,
      "theoretical_loss": 3.4522814137729396,
      "tokens_seen": 2046689280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836108324974925,
      "loss": 1.2688,
      "theoretical_loss": 3.4522723144869754,
      "tokens_seen": 2046754816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003835907723169509,
      "loss": 1.457,
      "theoretical_loss": 3.4522632155739363,
      "tokens_seen": 2046820352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038357071213640927,
      "loss": 1.4059,
      "theoretical_loss": 3.4522541170337964,
      "tokens_seen": 2046885888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038355065195586763,
      "loss": 1.2877,
      "theoretical_loss": 3.4522450188665275,
      "tokens_seen": 2046951424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383530591775326,
      "loss": 1.4025,
      "theoretical_loss": 3.452235921072103,
      "tokens_seen": 2047016960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003835105315947844,
      "loss": 1.2992,
      "theoretical_loss": 3.452226823650495,
      "tokens_seen": 2047082496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038349047141424277,
      "loss": 1.3467,
      "theoretical_loss": 3.4522177266016776,
      "tokens_seen": 2047148032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038347041123370113,
      "loss": 1.3536,
      "theoretical_loss": 3.452208629925622,
      "tokens_seen": 2047213568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003834503510531595,
      "loss": 1.4178,
      "theoretical_loss": 3.452199533622302,
      "tokens_seen": 2047279104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038343029087261786,
      "loss": 1.3349,
      "theoretical_loss": 3.4521904376916903,
      "tokens_seen": 2047344640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038341023069207627,
      "loss": 1.3552,
      "theoretical_loss": 3.4521813421337595,
      "tokens_seen": 2047410176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038339017051153463,
      "loss": 1.3091,
      "theoretical_loss": 3.4521722469484826,
      "tokens_seen": 2047475712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383370110330993,
      "loss": 1.3126,
      "theoretical_loss": 3.4521631521358316,
      "tokens_seen": 2047541248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038335005015045136,
      "loss": 1.3336,
      "theoretical_loss": 3.4521540576957808,
      "tokens_seen": 2047606784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003833299899699098,
      "loss": 1.3191,
      "theoretical_loss": 3.4521449636283013,
      "tokens_seen": 2047672320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038330992978936814,
      "loss": 1.2763,
      "theoretical_loss": 3.4521358699333677,
      "tokens_seen": 2047737856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832898696088265,
      "loss": 1.4055,
      "theoretical_loss": 3.4521267766109514,
      "tokens_seen": 2047803392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038326980942828486,
      "loss": 1.4188,
      "theoretical_loss": 3.452117683661026,
      "tokens_seen": 2047868928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832497492477433,
      "loss": 1.3475,
      "theoretical_loss": 3.4521085910835634,
      "tokens_seen": 2047934464
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038322968906720164,
      "loss": 1.3026,
      "theoretical_loss": 3.4520994988785376,
      "tokens_seen": 2048000000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038320962888666,
      "loss": 1.3318,
      "theoretical_loss": 3.4520904070459206,
      "tokens_seen": 2048065536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038318956870611836,
      "loss": 1.3502,
      "theoretical_loss": 3.4520813155856858,
      "tokens_seen": 2048131072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831695085255767,
      "loss": 1.4059,
      "theoretical_loss": 3.4520722244978055,
      "tokens_seen": 2048196608
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038314944834503514,
      "loss": 1.4382,
      "theoretical_loss": 3.4520631337822527,
      "tokens_seen": 2048262144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831293881644935,
      "loss": 1.3304,
      "theoretical_loss": 3.4520540434390012,
      "tokens_seen": 2048327680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038310932798395187,
      "loss": 1.3094,
      "theoretical_loss": 3.4520449534680218,
      "tokens_seen": 2048393216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038308926780341023,
      "loss": 1.3647,
      "theoretical_loss": 3.4520358638692894,
      "tokens_seen": 2048458752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038306920762286865,
      "loss": 1.2854,
      "theoretical_loss": 3.4520267746427757,
      "tokens_seen": 2048524288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383049147442327,
      "loss": 1.3308,
      "theoretical_loss": 3.4520176857884532,
      "tokens_seen": 2048589824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038302908726178537,
      "loss": 1.3543,
      "theoretical_loss": 3.452008597306296,
      "tokens_seen": 2048655360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038300902708124373,
      "loss": 1.3534,
      "theoretical_loss": 3.451999509196276,
      "tokens_seen": 2048720896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829889669007021,
      "loss": 1.3722,
      "theoretical_loss": 3.451990421458367,
      "tokens_seen": 2048786432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829689067201605,
      "loss": 1.3481,
      "theoretical_loss": 3.4519813340925407,
      "tokens_seen": 2048851968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829488465396189,
      "loss": 1.3788,
      "theoretical_loss": 3.451972247098771,
      "tokens_seen": 2048917504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038292878635907724,
      "loss": 1.3879,
      "theoretical_loss": 3.4519631604770296,
      "tokens_seen": 2048983040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829087261785356,
      "loss": 1.4259,
      "theoretical_loss": 3.4519540742272907,
      "tokens_seen": 2049048576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382888665997994,
      "loss": 1.2431,
      "theoretical_loss": 3.451944988349526,
      "tokens_seen": 2049114112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003828686058174524,
      "loss": 1.3701,
      "theoretical_loss": 3.451935902843709,
      "tokens_seen": 2049179648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038284854563691074,
      "loss": 1.275,
      "theoretical_loss": 3.4519268177098126,
      "tokens_seen": 2049245184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003828284854563691,
      "loss": 1.3213,
      "theoretical_loss": 3.4519177329478095,
      "tokens_seen": 2049310720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038280842527582746,
      "loss": 1.4012,
      "theoretical_loss": 3.451908648557673,
      "tokens_seen": 2049376256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827883650952859,
      "loss": 1.3776,
      "theoretical_loss": 3.451899564539375,
      "tokens_seen": 2049441792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038276830491474424,
      "loss": 1.4168,
      "theoretical_loss": 3.451890480892889,
      "tokens_seen": 2049507328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827482447342026,
      "loss": 1.3523,
      "theoretical_loss": 3.4518813976181884,
      "tokens_seen": 2049572864
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038272818455366097,
      "loss": 1.3579,
      "theoretical_loss": 3.4518723147152457,
      "tokens_seen": 2049638400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827081243731194,
      "loss": 1.3042,
      "theoretical_loss": 3.4518632321840332,
      "tokens_seen": 2049703936
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 747702,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.376419186592102,
      "objective/train/theoretical_loss": 3.4518541500245243,
      "objective/train/tokens_used": 420228576,
      "theoretical_loss": 3.4518541500245243,
      "tokens_seen": 2049769472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038268806419257775,
      "loss": 1.3937,
      "theoretical_loss": 3.4518541500245243,
      "tokens_seen": 2049769472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003826680040120361,
      "loss": 1.3062,
      "theoretical_loss": 3.4518450682366923,
      "tokens_seen": 2049835008
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038264794383149447,
      "loss": 1.3337,
      "theoretical_loss": 3.4518359868205097,
      "tokens_seen": 2049900544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038262788365095283,
      "loss": 1.3448,
      "theoretical_loss": 3.4518269057759494,
      "tokens_seen": 2049966080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038260782347041125,
      "loss": 1.2981,
      "theoretical_loss": 3.4518178251029843,
      "tokens_seen": 2050031616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003825877632898696,
      "loss": 1.4342,
      "theoretical_loss": 3.4518087448015873,
      "tokens_seen": 2050097152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382567703109328,
      "loss": 1.3958,
      "theoretical_loss": 3.4517996648717313,
      "tokens_seen": 2050162688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038254764292878634,
      "loss": 1.436,
      "theoretical_loss": 3.4517905853133897,
      "tokens_seen": 2050228224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038252758274824475,
      "loss": 1.4748,
      "theoretical_loss": 3.4517815061265344,
      "tokens_seen": 2050293760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003825075225677031,
      "loss": 1.2471,
      "theoretical_loss": 3.4517724273111394,
      "tokens_seen": 2050359296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824874623871615,
      "loss": 1.423,
      "theoretical_loss": 3.4517633488671766,
      "tokens_seen": 2050424832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038246740220661984,
      "loss": 1.2773,
      "theoretical_loss": 3.4517542707946203,
      "tokens_seen": 2050490368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824473420260782,
      "loss": 1.3578,
      "theoretical_loss": 3.4517451930934424,
      "tokens_seen": 2050555904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824272818455366,
      "loss": 1.3225,
      "theoretical_loss": 3.451736115763616,
      "tokens_seen": 2050621440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382407221664995,
      "loss": 1.4584,
      "theoretical_loss": 3.451727038805114,
      "tokens_seen": 2050686976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038238716148445334,
      "loss": 1.3118,
      "theoretical_loss": 3.4517179622179093,
      "tokens_seen": 2050752512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823671013039117,
      "loss": 1.3664,
      "theoretical_loss": 3.451708886001976,
      "tokens_seen": 2050818048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823470411233701,
      "loss": 1.2905,
      "theoretical_loss": 3.451699810157285,
      "tokens_seen": 2050883584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823269809428285,
      "loss": 1.3224,
      "theoretical_loss": 3.451690734683811,
      "tokens_seen": 2050949120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038230692076228685,
      "loss": 1.2581,
      "theoretical_loss": 3.4516816595815256,
      "tokens_seen": 2051014656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822868605817452,
      "loss": 1.5565,
      "theoretical_loss": 3.451672584850403,
      "tokens_seen": 2051080192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038226680040120357,
      "loss": 1.4085,
      "theoretical_loss": 3.451663510490415,
      "tokens_seen": 2051145728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382246740220662,
      "loss": 1.3184,
      "theoretical_loss": 3.451654436501536,
      "tokens_seen": 2051211264
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038222668004012035,
      "loss": 1.3814,
      "theoretical_loss": 3.451645362883738,
      "tokens_seen": 2051276800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822066198595787,
      "loss": 1.4038,
      "theoretical_loss": 3.451636289636993,
      "tokens_seen": 2051342336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821865596790371,
      "loss": 1.3888,
      "theoretical_loss": 3.4516272167612767,
      "tokens_seen": 2051407872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821664994984955,
      "loss": 1.3621,
      "theoretical_loss": 3.4516181442565594,
      "tokens_seen": 2051473408
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038214643931795385,
      "loss": 1.3239,
      "theoretical_loss": 3.4516090721228156,
      "tokens_seen": 2051538944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038212637913741227,
      "loss": 1.3418,
      "theoretical_loss": 3.451600000360017,
      "tokens_seen": 2051604480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038210631895687063,
      "loss": 1.41,
      "theoretical_loss": 3.4515909289681383,
      "tokens_seen": 2051670016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038208625877632905,
      "loss": 1.2409,
      "theoretical_loss": 3.451581857947151,
      "tokens_seen": 2051735552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003820661985957874,
      "loss": 1.3528,
      "theoretical_loss": 3.451572787297029,
      "tokens_seen": 2051801088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038204613841524577,
      "loss": 1.3039,
      "theoretical_loss": 3.451563717017745,
      "tokens_seen": 2051866624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038202607823470413,
      "loss": 1.3142,
      "theoretical_loss": 3.4515546471092717,
      "tokens_seen": 2051932160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003820060180541625,
      "loss": 1.387,
      "theoretical_loss": 3.4515455775715824,
      "tokens_seen": 2051997696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003819859578736209,
      "loss": 1.3922,
      "theoretical_loss": 3.4515365084046503,
      "tokens_seen": 2052063232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003819658976930793,
      "loss": 1.4019,
      "theoretical_loss": 3.451527439608448,
      "tokens_seen": 2052128768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038194583751253764,
      "loss": 1.2926,
      "theoretical_loss": 3.4515183711829485,
      "tokens_seen": 2052194304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000381925777331996,
      "loss": 1.3609,
      "theoretical_loss": 3.451509303128125,
      "tokens_seen": 2052259840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003819057171514544,
      "loss": 1.3647,
      "theoretical_loss": 3.451500235443951,
      "tokens_seen": 2052325376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818856569709128,
      "loss": 1.2831,
      "theoretical_loss": 3.4514911681303984,
      "tokens_seen": 2052390912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038186559679037114,
      "loss": 1.4023,
      "theoretical_loss": 3.451482101187441,
      "tokens_seen": 2052456448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818455366098295,
      "loss": 1.3395,
      "theoretical_loss": 3.4514730346150517,
      "tokens_seen": 2052521984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038182547642928786,
      "loss": 1.3827,
      "theoretical_loss": 3.4514639684132034,
      "tokens_seen": 2052587520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818054162487463,
      "loss": 1.355,
      "theoretical_loss": 3.4514549025818693,
      "tokens_seen": 2052653056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038178535606820464,
      "loss": 1.3568,
      "theoretical_loss": 3.451445837121022,
      "tokens_seen": 2052718592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000381765295887663,
      "loss": 1.4609,
      "theoretical_loss": 3.451436772030635,
      "tokens_seen": 2052784128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038174523570712137,
      "loss": 1.3821,
      "theoretical_loss": 3.451427707310682,
      "tokens_seen": 2052849664
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003817251755265798,
      "loss": 1.3841,
      "theoretical_loss": 3.451418642961134,
      "tokens_seen": 2052915200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038170511534603815,
      "loss": 1.3775,
      "theoretical_loss": 3.4514095789819654,
      "tokens_seen": 2052980736
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 748910,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4846608638763428,
      "objective/train/theoretical_loss": 3.45140051537315,
      "objective/train/tokens_used": 423505376,
      "theoretical_loss": 3.45140051537315,
      "tokens_seen": 2053046272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003816850551654965,
      "loss": 1.3551,
      "theoretical_loss": 3.45140051537315,
      "tokens_seen": 2053046272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038166499498495487,
      "loss": 1.4236,
      "theoretical_loss": 3.451391452134659,
      "tokens_seen": 2053111808
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038164493480441323,
      "loss": 1.379,
      "theoretical_loss": 3.451382389266467,
      "tokens_seen": 2053177344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038162487462387165,
      "loss": 1.3736,
      "theoretical_loss": 3.4513733267685462,
      "tokens_seen": 2053242880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038160481444333,
      "loss": 1.3852,
      "theoretical_loss": 3.45136426464087,
      "tokens_seen": 2053308416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815847542627884,
      "loss": 1.416,
      "theoretical_loss": 3.451355202883411,
      "tokens_seen": 2053373952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038156469408224674,
      "loss": 1.332,
      "theoretical_loss": 3.4513461414961433,
      "tokens_seen": 2053439488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038154463390170515,
      "loss": 1.446,
      "theoretical_loss": 3.451337080479039,
      "tokens_seen": 2053505024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815245737211635,
      "loss": 1.36,
      "theoretical_loss": 3.4513280198320713,
      "tokens_seen": 2053570560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815045135406219,
      "loss": 1.3645,
      "theoretical_loss": 3.4513189595552136,
      "tokens_seen": 2053636096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038148445336008024,
      "loss": 1.4119,
      "theoretical_loss": 3.4513098996484386,
      "tokens_seen": 2053701632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814643931795386,
      "loss": 1.3962,
      "theoretical_loss": 3.4513008401117196,
      "tokens_seen": 2053767168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000381444332998997,
      "loss": 1.2842,
      "theoretical_loss": 3.45129178094503,
      "tokens_seen": 2053832704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814242728184554,
      "loss": 1.3114,
      "theoretical_loss": 3.451282722148342,
      "tokens_seen": 2053898240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038140421263791374,
      "loss": 1.3962,
      "theoretical_loss": 3.4512736637216292,
      "tokens_seen": 2053963776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813841524573721,
      "loss": 1.2303,
      "theoretical_loss": 3.451264605664865,
      "tokens_seen": 2054029312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813640922768305,
      "loss": 1.3652,
      "theoretical_loss": 3.4512555479780223,
      "tokens_seen": 2054094848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813440320962889,
      "loss": 1.3699,
      "theoretical_loss": 3.451246490661074,
      "tokens_seen": 2054160384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038132397191574725,
      "loss": 1.3429,
      "theoretical_loss": 3.451237433713993,
      "tokens_seen": 2054225920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003813039117352056,
      "loss": 1.3754,
      "theoretical_loss": 3.451228377136753,
      "tokens_seen": 2054291456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038128385155466397,
      "loss": 1.4331,
      "theoretical_loss": 3.4512193209293267,
      "tokens_seen": 2054356992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812637913741224,
      "loss": 1.3229,
      "theoretical_loss": 3.451210265091687,
      "tokens_seen": 2054422528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038124373119358075,
      "loss": 1.394,
      "theoretical_loss": 3.4512012096238074,
      "tokens_seen": 2054488064
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812236710130391,
      "loss": 1.3932,
      "theoretical_loss": 3.4511921545256605,
      "tokens_seen": 2054553600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812036108324975,
      "loss": 1.3018,
      "theoretical_loss": 3.4511830997972206,
      "tokens_seen": 2054619136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003811835506519559,
      "loss": 1.2877,
      "theoretical_loss": 3.451174045438459,
      "tokens_seen": 2054684672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038116349047141425,
      "loss": 1.41,
      "theoretical_loss": 3.4511649914493505,
      "tokens_seen": 2054750208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003811434302908726,
      "loss": 1.4521,
      "theoretical_loss": 3.451155937829868,
      "tokens_seen": 2054815744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000381123370110331,
      "loss": 1.4979,
      "theoretical_loss": 3.451146884579983,
      "tokens_seen": 2054881280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038110330992978934,
      "loss": 1.4527,
      "theoretical_loss": 3.451137831699671,
      "tokens_seen": 2054946816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038108324974924776,
      "loss": 1.3333,
      "theoretical_loss": 3.451128779188903,
      "tokens_seen": 2055012352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810631895687061,
      "loss": 1.3992,
      "theoretical_loss": 3.4511197270476535,
      "tokens_seen": 2055077888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810431293881645,
      "loss": 1.3989,
      "theoretical_loss": 3.4511106752758947,
      "tokens_seen": 2055143424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038102306920762284,
      "loss": 1.3372,
      "theoretical_loss": 3.4511016238736008,
      "tokens_seen": 2055208960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038100300902708126,
      "loss": 1.3027,
      "theoretical_loss": 3.4510925728407438,
      "tokens_seen": 2055274496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003809829488465396,
      "loss": 1.355,
      "theoretical_loss": 3.4510835221772975,
      "tokens_seen": 2055340032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380962888665998,
      "loss": 1.3683,
      "theoretical_loss": 3.451074471883236,
      "tokens_seen": 2055405568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038094282848545635,
      "loss": 1.3575,
      "theoretical_loss": 3.45106542195853,
      "tokens_seen": 2055471104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038092276830491476,
      "loss": 1.383,
      "theoretical_loss": 3.4510563724031544,
      "tokens_seen": 2055536640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003809027081243731,
      "loss": 1.2824,
      "theoretical_loss": 3.4510473232170824,
      "tokens_seen": 2055602176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808826479438315,
      "loss": 1.4672,
      "theoretical_loss": 3.451038274400286,
      "tokens_seen": 2055667712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038086258776328985,
      "loss": 1.4254,
      "theoretical_loss": 3.4510292259527398,
      "tokens_seen": 2055733248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808425275827482,
      "loss": 1.3263,
      "theoretical_loss": 3.4510201778744163,
      "tokens_seen": 2055798784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808224674022066,
      "loss": 1.4186,
      "theoretical_loss": 3.451011130165288,
      "tokens_seen": 2055864320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380802407221665,
      "loss": 1.3537,
      "theoretical_loss": 3.451002082825329,
      "tokens_seen": 2055929856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038078234704112335,
      "loss": 1.3471,
      "theoretical_loss": 3.4509930358545127,
      "tokens_seen": 2055995392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807622868605817,
      "loss": 1.4668,
      "theoretical_loss": 3.450983989252811,
      "tokens_seen": 2056060928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038074222668004013,
      "loss": 1.4103,
      "theoretical_loss": 3.4509749430201984,
      "tokens_seen": 2056126464
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807221664994985,
      "loss": 1.3253,
      "theoretical_loss": 3.450965897156647,
      "tokens_seen": 2056192000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038070210631895685,
      "loss": 1.3438,
      "theoretical_loss": 3.4509568516621307,
      "tokens_seen": 2056257536
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 750166,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2493493556976318,
      "objective/train/theoretical_loss": 3.450947806536622,
      "objective/train/tokens_used": 426782176,
      "theoretical_loss": 3.450947806536622,
      "tokens_seen": 2056323072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806820461384152,
      "loss": 1.3098,
      "theoretical_loss": 3.450947806536622,
      "tokens_seen": 2056323072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806619859578736,
      "loss": 1.4908,
      "theoretical_loss": 3.4509387617800953,
      "tokens_seen": 2056388608
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380641925777332,
      "loss": 1.3482,
      "theoretical_loss": 3.4509297173925226,
      "tokens_seen": 2056454144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806218655967904,
      "loss": 1.316,
      "theoretical_loss": 3.4509206733738775,
      "tokens_seen": 2056519680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806018054162488,
      "loss": 1.3271,
      "theoretical_loss": 3.4509116297241333,
      "tokens_seen": 2056585216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038058174523570714,
      "loss": 1.3884,
      "theoretical_loss": 3.450902586443263,
      "tokens_seen": 2056650752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038056168505516555,
      "loss": 1.2964,
      "theoretical_loss": 3.4508935435312402,
      "tokens_seen": 2056716288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003805416248746239,
      "loss": 1.2602,
      "theoretical_loss": 3.450884500988038,
      "tokens_seen": 2056781824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003805215646940823,
      "loss": 1.3469,
      "theoretical_loss": 3.450875458813629,
      "tokens_seen": 2056847360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038050150451354064,
      "loss": 1.3771,
      "theoretical_loss": 3.450866417007987,
      "tokens_seen": 2056912896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380481444332999,
      "loss": 1.3656,
      "theoretical_loss": 3.450857375571085,
      "tokens_seen": 2056978432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804613841524574,
      "loss": 1.2998,
      "theoretical_loss": 3.450848334502896,
      "tokens_seen": 2057043968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804413239719158,
      "loss": 1.2712,
      "theoretical_loss": 3.4508392938033943,
      "tokens_seen": 2057109504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038042126379137414,
      "loss": 1.3993,
      "theoretical_loss": 3.4508302534725512,
      "tokens_seen": 2057175040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804012036108325,
      "loss": 1.3898,
      "theoretical_loss": 3.4508212135103418,
      "tokens_seen": 2057240576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803811434302909,
      "loss": 1.2993,
      "theoretical_loss": 3.4508121739167383,
      "tokens_seen": 2057306112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803610832497493,
      "loss": 1.4177,
      "theoretical_loss": 3.4508031346917143,
      "tokens_seen": 2057371648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038034102306920765,
      "loss": 1.3663,
      "theoretical_loss": 3.450794095835243,
      "tokens_seen": 2057437184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380320962888666,
      "loss": 1.3182,
      "theoretical_loss": 3.4507850573472973,
      "tokens_seen": 2057502720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038030090270812437,
      "loss": 1.1937,
      "theoretical_loss": 3.4507760192278507,
      "tokens_seen": 2057568256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802808425275828,
      "loss": 1.2975,
      "theoretical_loss": 3.4507669814768764,
      "tokens_seen": 2057633792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038026078234704115,
      "loss": 1.4395,
      "theoretical_loss": 3.450757944094348,
      "tokens_seen": 2057699328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802407221664995,
      "loss": 1.3469,
      "theoretical_loss": 3.450748907080238,
      "tokens_seen": 2057764864
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802206619859579,
      "loss": 1.2901,
      "theoretical_loss": 3.4507398704345205,
      "tokens_seen": 2057830400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802006018054163,
      "loss": 1.2503,
      "theoretical_loss": 3.4507308341571674,
      "tokens_seen": 2057895936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038018054162487465,
      "loss": 1.3031,
      "theoretical_loss": 3.4507217982481535,
      "tokens_seen": 2057961472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380160481444333,
      "loss": 1.467,
      "theoretical_loss": 3.450712762707452,
      "tokens_seen": 2058027008
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003801404212637914,
      "loss": 1.3075,
      "theoretical_loss": 3.4507037275350347,
      "tokens_seen": 2058092544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038012036108324974,
      "loss": 1.3497,
      "theoretical_loss": 3.450694692730876,
      "tokens_seen": 2058158080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038010030090270816,
      "loss": 1.3783,
      "theoretical_loss": 3.450685658294949,
      "tokens_seen": 2058223616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003800802407221665,
      "loss": 1.4522,
      "theoretical_loss": 3.450676624227227,
      "tokens_seen": 2058289152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003800601805416249,
      "loss": 1.3264,
      "theoretical_loss": 3.450667590527683,
      "tokens_seen": 2058354688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038004012036108324,
      "loss": 1.2982,
      "theoretical_loss": 3.45065855719629,
      "tokens_seen": 2058420224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038002006018054166,
      "loss": 1.3798,
      "theoretical_loss": 3.4506495242330226,
      "tokens_seen": 2058485760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038,
      "loss": 1.275,
      "theoretical_loss": 3.4506404916378526,
      "tokens_seen": 2058551296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003799799398194584,
      "loss": 1.2402,
      "theoretical_loss": 3.450631459410754,
      "tokens_seen": 2058616832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037995987963891675,
      "loss": 1.4422,
      "theoretical_loss": 3.4506224275517,
      "tokens_seen": 2058682368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037993981945837516,
      "loss": 1.2565,
      "theoretical_loss": 3.4506133960606635,
      "tokens_seen": 2058747904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003799197592778335,
      "loss": 1.3837,
      "theoretical_loss": 3.450604364937618,
      "tokens_seen": 2058813440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798996990972919,
      "loss": 1.4496,
      "theoretical_loss": 3.4505953341825375,
      "tokens_seen": 2058878976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037987963891675025,
      "loss": 1.4201,
      "theoretical_loss": 3.4505863037953945,
      "tokens_seen": 2058944512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798595787362086,
      "loss": 1.2644,
      "theoretical_loss": 3.4505772737761626,
      "tokens_seen": 2059010048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037983951855566703,
      "loss": 1.292,
      "theoretical_loss": 3.4505682441248147,
      "tokens_seen": 2059075584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798194583751254,
      "loss": 1.2584,
      "theoretical_loss": 3.4505592148413244,
      "tokens_seen": 2059141120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037979939819458375,
      "loss": 1.2263,
      "theoretical_loss": 3.4505501859256653,
      "tokens_seen": 2059206656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797793380140421,
      "loss": 1.3503,
      "theoretical_loss": 3.45054115737781,
      "tokens_seen": 2059272192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037975927783350053,
      "loss": 1.3637,
      "theoretical_loss": 3.450532129197733,
      "tokens_seen": 2059337728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797392176529589,
      "loss": 1.3186,
      "theoretical_loss": 3.4505231013854063,
      "tokens_seen": 2059403264
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037971915747241725,
      "loss": 1.4067,
      "theoretical_loss": 3.450514073940804,
      "tokens_seen": 2059468800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796990972918756,
      "loss": 1.3664,
      "theoretical_loss": 3.450505046863899,
      "tokens_seen": 2059534336
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 751330,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3781014680862427,
      "objective/train/theoretical_loss": 3.450496020154665,
      "objective/train/tokens_used": 430058976,
      "theoretical_loss": 3.450496020154665,
      "tokens_seen": 2059599872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379679037111334,
      "loss": 1.4111,
      "theoretical_loss": 3.450496020154665,
      "tokens_seen": 2059599872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796589769307924,
      "loss": 1.3682,
      "theoretical_loss": 3.450486993813075,
      "tokens_seen": 2059665408
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037963891675025076,
      "loss": 1.222,
      "theoretical_loss": 3.4504779678391024,
      "tokens_seen": 2059730944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796188565697091,
      "loss": 1.4279,
      "theoretical_loss": 3.4504689422327206,
      "tokens_seen": 2059796480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795987963891675,
      "loss": 1.3905,
      "theoretical_loss": 3.4504599169939034,
      "tokens_seen": 2059862016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795787362086259,
      "loss": 1.3263,
      "theoretical_loss": 3.450450892122623,
      "tokens_seen": 2059927552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037955867602808426,
      "loss": 1.43,
      "theoretical_loss": 3.4504418676188537,
      "tokens_seen": 2059993088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795386158475426,
      "loss": 1.2826,
      "theoretical_loss": 3.4504328434825684,
      "tokens_seen": 2060058624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379518555667001,
      "loss": 1.3222,
      "theoretical_loss": 3.450423819713741,
      "tokens_seen": 2060124160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037949849548645935,
      "loss": 1.3427,
      "theoretical_loss": 3.4504147963123444,
      "tokens_seen": 2060189696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037947843530591776,
      "loss": 1.3051,
      "theoretical_loss": 3.450405773278352,
      "tokens_seen": 2060255232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794583751253761,
      "loss": 1.3728,
      "theoretical_loss": 3.450396750611737,
      "tokens_seen": 2060320768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794383149448345,
      "loss": 1.3736,
      "theoretical_loss": 3.4503877283124726,
      "tokens_seen": 2060386304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037941825476429285,
      "loss": 1.3498,
      "theoretical_loss": 3.4503787063805333,
      "tokens_seen": 2060451840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037939819458375127,
      "loss": 1.4216,
      "theoretical_loss": 3.450369684815891,
      "tokens_seen": 2060517376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037937813440320963,
      "loss": 1.4388,
      "theoretical_loss": 3.45036066361852,
      "tokens_seen": 2060582912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379358074222668,
      "loss": 1.3142,
      "theoretical_loss": 3.450351642788393,
      "tokens_seen": 2060648448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037933801404212635,
      "loss": 1.3774,
      "theoretical_loss": 3.4503426223254845,
      "tokens_seen": 2060713984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003793179538615847,
      "loss": 1.2908,
      "theoretical_loss": 3.450333602229766,
      "tokens_seen": 2060779520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037929789368104313,
      "loss": 1.376,
      "theoretical_loss": 3.4503245825012128,
      "tokens_seen": 2060845056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792778335005015,
      "loss": 1.3131,
      "theoretical_loss": 3.4503155631397977,
      "tokens_seen": 2060910592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037925777331995986,
      "loss": 1.3065,
      "theoretical_loss": 3.4503065441454934,
      "tokens_seen": 2060976128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792377131394182,
      "loss": 1.3306,
      "theoretical_loss": 3.4502975255182733,
      "tokens_seen": 2061041664
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037921765295887664,
      "loss": 1.449,
      "theoretical_loss": 3.450288507258112,
      "tokens_seen": 2061107200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379197592778335,
      "loss": 1.3525,
      "theoretical_loss": 3.4502794893649815,
      "tokens_seen": 2061172736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037917753259779336,
      "loss": 1.3103,
      "theoretical_loss": 3.4502704718388566,
      "tokens_seen": 2061238272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791574724172517,
      "loss": 1.2614,
      "theoretical_loss": 3.4502614546797092,
      "tokens_seen": 2061303808
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791374122367101,
      "loss": 1.3807,
      "theoretical_loss": 3.4502524378875137,
      "tokens_seen": 2061369344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037911735205616856,
      "loss": 1.3671,
      "theoretical_loss": 3.4502434214622433,
      "tokens_seen": 2061434880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790972918756269,
      "loss": 1.3784,
      "theoretical_loss": 3.4502344054038714,
      "tokens_seen": 2061500416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790772316950853,
      "loss": 1.3529,
      "theoretical_loss": 3.4502253897123705,
      "tokens_seen": 2061565952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037905717151454364,
      "loss": 1.249,
      "theoretical_loss": 3.4502163743877157,
      "tokens_seen": 2061631488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037903711133400206,
      "loss": 1.3148,
      "theoretical_loss": 3.4502073594298794,
      "tokens_seen": 2061697024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790170511534604,
      "loss": 1.3856,
      "theoretical_loss": 3.450198344838835,
      "tokens_seen": 2061762560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003789969909729188,
      "loss": 1.3319,
      "theoretical_loss": 3.450189330614556,
      "tokens_seen": 2061828096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037897693079237715,
      "loss": 1.3713,
      "theoretical_loss": 3.4501803167570158,
      "tokens_seen": 2061893632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037895687061183556,
      "loss": 1.3886,
      "theoretical_loss": 3.4501713032661883,
      "tokens_seen": 2061959168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003789368104312939,
      "loss": 1.3708,
      "theoretical_loss": 3.450162290142046,
      "tokens_seen": 2062024704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003789167502507523,
      "loss": 1.4222,
      "theoretical_loss": 3.4501532773845636,
      "tokens_seen": 2062090240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037889669007021065,
      "loss": 1.3278,
      "theoretical_loss": 3.4501442649937135,
      "tokens_seen": 2062155776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378876629889669,
      "loss": 1.3538,
      "theoretical_loss": 3.4501352529694693,
      "tokens_seen": 2062221312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037885656970912743,
      "loss": 1.3442,
      "theoretical_loss": 3.450126241311805,
      "tokens_seen": 2062286848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003788365095285858,
      "loss": 1.3701,
      "theoretical_loss": 3.450117230020693,
      "tokens_seen": 2062352384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037881644934804415,
      "loss": 1.4017,
      "theoretical_loss": 3.4501082190961077,
      "tokens_seen": 2062417920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787963891675025,
      "loss": 1.3094,
      "theoretical_loss": 3.450099208538022,
      "tokens_seen": 2062483456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037877632898696093,
      "loss": 1.2367,
      "theoretical_loss": 3.45009019834641,
      "tokens_seen": 2062548992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787562688064193,
      "loss": 1.3556,
      "theoretical_loss": 3.450081188521244,
      "tokens_seen": 2062614528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037873620862587766,
      "loss": 1.2267,
      "theoretical_loss": 3.4500721790624986,
      "tokens_seen": 2062680064
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378716148445336,
      "loss": 1.3159,
      "theoretical_loss": 3.4500631699701465,
      "tokens_seen": 2062745600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786960882647944,
      "loss": 1.3928,
      "theoretical_loss": 3.450054161244162,
      "tokens_seen": 2062811136
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 752534,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.330669641494751,
      "objective/train/theoretical_loss": 3.4500451528845177,
      "objective/train/tokens_used": 433335776,
      "theoretical_loss": 3.4500451528845177,
      "tokens_seen": 2062876672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786760280842528,
      "loss": 1.3501,
      "theoretical_loss": 3.4500451528845177,
      "tokens_seen": 2062876672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037865596790371116,
      "loss": 1.3431,
      "theoretical_loss": 3.450036144891188,
      "tokens_seen": 2062942208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786359077231695,
      "loss": 1.3088,
      "theoretical_loss": 3.450027137264145,
      "tokens_seen": 2063007744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786158475426279,
      "loss": 1.3189,
      "theoretical_loss": 3.450018130003363,
      "tokens_seen": 2063073280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785957873620863,
      "loss": 1.4687,
      "theoretical_loss": 3.4500091231088157,
      "tokens_seen": 2063138816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037857572718154466,
      "loss": 1.3607,
      "theoretical_loss": 3.450000116580476,
      "tokens_seen": 2063204352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378555667001003,
      "loss": 1.3475,
      "theoretical_loss": 3.449991110418318,
      "tokens_seen": 2063269888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785356068204614,
      "loss": 1.4041,
      "theoretical_loss": 3.4499821046223147,
      "tokens_seen": 2063335424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037851554663991975,
      "loss": 1.3203,
      "theoretical_loss": 3.44997309919244,
      "tokens_seen": 2063400960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037849548645937816,
      "loss": 1.3866,
      "theoretical_loss": 3.449964094128667,
      "tokens_seen": 2063466496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784754262788365,
      "loss": 1.3617,
      "theoretical_loss": 3.4499550894309694,
      "tokens_seen": 2063532032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784553660982949,
      "loss": 1.2786,
      "theoretical_loss": 3.44994608509932,
      "tokens_seen": 2063597568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037843530591775325,
      "loss": 1.3252,
      "theoretical_loss": 3.449937081133694,
      "tokens_seen": 2063663104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037841524573721167,
      "loss": 1.3894,
      "theoretical_loss": 3.4499280775340626,
      "tokens_seen": 2063728640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037839518555667003,
      "loss": 1.3072,
      "theoretical_loss": 3.449919074300401,
      "tokens_seen": 2063794176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783751253761284,
      "loss": 1.2985,
      "theoretical_loss": 3.449910071432682,
      "tokens_seen": 2063859712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037835506519558675,
      "loss": 1.3177,
      "theoretical_loss": 3.4499010689308798,
      "tokens_seen": 2063925248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783350050150451,
      "loss": 1.3237,
      "theoretical_loss": 3.449892066794967,
      "tokens_seen": 2063990784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037831494483450353,
      "loss": 1.3256,
      "theoretical_loss": 3.4498830650249177,
      "tokens_seen": 2064056320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782948846539619,
      "loss": 1.3122,
      "theoretical_loss": 3.4498740636207055,
      "tokens_seen": 2064121856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037827482447342026,
      "loss": 1.3388,
      "theoretical_loss": 3.4498650625823033,
      "tokens_seen": 2064187392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782547642928786,
      "loss": 1.4253,
      "theoretical_loss": 3.449856061909685,
      "tokens_seen": 2064252928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037823470411233704,
      "loss": 1.3886,
      "theoretical_loss": 3.4498470616028243,
      "tokens_seen": 2064318464
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782146439317954,
      "loss": 1.3155,
      "theoretical_loss": 3.4498380616616946,
      "tokens_seen": 2064384000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037819458375125376,
      "loss": 1.3096,
      "theoretical_loss": 3.449829062086269,
      "tokens_seen": 2064449536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781745235707121,
      "loss": 1.3848,
      "theoretical_loss": 3.4498200628765217,
      "tokens_seen": 2064515072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781544633901705,
      "loss": 1.2935,
      "theoretical_loss": 3.449811064032426,
      "tokens_seen": 2064580608
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781344032096289,
      "loss": 1.3963,
      "theoretical_loss": 3.449802065553955,
      "tokens_seen": 2064646144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037811434302908726,
      "loss": 1.4688,
      "theoretical_loss": 3.4497930674410826,
      "tokens_seen": 2064711680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003780942828485456,
      "loss": 1.3633,
      "theoretical_loss": 3.4497840696937825,
      "tokens_seen": 2064777216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378074222668004,
      "loss": 1.4294,
      "theoretical_loss": 3.449775072312028,
      "tokens_seen": 2064842752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003780541624874624,
      "loss": 1.4046,
      "theoretical_loss": 3.449766075295793,
      "tokens_seen": 2064908288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037803410230692077,
      "loss": 1.3639,
      "theoretical_loss": 3.449757078645051,
      "tokens_seen": 2064973824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037801404212637913,
      "loss": 1.3637,
      "theoretical_loss": 3.4497480823597746,
      "tokens_seen": 2065039360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003779939819458375,
      "loss": 1.3935,
      "theoretical_loss": 3.449739086439939,
      "tokens_seen": 2065104896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037797392176529585,
      "loss": 1.4495,
      "theoretical_loss": 3.449730090885516,
      "tokens_seen": 2065170432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037795386158475427,
      "loss": 1.3631,
      "theoretical_loss": 3.4497210956964803,
      "tokens_seen": 2065235968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037793380140421263,
      "loss": 1.4188,
      "theoretical_loss": 3.4497121008728056,
      "tokens_seen": 2065301504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377913741223671,
      "loss": 1.2886,
      "theoretical_loss": 3.4497031064144648,
      "tokens_seen": 2065367040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037789368104312936,
      "loss": 1.3587,
      "theoretical_loss": 3.4496941123214313,
      "tokens_seen": 2065432576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778736208625878,
      "loss": 1.2964,
      "theoretical_loss": 3.44968511859368,
      "tokens_seen": 2065498112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037785356068204614,
      "loss": 1.2949,
      "theoretical_loss": 3.449676125231183,
      "tokens_seen": 2065563648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778335005015045,
      "loss": 1.3425,
      "theoretical_loss": 3.4496671322339143,
      "tokens_seen": 2065629184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037781344032096286,
      "loss": 1.3075,
      "theoretical_loss": 3.4496581396018478,
      "tokens_seen": 2065694720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777933801404213,
      "loss": 1.3222,
      "theoretical_loss": 3.4496491473349566,
      "tokens_seen": 2065760256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037777331995987964,
      "loss": 1.5782,
      "theoretical_loss": 3.4496401554332152,
      "tokens_seen": 2065825792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377753259779338,
      "loss": 1.31,
      "theoretical_loss": 3.449631163896596,
      "tokens_seen": 2065891328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037773319959879636,
      "loss": 1.245,
      "theoretical_loss": 3.449622172725074,
      "tokens_seen": 2065956864
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777131394182547,
      "loss": 1.3825,
      "theoretical_loss": 3.4496131819186213,
      "tokens_seen": 2066022400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037769307923771314,
      "loss": 1.3257,
      "theoretical_loss": 3.449604191477212,
      "tokens_seen": 2066087936
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 753515,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3310260772705078,
      "objective/train/theoretical_loss": 3.4495952014008204,
      "objective/train/tokens_used": 436612576,
      "theoretical_loss": 3.4495952014008204,
      "tokens_seen": 2066153472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776730190571715,
      "loss": 1.2863,
      "theoretical_loss": 3.4495952014008204,
      "tokens_seen": 2066153472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037765295887662987,
      "loss": 1.2429,
      "theoretical_loss": 3.44958621168942,
      "tokens_seen": 2066219008
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037763289869608823,
      "loss": 1.3256,
      "theoretical_loss": 3.449577222342983,
      "tokens_seen": 2066284544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776128385155467,
      "loss": 1.3724,
      "theoretical_loss": 3.449568233361484,
      "tokens_seen": 2066350080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037759277833500506,
      "loss": 1.3123,
      "theoretical_loss": 3.449559244744897,
      "tokens_seen": 2066415616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775727181544634,
      "loss": 1.3765,
      "theoretical_loss": 3.449550256493195,
      "tokens_seen": 2066481152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775526579739218,
      "loss": 1.4511,
      "theoretical_loss": 3.4495412686063522,
      "tokens_seen": 2066546688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037753259779338015,
      "loss": 1.3386,
      "theoretical_loss": 3.449532281084342,
      "tokens_seen": 2066612224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037751253761283857,
      "loss": 1.3284,
      "theoretical_loss": 3.449523293927137,
      "tokens_seen": 2066677760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037749247743229693,
      "loss": 1.3404,
      "theoretical_loss": 3.4495143071347125,
      "tokens_seen": 2066743296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003774724172517553,
      "loss": 1.2295,
      "theoretical_loss": 3.449505320707041,
      "tokens_seen": 2066808832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037745235707121365,
      "loss": 1.305,
      "theoretical_loss": 3.4494963346440968,
      "tokens_seen": 2066874368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037743229689067207,
      "loss": 1.3815,
      "theoretical_loss": 3.449487348945853,
      "tokens_seen": 2066939904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037741223671013043,
      "loss": 1.462,
      "theoretical_loss": 3.449478363612283,
      "tokens_seen": 2067005440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773921765295888,
      "loss": 1.3252,
      "theoretical_loss": 3.4494693786433612,
      "tokens_seen": 2067070976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037737211634904715,
      "loss": 1.3532,
      "theoretical_loss": 3.449460394039061,
      "tokens_seen": 2067136512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773520561685055,
      "loss": 1.3121,
      "theoretical_loss": 3.449451409799356,
      "tokens_seen": 2067202048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037733199598796393,
      "loss": 1.2725,
      "theoretical_loss": 3.4494424259242193,
      "tokens_seen": 2067267584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773119358074223,
      "loss": 1.3043,
      "theoretical_loss": 3.4494334424136257,
      "tokens_seen": 2067333120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037729187562688066,
      "loss": 1.2655,
      "theoretical_loss": 3.4494244592675476,
      "tokens_seen": 2067398656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377271815446339,
      "loss": 1.3997,
      "theoretical_loss": 3.4494154764859593,
      "tokens_seen": 2067464192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037725175526579744,
      "loss": 1.3476,
      "theoretical_loss": 3.449406494068835,
      "tokens_seen": 2067529728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003772316950852558,
      "loss": 1.4015,
      "theoretical_loss": 3.449397512016147,
      "tokens_seen": 2067595264
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037721163490471416,
      "loss": 1.37,
      "theoretical_loss": 3.4493885303278704,
      "tokens_seen": 2067660800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771915747241725,
      "loss": 1.3407,
      "theoretical_loss": 3.449379549003978,
      "tokens_seen": 2067726336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771715145436309,
      "loss": 1.3115,
      "theoretical_loss": 3.4493705680444435,
      "tokens_seen": 2067791872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771514543630893,
      "loss": 1.332,
      "theoretical_loss": 3.449361587449241,
      "tokens_seen": 2067857408
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037713139418254766,
      "loss": 1.2769,
      "theoretical_loss": 3.4493526072183434,
      "tokens_seen": 2067922944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377111334002006,
      "loss": 1.4331,
      "theoretical_loss": 3.4493436273517255,
      "tokens_seen": 2067988480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770912738214644,
      "loss": 1.368,
      "theoretical_loss": 3.4493346478493603,
      "tokens_seen": 2068054016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770712136409228,
      "loss": 1.3091,
      "theoretical_loss": 3.4493256687112215,
      "tokens_seen": 2068119552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037705115346038117,
      "loss": 1.3052,
      "theoretical_loss": 3.449316689937282,
      "tokens_seen": 2068185088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037703109327983953,
      "loss": 1.3963,
      "theoretical_loss": 3.4493077115275175,
      "tokens_seen": 2068250624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770110330992979,
      "loss": 1.4061,
      "theoretical_loss": 3.4492987334819,
      "tokens_seen": 2068316160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037699097291875625,
      "loss": 1.3042,
      "theoretical_loss": 3.449289755800404,
      "tokens_seen": 2068381696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037697091273821467,
      "loss": 1.4105,
      "theoretical_loss": 3.4492807784830024,
      "tokens_seen": 2068447232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037695085255767303,
      "loss": 1.278,
      "theoretical_loss": 3.4492718015296697,
      "tokens_seen": 2068512768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003769307923771314,
      "loss": 1.4581,
      "theoretical_loss": 3.449262824940379,
      "tokens_seen": 2068578304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037691073219658976,
      "loss": 1.3948,
      "theoretical_loss": 3.449253848715105,
      "tokens_seen": 2068643840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768906720160482,
      "loss": 1.3494,
      "theoretical_loss": 3.4492448728538196,
      "tokens_seen": 2068709376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037687061183550654,
      "loss": 1.3224,
      "theoretical_loss": 3.4492358973564987,
      "tokens_seen": 2068774912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768505516549649,
      "loss": 1.4357,
      "theoretical_loss": 3.4492269222231147,
      "tokens_seen": 2068840448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037683049147442326,
      "loss": 1.2935,
      "theoretical_loss": 3.4492179474536413,
      "tokens_seen": 2068905984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768104312938817,
      "loss": 1.3937,
      "theoretical_loss": 3.4492089730480524,
      "tokens_seen": 2068971520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037679037111334004,
      "loss": 1.3955,
      "theoretical_loss": 3.4491999990063222,
      "tokens_seen": 2069037056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003767703109327984,
      "loss": 1.4024,
      "theoretical_loss": 3.4491910253284237,
      "tokens_seen": 2069102592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037675025075225676,
      "loss": 1.3335,
      "theoretical_loss": 3.449182052014331,
      "tokens_seen": 2069168128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003767301905717151,
      "loss": 1.293,
      "theoretical_loss": 3.4491730790640176,
      "tokens_seen": 2069233664
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037671013039117354,
      "loss": 1.3413,
      "theoretical_loss": 3.4491641064774576,
      "tokens_seen": 2069299200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766900702106319,
      "loss": 1.4542,
      "theoretical_loss": 3.449155134254625,
      "tokens_seen": 2069364736
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 754612,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3888747692108154,
      "objective/train/theoretical_loss": 3.4491461623954924,
      "objective/train/tokens_used": 439889376,
      "theoretical_loss": 3.4491461623954924,
      "tokens_seen": 2069430272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037667001003009027,
      "loss": 1.3643,
      "theoretical_loss": 3.4491461623954924,
      "tokens_seen": 2069430272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037664994984954863,
      "loss": 1.2622,
      "theoretical_loss": 3.449137190900035,
      "tokens_seen": 2069495808
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037662988966900705,
      "loss": 1.3799,
      "theoretical_loss": 3.449128219768225,
      "tokens_seen": 2069561344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766098294884654,
      "loss": 1.3806,
      "theoretical_loss": 3.449119249000037,
      "tokens_seen": 2069626880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037658976930792377,
      "loss": 1.4275,
      "theoretical_loss": 3.4491102785954446,
      "tokens_seen": 2069692416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037656970912738213,
      "loss": 1.3538,
      "theoretical_loss": 3.4491013085544218,
      "tokens_seen": 2069757952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765496489468405,
      "loss": 1.3409,
      "theoretical_loss": 3.449092338876942,
      "tokens_seen": 2069823488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765295887662989,
      "loss": 1.4112,
      "theoretical_loss": 3.4490833695629792,
      "tokens_seen": 2069889024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765095285857573,
      "loss": 1.3635,
      "theoretical_loss": 3.449074400612507,
      "tokens_seen": 2069954560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037648946840521564,
      "loss": 1.3504,
      "theoretical_loss": 3.449065432025499,
      "tokens_seen": 2070020096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000376469408224674,
      "loss": 1.3353,
      "theoretical_loss": 3.44905646380193,
      "tokens_seen": 2070085632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003764493480441324,
      "loss": 1.3992,
      "theoretical_loss": 3.4490474959417723,
      "tokens_seen": 2070151168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003764292878635908,
      "loss": 1.3999,
      "theoretical_loss": 3.4490385284450005,
      "tokens_seen": 2070216704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037640922768304914,
      "loss": 1.287,
      "theoretical_loss": 3.4490295613115878,
      "tokens_seen": 2070282240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763891675025075,
      "loss": 1.4165,
      "theoretical_loss": 3.449020594541509,
      "tokens_seen": 2070347776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037636910732196586,
      "loss": 1.3939,
      "theoretical_loss": 3.4490116281347367,
      "tokens_seen": 2070413312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763490471414243,
      "loss": 1.352,
      "theoretical_loss": 3.449002662091246,
      "tokens_seen": 2070478848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037632898696088264,
      "loss": 1.3397,
      "theoretical_loss": 3.4489936964110095,
      "tokens_seen": 2070544384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000376308926780341,
      "loss": 1.3467,
      "theoretical_loss": 3.448984731094001,
      "tokens_seen": 2070609920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037628886659979937,
      "loss": 1.4159,
      "theoretical_loss": 3.448975766140195,
      "tokens_seen": 2070675456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003762688064192578,
      "loss": 1.3443,
      "theoretical_loss": 3.448966801549565,
      "tokens_seen": 2070740992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037624874623871615,
      "loss": 1.4185,
      "theoretical_loss": 3.448957837322085,
      "tokens_seen": 2070806528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003762286860581745,
      "loss": 1.2558,
      "theoretical_loss": 3.448948873457728,
      "tokens_seen": 2070872064
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037620862587763287,
      "loss": 1.376,
      "theoretical_loss": 3.4489399099564695,
      "tokens_seen": 2070937600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037618856569709123,
      "loss": 1.2335,
      "theoretical_loss": 3.4489309468182814,
      "tokens_seen": 2071003136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037616850551654965,
      "loss": 1.345,
      "theoretical_loss": 3.448921984043138,
      "tokens_seen": 2071068672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000376148445336008,
      "loss": 1.4028,
      "theoretical_loss": 3.4489130216310135,
      "tokens_seen": 2071134208
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037612838515546637,
      "loss": 1.4193,
      "theoretical_loss": 3.4489040595818823,
      "tokens_seen": 2071199744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003761083249749248,
      "loss": 1.4051,
      "theoretical_loss": 3.4488950978957167,
      "tokens_seen": 2071265280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003760882647943832,
      "loss": 1.4004,
      "theoretical_loss": 3.448886136572492,
      "tokens_seen": 2071330816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037606820461384157,
      "loss": 1.4277,
      "theoretical_loss": 3.448877175612181,
      "tokens_seen": 2071396352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037604814443329993,
      "loss": 1.3254,
      "theoretical_loss": 3.448868215014758,
      "tokens_seen": 2071461888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003760280842527583,
      "loss": 1.251,
      "theoretical_loss": 3.4488592547801966,
      "tokens_seen": 2071527424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037600802407221665,
      "loss": 1.4546,
      "theoretical_loss": 3.4488502949084703,
      "tokens_seen": 2071592960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037598796389167507,
      "loss": 1.3047,
      "theoretical_loss": 3.448841335399554,
      "tokens_seen": 2071658496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037596790371113343,
      "loss": 1.3826,
      "theoretical_loss": 3.4488323762534208,
      "tokens_seen": 2071724032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003759478435305918,
      "loss": 1.2144,
      "theoretical_loss": 3.4488234174700443,
      "tokens_seen": 2071789568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037592778335005016,
      "loss": 1.3958,
      "theoretical_loss": 3.448814459049399,
      "tokens_seen": 2071855104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003759077231695086,
      "loss": 1.2211,
      "theoretical_loss": 3.448805500991458,
      "tokens_seen": 2071920640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037588766298896694,
      "loss": 1.4389,
      "theoretical_loss": 3.4487965432961962,
      "tokens_seen": 2071986176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758676028084253,
      "loss": 1.2737,
      "theoretical_loss": 3.4487875859635864,
      "tokens_seen": 2072051712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037584754262788366,
      "loss": 1.2849,
      "theoretical_loss": 3.4487786289936024,
      "tokens_seen": 2072117248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758274824473421,
      "loss": 1.3343,
      "theoretical_loss": 3.448769672386219,
      "tokens_seen": 2072182784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037580742226680044,
      "loss": 1.3565,
      "theoretical_loss": 3.4487607161414093,
      "tokens_seen": 2072248320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757873620862588,
      "loss": 1.2022,
      "theoretical_loss": 3.4487517602591478,
      "tokens_seen": 2072313856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037576730190571716,
      "loss": 1.364,
      "theoretical_loss": 3.4487428047394073,
      "tokens_seen": 2072379392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757472417251755,
      "loss": 1.4593,
      "theoretical_loss": 3.4487338495821627,
      "tokens_seen": 2072444928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037572718154463394,
      "loss": 1.4096,
      "theoretical_loss": 3.4487248947873876,
      "tokens_seen": 2072510464
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757071213640923,
      "loss": 1.3664,
      "theoretical_loss": 3.4487159403550556,
      "tokens_seen": 2072576000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037568706118355067,
      "loss": 1.215,
      "theoretical_loss": 3.4487069862851407,
      "tokens_seen": 2072641536
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 755744,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3763511180877686,
      "objective/train/theoretical_loss": 3.448698032577617,
      "objective/train/tokens_used": 443166176,
      "theoretical_loss": 3.448698032577617,
      "tokens_seen": 2072707072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037566700100300903,
      "loss": 1.3839,
      "theoretical_loss": 3.448698032577617,
      "tokens_seen": 2072707072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037564694082246745,
      "loss": 1.3018,
      "theoretical_loss": 3.448689079232458,
      "tokens_seen": 2072772608
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003756268806419258,
      "loss": 1.3844,
      "theoretical_loss": 3.448680126249638,
      "tokens_seen": 2072838144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037560682046138417,
      "loss": 1.4293,
      "theoretical_loss": 3.44867117362913,
      "tokens_seen": 2072903680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037558676028084253,
      "loss": 1.4653,
      "theoretical_loss": 3.448662221370909,
      "tokens_seen": 2072969216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755667001003009,
      "loss": 1.3939,
      "theoretical_loss": 3.4486532694749483,
      "tokens_seen": 2073034752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755466399197593,
      "loss": 1.3372,
      "theoretical_loss": 3.4486443179412216,
      "tokens_seen": 2073100288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755265797392177,
      "loss": 1.2907,
      "theoretical_loss": 3.4486353667697034,
      "tokens_seen": 2073165824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037550651955867604,
      "loss": 1.4508,
      "theoretical_loss": 3.4486264159603675,
      "tokens_seen": 2073231360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754864593781344,
      "loss": 1.418,
      "theoretical_loss": 3.448617465513187,
      "tokens_seen": 2073296896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754663991975928,
      "loss": 1.3827,
      "theoretical_loss": 3.4486085154281367,
      "tokens_seen": 2073362432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754463390170512,
      "loss": 1.3121,
      "theoretical_loss": 3.4485995657051904,
      "tokens_seen": 2073427968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037542627883650954,
      "loss": 1.2852,
      "theoretical_loss": 3.4485906163443216,
      "tokens_seen": 2073493504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754062186559679,
      "loss": 1.2771,
      "theoretical_loss": 3.448581667345504,
      "tokens_seen": 2073559040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037538615847542626,
      "loss": 1.3048,
      "theoretical_loss": 3.4485727187087125,
      "tokens_seen": 2073624576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003753660982948847,
      "loss": 1.3124,
      "theoretical_loss": 3.4485637704339203,
      "tokens_seen": 2073690112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037534603811434304,
      "loss": 1.2474,
      "theoretical_loss": 3.4485548225211007,
      "tokens_seen": 2073755648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003753259779338014,
      "loss": 1.4349,
      "theoretical_loss": 3.448545874970229,
      "tokens_seen": 2073821184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037530591775325977,
      "loss": 1.2436,
      "theoretical_loss": 3.4485369277812783,
      "tokens_seen": 2073886720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752858575727182,
      "loss": 1.2962,
      "theoretical_loss": 3.448527980954223,
      "tokens_seen": 2073952256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037526579739217655,
      "loss": 1.3145,
      "theoretical_loss": 3.4485190344890366,
      "tokens_seen": 2074017792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752457372116349,
      "loss": 1.3318,
      "theoretical_loss": 3.448510088385693,
      "tokens_seen": 2074083328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037522567703109327,
      "loss": 1.2096,
      "theoretical_loss": 3.448501142644166,
      "tokens_seen": 2074148864
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037520561685055163,
      "loss": 1.3782,
      "theoretical_loss": 3.4484921972644305,
      "tokens_seen": 2074214400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037518555667001005,
      "loss": 1.3167,
      "theoretical_loss": 3.4484832522464597,
      "tokens_seen": 2074279936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003751654964894684,
      "loss": 1.3409,
      "theoretical_loss": 3.448474307590227,
      "tokens_seen": 2074345472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003751454363089268,
      "loss": 1.4329,
      "theoretical_loss": 3.448465363295707,
      "tokens_seen": 2074411008
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037512537612838514,
      "loss": 1.4182,
      "theoretical_loss": 3.4484564193628744,
      "tokens_seen": 2074476544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037510531594784355,
      "loss": 1.3918,
      "theoretical_loss": 3.4484474757917014,
      "tokens_seen": 2074542080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750852557673019,
      "loss": 1.4584,
      "theoretical_loss": 3.4484385325821636,
      "tokens_seen": 2074607616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750651955867603,
      "loss": 1.3237,
      "theoretical_loss": 3.4484295897342334,
      "tokens_seen": 2074673152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037504513540621864,
      "loss": 1.298,
      "theoretical_loss": 3.448420647247886,
      "tokens_seen": 2074738688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000375025075225677,
      "loss": 1.217,
      "theoretical_loss": 3.448411705123095,
      "tokens_seen": 2074804224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750050150451354,
      "loss": 1.4113,
      "theoretical_loss": 3.4484027633598346,
      "tokens_seen": 2074869760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749849548645938,
      "loss": 1.3677,
      "theoretical_loss": 3.4483938219580783,
      "tokens_seen": 2074935296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037496489468405214,
      "loss": 1.2873,
      "theoretical_loss": 3.4483848809178,
      "tokens_seen": 2075000832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749448345035105,
      "loss": 1.2899,
      "theoretical_loss": 3.448375940238974,
      "tokens_seen": 2075066368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749247743229689,
      "loss": 1.2029,
      "theoretical_loss": 3.448366999921575,
      "tokens_seen": 2075131904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749047141424273,
      "loss": 1.388,
      "theoretical_loss": 3.448358059965575,
      "tokens_seen": 2075197440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037488465396188564,
      "loss": 1.4259,
      "theoretical_loss": 3.44834912037095,
      "tokens_seen": 2075262976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374864593781344,
      "loss": 1.3073,
      "theoretical_loss": 3.4483401811376724,
      "tokens_seen": 2075328512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037484453360080237,
      "loss": 1.3622,
      "theoretical_loss": 3.4483312422657173,
      "tokens_seen": 2075394048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003748244734202608,
      "loss": 1.2799,
      "theoretical_loss": 3.4483223037550585,
      "tokens_seen": 2075459584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037480441323971915,
      "loss": 1.2492,
      "theoretical_loss": 3.4483133656056695,
      "tokens_seen": 2075525120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003747843530591775,
      "loss": 1.4148,
      "theoretical_loss": 3.4483044278175248,
      "tokens_seen": 2075590656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037476429287863587,
      "loss": 1.2438,
      "theoretical_loss": 3.448295490390598,
      "tokens_seen": 2075656192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003747442326980943,
      "loss": 1.4106,
      "theoretical_loss": 3.448286553324863,
      "tokens_seen": 2075721728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037472417251755265,
      "loss": 1.3273,
      "theoretical_loss": 3.4482776166202944,
      "tokens_seen": 2075787264
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374704112337011,
      "loss": 1.3045,
      "theoretical_loss": 3.4482686802768656,
      "tokens_seen": 2075852800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003746840521564694,
      "loss": 1.3206,
      "theoretical_loss": 3.4482597442945515,
      "tokens_seen": 2075918336
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 756913,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1429805755615234,
      "objective/train/theoretical_loss": 3.4482508086733246,
      "objective/train/tokens_used": 446442976,
      "theoretical_loss": 3.4482508086733246,
      "tokens_seen": 2075983872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037466399197592774,
      "loss": 1.2074,
      "theoretical_loss": 3.4482508086733246,
      "tokens_seen": 2075983872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037464393179538615,
      "loss": 1.4924,
      "theoretical_loss": 3.448241873413161,
      "tokens_seen": 2076049408
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037462387161484457,
      "loss": 1.2696,
      "theoretical_loss": 3.4482329385140327,
      "tokens_seen": 2076114944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037460381143430293,
      "loss": 1.3054,
      "theoretical_loss": 3.4482240039759144,
      "tokens_seen": 2076180480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745837512537613,
      "loss": 1.3742,
      "theoretical_loss": 3.4482150697987803,
      "tokens_seen": 2076246016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745636910732197,
      "loss": 1.3688,
      "theoretical_loss": 3.4482061359826046,
      "tokens_seen": 2076311552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745436308926781,
      "loss": 1.4298,
      "theoretical_loss": 3.448197202527361,
      "tokens_seen": 2076377088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037452357071213644,
      "loss": 1.3149,
      "theoretical_loss": 3.4481882694330235,
      "tokens_seen": 2076442624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745035105315948,
      "loss": 1.2994,
      "theoretical_loss": 3.4481793366995666,
      "tokens_seen": 2076508160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744834503510532,
      "loss": 1.3241,
      "theoretical_loss": 3.448170404326964,
      "tokens_seen": 2076573696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744633901705116,
      "loss": 1.2989,
      "theoretical_loss": 3.4481614723151894,
      "tokens_seen": 2076639232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037444332998996994,
      "loss": 1.3164,
      "theoretical_loss": 3.448152540664217,
      "tokens_seen": 2076704768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744232698094283,
      "loss": 1.4324,
      "theoretical_loss": 3.4481436093740214,
      "tokens_seen": 2076770304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037440320962888666,
      "loss": 1.2686,
      "theoretical_loss": 3.448134678444576,
      "tokens_seen": 2076835840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743831494483451,
      "loss": 1.3958,
      "theoretical_loss": 3.4481257478758556,
      "tokens_seen": 2076901376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037436308926780344,
      "loss": 1.3549,
      "theoretical_loss": 3.4481168176678327,
      "tokens_seen": 2076966912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743430290872618,
      "loss": 1.3391,
      "theoretical_loss": 3.448107887820483,
      "tokens_seen": 2077032448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037432296890672017,
      "loss": 1.3458,
      "theoretical_loss": 3.44809895833378,
      "tokens_seen": 2077097984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743029087261786,
      "loss": 1.2878,
      "theoretical_loss": 3.4480900292076977,
      "tokens_seen": 2077163520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037428284854563695,
      "loss": 1.2474,
      "theoretical_loss": 3.4480811004422103,
      "tokens_seen": 2077229056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742627883650953,
      "loss": 1.3133,
      "theoretical_loss": 3.4480721720372913,
      "tokens_seen": 2077294592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037424272818455367,
      "loss": 1.2687,
      "theoretical_loss": 3.448063243992915,
      "tokens_seen": 2077360128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037422266800401203,
      "loss": 1.4019,
      "theoretical_loss": 3.4480543163090562,
      "tokens_seen": 2077425664
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037420260782347045,
      "loss": 1.2623,
      "theoretical_loss": 3.4480453889856886,
      "tokens_seen": 2077491200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741825476429288,
      "loss": 1.3361,
      "theoretical_loss": 3.4480364620227855,
      "tokens_seen": 2077556736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741624874623872,
      "loss": 1.3704,
      "theoretical_loss": 3.448027535420322,
      "tokens_seen": 2077622272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037414242728184554,
      "loss": 1.3427,
      "theoretical_loss": 3.448018609178271,
      "tokens_seen": 2077687808
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037412236710130395,
      "loss": 1.315,
      "theoretical_loss": 3.448009683296608,
      "tokens_seen": 2077753344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741023069207623,
      "loss": 1.2976,
      "theoretical_loss": 3.448000757775306,
      "tokens_seen": 2077818880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740822467402207,
      "loss": 1.2403,
      "theoretical_loss": 3.4479918326143397,
      "tokens_seen": 2077884416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037406218655967904,
      "loss": 1.3891,
      "theoretical_loss": 3.447982907813683,
      "tokens_seen": 2077949952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740421263791374,
      "loss": 1.3874,
      "theoretical_loss": 3.44797398337331,
      "tokens_seen": 2078015488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740220661985958,
      "loss": 1.3062,
      "theoretical_loss": 3.4479650592931947,
      "tokens_seen": 2078081024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740020060180542,
      "loss": 1.4022,
      "theoretical_loss": 3.4479561355733113,
      "tokens_seen": 2078146560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037398194583751254,
      "loss": 1.3615,
      "theoretical_loss": 3.447947212213634,
      "tokens_seen": 2078212096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739618856569709,
      "loss": 1.3038,
      "theoretical_loss": 3.447938289214137,
      "tokens_seen": 2078277632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739418254764293,
      "loss": 1.3652,
      "theoretical_loss": 3.447929366574793,
      "tokens_seen": 2078343168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739217652958877,
      "loss": 1.4028,
      "theoretical_loss": 3.447920444295578,
      "tokens_seen": 2078408704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037390170511534605,
      "loss": 1.2863,
      "theoretical_loss": 3.4479115223764656,
      "tokens_seen": 2078474240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738816449348044,
      "loss": 1.3141,
      "theoretical_loss": 3.4479026008174296,
      "tokens_seen": 2078539776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037386158475426277,
      "loss": 1.3018,
      "theoretical_loss": 3.447893679618444,
      "tokens_seen": 2078605312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738415245737212,
      "loss": 1.3015,
      "theoretical_loss": 3.447884758779483,
      "tokens_seen": 2078670848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037382146439317955,
      "loss": 1.3282,
      "theoretical_loss": 3.4478758383005212,
      "tokens_seen": 2078736384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738014042126379,
      "loss": 1.2612,
      "theoretical_loss": 3.447866918181532,
      "tokens_seen": 2078801920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037378134403209627,
      "loss": 1.337,
      "theoretical_loss": 3.4478579984224904,
      "tokens_seen": 2078867456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737612838515547,
      "loss": 1.3796,
      "theoretical_loss": 3.44784907902337,
      "tokens_seen": 2078932992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037374122367101305,
      "loss": 1.3387,
      "theoretical_loss": 3.4478401599841444,
      "tokens_seen": 2078998528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737211634904714,
      "loss": 1.3103,
      "theoretical_loss": 3.447831241304789,
      "tokens_seen": 2079064064
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737011033099298,
      "loss": 1.3103,
      "theoretical_loss": 3.4478223229852767,
      "tokens_seen": 2079129600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037368104312938814,
      "loss": 1.4519,
      "theoretical_loss": 3.447813405025582,
      "tokens_seen": 2079195136
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 758165,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.335489273071289,
      "objective/train/theoretical_loss": 3.44780448742568,
      "objective/train/tokens_used": 449719776,
      "theoretical_loss": 3.44780448742568,
      "tokens_seen": 2079260672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037366098294884655,
      "loss": 1.3579,
      "theoretical_loss": 3.44780448742568,
      "tokens_seen": 2079260672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003736409227683049,
      "loss": 1.2901,
      "theoretical_loss": 3.4477955701855434,
      "tokens_seen": 2079326208
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003736208625877633,
      "loss": 1.4038,
      "theoretical_loss": 3.447786653305147,
      "tokens_seen": 2079391744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037360080240722164,
      "loss": 1.2924,
      "theoretical_loss": 3.4477777367844653,
      "tokens_seen": 2079457280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037358074222668006,
      "loss": 1.4341,
      "theoretical_loss": 3.4477688206234722,
      "tokens_seen": 2079522816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735606820461384,
      "loss": 1.4142,
      "theoretical_loss": 3.4477599048221412,
      "tokens_seen": 2079588352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735406218655968,
      "loss": 1.3315,
      "theoretical_loss": 3.4477509893804474,
      "tokens_seen": 2079653888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037352056168505514,
      "loss": 1.4694,
      "theoretical_loss": 3.4477420742983647,
      "tokens_seen": 2079719424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037350050150451356,
      "loss": 1.2809,
      "theoretical_loss": 3.4477331595758667,
      "tokens_seen": 2079784960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734804413239719,
      "loss": 1.3458,
      "theoretical_loss": 3.447724245212928,
      "tokens_seen": 2079850496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734603811434303,
      "loss": 1.427,
      "theoretical_loss": 3.4477153312095234,
      "tokens_seen": 2079916032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037344032096288865,
      "loss": 1.3898,
      "theoretical_loss": 3.447706417565626,
      "tokens_seen": 2079981568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373420260782347,
      "loss": 1.3154,
      "theoretical_loss": 3.4476975042812104,
      "tokens_seen": 2080047104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734002006018054,
      "loss": 1.3298,
      "theoretical_loss": 3.4476885913562514,
      "tokens_seen": 2080112640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733801404212638,
      "loss": 1.3451,
      "theoretical_loss": 3.447679678790722,
      "tokens_seen": 2080178176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037336008024072215,
      "loss": 1.3329,
      "theoretical_loss": 3.4476707665845967,
      "tokens_seen": 2080243712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733400200601805,
      "loss": 1.3843,
      "theoretical_loss": 3.4476618547378504,
      "tokens_seen": 2080309248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037331995987963893,
      "loss": 1.3055,
      "theoretical_loss": 3.447652943250457,
      "tokens_seen": 2080374784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732998996990973,
      "loss": 1.4442,
      "theoretical_loss": 3.44764403212239,
      "tokens_seen": 2080440320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037327983951855565,
      "loss": 1.3153,
      "theoretical_loss": 3.4476351213536245,
      "tokens_seen": 2080505856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373259779338014,
      "loss": 1.2781,
      "theoretical_loss": 3.447626210944134,
      "tokens_seen": 2080571392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732397191574724,
      "loss": 1.4366,
      "theoretical_loss": 3.447617300893893,
      "tokens_seen": 2080636928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732196589769308,
      "loss": 1.337,
      "theoretical_loss": 3.447608391202876,
      "tokens_seen": 2080702464
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037319959879638916,
      "loss": 1.3664,
      "theoretical_loss": 3.447599481871057,
      "tokens_seen": 2080768000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731795386158475,
      "loss": 1.2406,
      "theoretical_loss": 3.44759057289841,
      "tokens_seen": 2080833536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731594784353059,
      "loss": 1.2714,
      "theoretical_loss": 3.4475816642849093,
      "tokens_seen": 2080899072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731394182547643,
      "loss": 1.3488,
      "theoretical_loss": 3.447572756030529,
      "tokens_seen": 2080964608
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731193580742227,
      "loss": 1.3154,
      "theoretical_loss": 3.4475638481352435,
      "tokens_seen": 2081030144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730992978936811,
      "loss": 1.2924,
      "theoretical_loss": 3.4475549405990273,
      "tokens_seen": 2081095680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037307923771313944,
      "loss": 1.4311,
      "theoretical_loss": 3.4475460334218537,
      "tokens_seen": 2081161216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730591775325978,
      "loss": 1.4018,
      "theoretical_loss": 3.447537126603698,
      "tokens_seen": 2081226752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730391173520562,
      "loss": 1.3416,
      "theoretical_loss": 3.447528220144534,
      "tokens_seen": 2081292288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730190571715146,
      "loss": 1.3829,
      "theoretical_loss": 3.447519314044335,
      "tokens_seen": 2081357824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037299899699097294,
      "loss": 1.3314,
      "theoretical_loss": 3.447510408303077,
      "tokens_seen": 2081423360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729789368104313,
      "loss": 1.4558,
      "theoretical_loss": 3.4475015029207334,
      "tokens_seen": 2081488896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729588766298897,
      "loss": 1.3249,
      "theoretical_loss": 3.447492597897278,
      "tokens_seen": 2081554432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729388164493481,
      "loss": 1.3054,
      "theoretical_loss": 3.4474836932326856,
      "tokens_seen": 2081619968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037291875626880645,
      "loss": 1.1993,
      "theoretical_loss": 3.4474747889269297,
      "tokens_seen": 2081685504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728986960882648,
      "loss": 1.3733,
      "theoretical_loss": 3.4474658849799855,
      "tokens_seen": 2081751040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037287863590772317,
      "loss": 1.2718,
      "theoretical_loss": 3.447456981391827,
      "tokens_seen": 2081816576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728585757271816,
      "loss": 1.39,
      "theoretical_loss": 3.447448078162428,
      "tokens_seen": 2081882112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037283851554663995,
      "loss": 1.3114,
      "theoretical_loss": 3.4474391752917635,
      "tokens_seen": 2081947648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728184553660983,
      "loss": 1.3593,
      "theoretical_loss": 3.4474302727798065,
      "tokens_seen": 2082013184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727983951855567,
      "loss": 1.3305,
      "theoretical_loss": 3.4474213706265324,
      "tokens_seen": 2082078720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727783350050151,
      "loss": 1.4255,
      "theoretical_loss": 3.4474124688319154,
      "tokens_seen": 2082144256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037275827482447345,
      "loss": 1.3578,
      "theoretical_loss": 3.4474035673959293,
      "tokens_seen": 2082209792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727382146439318,
      "loss": 1.3362,
      "theoretical_loss": 3.4473946663185484,
      "tokens_seen": 2082275328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727181544633902,
      "loss": 1.3871,
      "theoretical_loss": 3.4473857655997473,
      "tokens_seen": 2082340864
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037269809428284854,
      "loss": 1.3939,
      "theoretical_loss": 3.4473768652395,
      "tokens_seen": 2082406400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037267803410230696,
      "loss": 1.3154,
      "theoretical_loss": 3.4473679652377807,
      "tokens_seen": 2082471936
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 759184,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.221189022064209,
      "objective/train/theoretical_loss": 3.447359065594564,
      "objective/train/tokens_used": 452996576,
      "theoretical_loss": 3.447359065594564,
      "tokens_seen": 2082537472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726579739217653,
      "loss": 1.2887,
      "theoretical_loss": 3.447359065594564,
      "tokens_seen": 2082537472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726379137412237,
      "loss": 1.4102,
      "theoretical_loss": 3.447350166309824,
      "tokens_seen": 2082603008
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037261785356068204,
      "loss": 1.2748,
      "theoretical_loss": 3.4473412673835346,
      "tokens_seen": 2082668544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037259779338014046,
      "loss": 1.3869,
      "theoretical_loss": 3.4473323688156707,
      "tokens_seen": 2082734080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725777331995988,
      "loss": 1.3816,
      "theoretical_loss": 3.4473234706062064,
      "tokens_seen": 2082799616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725576730190572,
      "loss": 1.2924,
      "theoretical_loss": 3.4473145727551158,
      "tokens_seen": 2082865152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037253761283851554,
      "loss": 1.3734,
      "theoretical_loss": 3.447305675262373,
      "tokens_seen": 2082930688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037251755265797396,
      "loss": 1.2656,
      "theoretical_loss": 3.4472967781279533,
      "tokens_seen": 2082996224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724974924774323,
      "loss": 1.2967,
      "theoretical_loss": 3.44728788135183,
      "tokens_seen": 2083061760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724774322968907,
      "loss": 1.3389,
      "theoretical_loss": 3.4472789849339778,
      "tokens_seen": 2083127296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037245737211634905,
      "loss": 1.4214,
      "theoretical_loss": 3.447270088874371,
      "tokens_seen": 2083192832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724373119358074,
      "loss": 1.4837,
      "theoretical_loss": 3.4472611931729835,
      "tokens_seen": 2083258368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724172517552658,
      "loss": 1.3538,
      "theoretical_loss": 3.4472522978297904,
      "tokens_seen": 2083323904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723971915747242,
      "loss": 1.3755,
      "theoretical_loss": 3.447243402844765,
      "tokens_seen": 2083389440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037237713139418255,
      "loss": 1.3115,
      "theoretical_loss": 3.447234508217883,
      "tokens_seen": 2083454976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723570712136409,
      "loss": 1.3346,
      "theoretical_loss": 3.447225613949117,
      "tokens_seen": 2083520512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037233701103309933,
      "loss": 1.3728,
      "theoretical_loss": 3.4472167200384423,
      "tokens_seen": 2083586048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723169508525577,
      "loss": 1.3903,
      "theoretical_loss": 3.4472078264858332,
      "tokens_seen": 2083651584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037229689067201605,
      "loss": 1.3117,
      "theoretical_loss": 3.447198933291264,
      "tokens_seen": 2083717120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722768304914744,
      "loss": 1.3741,
      "theoretical_loss": 3.447190040454709,
      "tokens_seen": 2083782656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722567703109328,
      "loss": 1.329,
      "theoretical_loss": 3.447181147976142,
      "tokens_seen": 2083848192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722367101303912,
      "loss": 1.3859,
      "theoretical_loss": 3.4471722558555387,
      "tokens_seen": 2083913728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037221664994984956,
      "loss": 1.4165,
      "theoretical_loss": 3.447163364092872,
      "tokens_seen": 2083979264
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721965897693079,
      "loss": 1.296,
      "theoretical_loss": 3.4471544726881165,
      "tokens_seen": 2084044800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721765295887663,
      "loss": 1.2957,
      "theoretical_loss": 3.4471455816412475,
      "tokens_seen": 2084110336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721564694082247,
      "loss": 1.3047,
      "theoretical_loss": 3.447136690952238,
      "tokens_seen": 2084175872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037213640922768306,
      "loss": 1.3856,
      "theoretical_loss": 3.4471278006210637,
      "tokens_seen": 2084241408
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721163490471414,
      "loss": 1.3583,
      "theoretical_loss": 3.447118910647698,
      "tokens_seen": 2084306944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003720962888665998,
      "loss": 1.3197,
      "theoretical_loss": 3.447110021032115,
      "tokens_seen": 2084372480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037207622868605815,
      "loss": 1.3628,
      "theoretical_loss": 3.44710113177429,
      "tokens_seen": 2084438016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037205616850551656,
      "loss": 1.3225,
      "theoretical_loss": 3.447092242874197,
      "tokens_seen": 2084503552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003720361083249749,
      "loss": 1.3943,
      "theoretical_loss": 3.4470833543318102,
      "tokens_seen": 2084569088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003720160481444333,
      "loss": 1.4609,
      "theoretical_loss": 3.4470744661471038,
      "tokens_seen": 2084634624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037199598796389165,
      "loss": 1.3049,
      "theoretical_loss": 3.4470655783200526,
      "tokens_seen": 2084700160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037197592778335007,
      "loss": 1.2727,
      "theoretical_loss": 3.447056690850631,
      "tokens_seen": 2084765696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037195586760280843,
      "loss": 1.3587,
      "theoretical_loss": 3.447047803738812,
      "tokens_seen": 2084831232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003719358074222668,
      "loss": 1.4363,
      "theoretical_loss": 3.447038916984572,
      "tokens_seen": 2084896768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037191574724172515,
      "loss": 1.4422,
      "theoretical_loss": 3.4470300305878845,
      "tokens_seen": 2084962304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718956870611835,
      "loss": 1.2346,
      "theoretical_loss": 3.4470211445487235,
      "tokens_seen": 2085027840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037187562688064193,
      "loss": 1.3941,
      "theoretical_loss": 3.447012258867064,
      "tokens_seen": 2085093376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718555667001003,
      "loss": 1.2277,
      "theoretical_loss": 3.44700337354288,
      "tokens_seen": 2085158912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037183550651955866,
      "loss": 1.4075,
      "theoretical_loss": 3.446994488576146,
      "tokens_seen": 2085224448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000371815446339017,
      "loss": 1.2745,
      "theoretical_loss": 3.4469856039668363,
      "tokens_seen": 2085289984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037179538615847544,
      "loss": 1.2611,
      "theoretical_loss": 3.4469767197149253,
      "tokens_seen": 2085355520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717753259779338,
      "loss": 1.3076,
      "theoretical_loss": 3.446967835820388,
      "tokens_seen": 2085421056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037175526579739216,
      "loss": 1.279,
      "theoretical_loss": 3.4469589522831976,
      "tokens_seen": 2085486592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717352056168505,
      "loss": 1.3766,
      "theoretical_loss": 3.4469500691033295,
      "tokens_seen": 2085552128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717151454363089,
      "loss": 1.4607,
      "theoretical_loss": 3.4469411862807577,
      "tokens_seen": 2085617664
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716950852557673,
      "loss": 1.3921,
      "theoretical_loss": 3.4469323038154567,
      "tokens_seen": 2085683200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037167502507522566,
      "loss": 1.4168,
      "theoretical_loss": 3.4469234217074005,
      "tokens_seen": 2085748736
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 760445,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3558768033981323,
      "objective/train/theoretical_loss": 3.4469145399565644,
      "objective/train/tokens_used": 456273376,
      "theoretical_loss": 3.4469145399565644,
      "tokens_seen": 2085814272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000371654964894684,
      "loss": 1.3861,
      "theoretical_loss": 3.4469145399565644,
      "tokens_seen": 2085814272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716349047141424,
      "loss": 1.2515,
      "theoretical_loss": 3.4469056585629216,
      "tokens_seen": 2085879808
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037161484453360086,
      "loss": 1.3498,
      "theoretical_loss": 3.4468967775264474,
      "tokens_seen": 2085945344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003715947843530592,
      "loss": 1.2705,
      "theoretical_loss": 3.4468878968471164,
      "tokens_seen": 2086010880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003715747241725176,
      "loss": 1.5168,
      "theoretical_loss": 3.4468790165249024,
      "tokens_seen": 2086076416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037155466399197595,
      "loss": 1.3204,
      "theoretical_loss": 3.44687013655978,
      "tokens_seen": 2086141952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037153460381143436,
      "loss": 1.4237,
      "theoretical_loss": 3.4468612569517236,
      "tokens_seen": 2086207488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003715145436308927,
      "loss": 1.3912,
      "theoretical_loss": 3.446852377700708,
      "tokens_seen": 2086273024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714944834503511,
      "loss": 1.4818,
      "theoretical_loss": 3.4468434988067065,
      "tokens_seen": 2086338560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037147442326980945,
      "loss": 1.4445,
      "theoretical_loss": 3.446834620269695,
      "tokens_seen": 2086404096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714543630892678,
      "loss": 1.4011,
      "theoretical_loss": 3.446825742089647,
      "tokens_seen": 2086469632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037143430290872623,
      "loss": 1.2855,
      "theoretical_loss": 3.4468168642665376,
      "tokens_seen": 2086535168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714142427281846,
      "loss": 1.3122,
      "theoretical_loss": 3.4468079868003407,
      "tokens_seen": 2086600704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037139418254764295,
      "loss": 1.3643,
      "theoretical_loss": 3.446799109691031,
      "tokens_seen": 2086666240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003713741223671013,
      "loss": 1.1952,
      "theoretical_loss": 3.446790232938582,
      "tokens_seen": 2086731776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037135406218655973,
      "loss": 1.377,
      "theoretical_loss": 3.44678135654297,
      "tokens_seen": 2086797312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003713340020060181,
      "loss": 1.3972,
      "theoretical_loss": 3.446772480504168,
      "tokens_seen": 2086862848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037131394182547645,
      "loss": 1.3611,
      "theoretical_loss": 3.4467636048221513,
      "tokens_seen": 2086928384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003712938816449348,
      "loss": 1.4277,
      "theoretical_loss": 3.4467547294968934,
      "tokens_seen": 2086993920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003712738214643932,
      "loss": 1.385,
      "theoretical_loss": 3.44674585452837,
      "tokens_seen": 2087059456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003712537612838516,
      "loss": 1.2567,
      "theoretical_loss": 3.446736979916554,
      "tokens_seen": 2087124992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037123370110330996,
      "loss": 1.3637,
      "theoretical_loss": 3.446728105661421,
      "tokens_seen": 2087190528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003712136409227683,
      "loss": 1.262,
      "theoretical_loss": 3.4467192317629456,
      "tokens_seen": 2087256064
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711935807422267,
      "loss": 1.3156,
      "theoretical_loss": 3.446710358221102,
      "tokens_seen": 2087321600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711735205616851,
      "loss": 1.3109,
      "theoretical_loss": 3.446701485035864,
      "tokens_seen": 2087387136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037115346038114346,
      "loss": 1.2819,
      "theoretical_loss": 3.4466926122072064,
      "tokens_seen": 2087452672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711334002006018,
      "loss": 1.3675,
      "theoretical_loss": 3.4466837397351044,
      "tokens_seen": 2087518208
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711133400200602,
      "loss": 1.4508,
      "theoretical_loss": 3.4466748676195316,
      "tokens_seen": 2087583744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037109327983951855,
      "loss": 1.3272,
      "theoretical_loss": 3.4466659958604633,
      "tokens_seen": 2087649280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037107321965897696,
      "loss": 1.3219,
      "theoretical_loss": 3.4466571244578734,
      "tokens_seen": 2087714816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003710531594784353,
      "loss": 1.4201,
      "theoretical_loss": 3.446648253411736,
      "tokens_seen": 2087780352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003710330992978937,
      "loss": 1.3165,
      "theoretical_loss": 3.4466393827220267,
      "tokens_seen": 2087845888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037101303911735205,
      "loss": 1.3381,
      "theoretical_loss": 3.4466305123887193,
      "tokens_seen": 2087911424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037099297893681047,
      "loss": 1.3481,
      "theoretical_loss": 3.446621642411788,
      "tokens_seen": 2087976960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037097291875626883,
      "loss": 1.2969,
      "theoretical_loss": 3.446612772791208,
      "tokens_seen": 2088042496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003709528585757272,
      "loss": 1.3704,
      "theoretical_loss": 3.4466039035269533,
      "tokens_seen": 2088108032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037093279839518555,
      "loss": 1.3183,
      "theoretical_loss": 3.4465950346189986,
      "tokens_seen": 2088173568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003709127382146439,
      "loss": 1.3428,
      "theoretical_loss": 3.446586166067319,
      "tokens_seen": 2088239104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037089267803410233,
      "loss": 1.3269,
      "theoretical_loss": 3.4465772978718876,
      "tokens_seen": 2088304640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708726178535607,
      "loss": 1.3119,
      "theoretical_loss": 3.44656843003268,
      "tokens_seen": 2088370176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037085255767301906,
      "loss": 1.3168,
      "theoretical_loss": 3.4465595625496706,
      "tokens_seen": 2088435712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708324974924774,
      "loss": 1.2566,
      "theoretical_loss": 3.4465506954228338,
      "tokens_seen": 2088501248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037081243731193584,
      "loss": 1.3596,
      "theoretical_loss": 3.4465418286521436,
      "tokens_seen": 2088566784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707923771313942,
      "loss": 1.4234,
      "theoretical_loss": 3.4465329622375753,
      "tokens_seen": 2088632320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037077231695085256,
      "loss": 1.2814,
      "theoretical_loss": 3.446524096179103,
      "tokens_seen": 2088697856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707522567703109,
      "loss": 1.2835,
      "theoretical_loss": 3.4465152304767015,
      "tokens_seen": 2088763392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707321965897693,
      "loss": 1.3687,
      "theoretical_loss": 3.446506365130345,
      "tokens_seen": 2088828928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707121364092277,
      "loss": 1.3366,
      "theoretical_loss": 3.446497500140008,
      "tokens_seen": 2088894464
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037069207622868606,
      "loss": 1.4386,
      "theoretical_loss": 3.4464886355056654,
      "tokens_seen": 2088960000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003706720160481444,
      "loss": 1.3665,
      "theoretical_loss": 3.446479771227292,
      "tokens_seen": 2089025536
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 761616,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3948633670806885,
      "objective/train/theoretical_loss": 3.446470907304861,
      "objective/train/tokens_used": 459550176,
      "theoretical_loss": 3.446470907304861,
      "tokens_seen": 2089091072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003706519558676028,
      "loss": 1.3112,
      "theoretical_loss": 3.446470907304861,
      "tokens_seen": 2089091072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003706318956870612,
      "loss": 1.3293,
      "theoretical_loss": 3.4464620437383484,
      "tokens_seen": 2089156608
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037061183550651957,
      "loss": 1.293,
      "theoretical_loss": 3.446453180527728,
      "tokens_seen": 2089222144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037059177532597793,
      "loss": 1.413,
      "theoretical_loss": 3.4464443176729747,
      "tokens_seen": 2089287680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003705717151454363,
      "loss": 1.3553,
      "theoretical_loss": 3.446435455174063,
      "tokens_seen": 2089353216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037055165496489465,
      "loss": 1.3972,
      "theoretical_loss": 3.446426593030967,
      "tokens_seen": 2089418752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037053159478435307,
      "loss": 1.3368,
      "theoretical_loss": 3.4464177312436615,
      "tokens_seen": 2089484288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037051153460381143,
      "loss": 1.3972,
      "theoretical_loss": 3.446408869812122,
      "tokens_seen": 2089549824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704914744232698,
      "loss": 1.2802,
      "theoretical_loss": 3.446400008736321,
      "tokens_seen": 2089615360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037047141424272816,
      "loss": 1.3065,
      "theoretical_loss": 3.446391148016235,
      "tokens_seen": 2089680896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704513540621866,
      "loss": 1.2517,
      "theoretical_loss": 3.4463822876518377,
      "tokens_seen": 2089746432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037043129388164494,
      "loss": 1.293,
      "theoretical_loss": 3.4463734276431035,
      "tokens_seen": 2089811968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704112337011033,
      "loss": 1.3577,
      "theoretical_loss": 3.446364567990008,
      "tokens_seen": 2089877504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037039117352056166,
      "loss": 1.3469,
      "theoretical_loss": 3.4463557086925243,
      "tokens_seen": 2089943040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703711133400201,
      "loss": 1.3559,
      "theoretical_loss": 3.446346849750628,
      "tokens_seen": 2090008576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037035105315947844,
      "loss": 1.266,
      "theoretical_loss": 3.4463379911642935,
      "tokens_seen": 2090074112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703309929789368,
      "loss": 1.34,
      "theoretical_loss": 3.446329132933495,
      "tokens_seen": 2090139648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037031093279839516,
      "loss": 1.4095,
      "theoretical_loss": 3.4463202750582074,
      "tokens_seen": 2090205184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702908726178535,
      "loss": 1.2532,
      "theoretical_loss": 3.4463114175384053,
      "tokens_seen": 2090270720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037027081243731194,
      "loss": 1.3557,
      "theoretical_loss": 3.446302560374064,
      "tokens_seen": 2090336256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702507522567703,
      "loss": 1.3104,
      "theoretical_loss": 3.446293703565156,
      "tokens_seen": 2090401792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037023069207622867,
      "loss": 1.3731,
      "theoretical_loss": 3.446284847111658,
      "tokens_seen": 2090467328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037021063189568703,
      "loss": 1.3606,
      "theoretical_loss": 3.4462759910135436,
      "tokens_seen": 2090532864
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037019057171514544,
      "loss": 1.1928,
      "theoretical_loss": 3.4462671352707877,
      "tokens_seen": 2090598400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003701705115346038,
      "loss": 1.3777,
      "theoretical_loss": 3.4462582798833648,
      "tokens_seen": 2090663936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037015045135406217,
      "loss": 1.4143,
      "theoretical_loss": 3.4462494248512496,
      "tokens_seen": 2090729472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037013039117352053,
      "loss": 1.3055,
      "theoretical_loss": 3.4462405701744165,
      "tokens_seen": 2090795008
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037011033099297895,
      "loss": 1.3743,
      "theoretical_loss": 3.44623171585284,
      "tokens_seen": 2090860544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037009027081243736,
      "loss": 1.3612,
      "theoretical_loss": 3.4462228618864956,
      "tokens_seen": 2090926080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003700702106318957,
      "loss": 1.4749,
      "theoretical_loss": 3.4462140082753567,
      "tokens_seen": 2090991616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003700501504513541,
      "loss": 1.2325,
      "theoretical_loss": 3.446205155019399,
      "tokens_seen": 2091057152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037003009027081245,
      "loss": 1.3897,
      "theoretical_loss": 3.4461963021185964,
      "tokens_seen": 2091122688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037001003009027087,
      "loss": 1.3854,
      "theoretical_loss": 3.4461874495729234,
      "tokens_seen": 2091188224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036998996990972923,
      "loss": 1.3896,
      "theoretical_loss": 3.4461785973823553,
      "tokens_seen": 2091253760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003699699097291876,
      "loss": 1.3562,
      "theoretical_loss": 3.4461697455468663,
      "tokens_seen": 2091319296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036994984954864595,
      "loss": 1.393,
      "theoretical_loss": 3.4461608940664314,
      "tokens_seen": 2091384832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003699297893681043,
      "loss": 1.347,
      "theoretical_loss": 3.446152042941024,
      "tokens_seen": 2091450368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036990972918756273,
      "loss": 1.5436,
      "theoretical_loss": 3.446143192170621,
      "tokens_seen": 2091515904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698896690070211,
      "loss": 1.363,
      "theoretical_loss": 3.4461343417551946,
      "tokens_seen": 2091581440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036986960882647946,
      "loss": 1.2586,
      "theoretical_loss": 3.446125491694721,
      "tokens_seen": 2091646976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698495486459378,
      "loss": 1.292,
      "theoretical_loss": 3.4461166419891747,
      "tokens_seen": 2091712512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036982948846539624,
      "loss": 1.3086,
      "theoretical_loss": 3.4461077926385295,
      "tokens_seen": 2091778048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698094282848546,
      "loss": 1.4264,
      "theoretical_loss": 3.446098943642761,
      "tokens_seen": 2091843584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036978936810431296,
      "loss": 1.3687,
      "theoretical_loss": 3.446090095001843,
      "tokens_seen": 2091909120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697693079237713,
      "loss": 1.367,
      "theoretical_loss": 3.4460812467157513,
      "tokens_seen": 2091974656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697492477432297,
      "loss": 1.3264,
      "theoretical_loss": 3.4460723987844593,
      "tokens_seen": 2092040192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697291875626881,
      "loss": 1.3454,
      "theoretical_loss": 3.446063551207942,
      "tokens_seen": 2092105728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036970912738214646,
      "loss": 1.3447,
      "theoretical_loss": 3.446054703986175,
      "tokens_seen": 2092171264
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696890672016048,
      "loss": 1.3832,
      "theoretical_loss": 3.446045857119132,
      "tokens_seen": 2092236800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696690070210632,
      "loss": 1.3116,
      "theoretical_loss": 3.446037010606788,
      "tokens_seen": 2092302336
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 762773,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.473339319229126,
      "objective/train/theoretical_loss": 3.446028164449117,
      "objective/train/tokens_used": 462826976,
      "theoretical_loss": 3.446028164449117,
      "tokens_seen": 2092367872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696489468405216,
      "loss": 1.4383,
      "theoretical_loss": 3.446028164449117,
      "tokens_seen": 2092367872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036962888665997997,
      "loss": 1.293,
      "theoretical_loss": 3.446019318646095,
      "tokens_seen": 2092433408
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036960882647943833,
      "loss": 1.3879,
      "theoretical_loss": 3.4460104731976955,
      "tokens_seen": 2092498944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695887662988967,
      "loss": 1.4025,
      "theoretical_loss": 3.4460016281038937,
      "tokens_seen": 2092564480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036956870611835505,
      "loss": 1.2483,
      "theoretical_loss": 3.445992783364664,
      "tokens_seen": 2092630016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036954864593781347,
      "loss": 1.4044,
      "theoretical_loss": 3.445983938979982,
      "tokens_seen": 2092695552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036952858575727183,
      "loss": 1.3442,
      "theoretical_loss": 3.4459750949498207,
      "tokens_seen": 2092761088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695085255767302,
      "loss": 1.4085,
      "theoretical_loss": 3.4459662512741565,
      "tokens_seen": 2092826624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036948846539618856,
      "loss": 1.3335,
      "theoretical_loss": 3.445957407952963,
      "tokens_seen": 2092892160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369468405215647,
      "loss": 1.3896,
      "theoretical_loss": 3.4459485649862147,
      "tokens_seen": 2092957696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036944834503510534,
      "loss": 1.3336,
      "theoretical_loss": 3.4459397223738875,
      "tokens_seen": 2093023232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003694282848545637,
      "loss": 1.3698,
      "theoretical_loss": 3.445930880115955,
      "tokens_seen": 2093088768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036940822467402206,
      "loss": 1.3368,
      "theoretical_loss": 3.4459220382123927,
      "tokens_seen": 2093154304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693881644934805,
      "loss": 1.3133,
      "theoretical_loss": 3.445913196663175,
      "tokens_seen": 2093219840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036936810431293884,
      "loss": 1.316,
      "theoretical_loss": 3.4459043554682762,
      "tokens_seen": 2093285376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693480441323972,
      "loss": 1.4123,
      "theoretical_loss": 3.4458955146276713,
      "tokens_seen": 2093350912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036932798395185556,
      "loss": 1.2945,
      "theoretical_loss": 3.4458866741413354,
      "tokens_seen": 2093416448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693079237713139,
      "loss": 1.3152,
      "theoretical_loss": 3.445877834009243,
      "tokens_seen": 2093481984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036928786359077234,
      "loss": 1.2877,
      "theoretical_loss": 3.4458689942313683,
      "tokens_seen": 2093547520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003692678034102307,
      "loss": 1.3279,
      "theoretical_loss": 3.4458601548076864,
      "tokens_seen": 2093613056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036924774322968907,
      "loss": 1.3856,
      "theoretical_loss": 3.445851315738172,
      "tokens_seen": 2093678592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036922768304914743,
      "loss": 1.2926,
      "theoretical_loss": 3.4458424770227998,
      "tokens_seen": 2093744128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036920762286860585,
      "loss": 1.339,
      "theoretical_loss": 3.445833638661545,
      "tokens_seen": 2093809664
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691875626880642,
      "loss": 1.3602,
      "theoretical_loss": 3.4458248006543815,
      "tokens_seen": 2093875200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036916750250752257,
      "loss": 1.3063,
      "theoretical_loss": 3.4458159630012846,
      "tokens_seen": 2093940736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036914744232698093,
      "loss": 1.2788,
      "theoretical_loss": 3.445807125702229,
      "tokens_seen": 2094006272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691273821464393,
      "loss": 1.2991,
      "theoretical_loss": 3.4457982887571887,
      "tokens_seen": 2094071808
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691073219658977,
      "loss": 1.4432,
      "theoretical_loss": 3.4457894521661396,
      "tokens_seen": 2094137344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690872617853561,
      "loss": 1.4004,
      "theoretical_loss": 3.445780615929056,
      "tokens_seen": 2094202880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036906720160481444,
      "loss": 1.27,
      "theoretical_loss": 3.445771780045912,
      "tokens_seen": 2094268416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690471414242728,
      "loss": 1.2412,
      "theoretical_loss": 3.4457629445166833,
      "tokens_seen": 2094333952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690270812437312,
      "loss": 1.3104,
      "theoretical_loss": 3.445754109341344,
      "tokens_seen": 2094399488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690070210631896,
      "loss": 1.4074,
      "theoretical_loss": 3.4457452745198696,
      "tokens_seen": 2094465024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036898696088264794,
      "loss": 1.3001,
      "theoretical_loss": 3.445736440052234,
      "tokens_seen": 2094530560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003689669007021063,
      "loss": 1.4295,
      "theoretical_loss": 3.445727605938412,
      "tokens_seen": 2094596096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036894684052156466,
      "loss": 1.3814,
      "theoretical_loss": 3.445718772178379,
      "tokens_seen": 2094661632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003689267803410231,
      "loss": 1.2488,
      "theoretical_loss": 3.445709938772109,
      "tokens_seen": 2094727168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036890672016048144,
      "loss": 1.3465,
      "theoretical_loss": 3.4457011057195777,
      "tokens_seen": 2094792704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688866599799398,
      "loss": 1.3506,
      "theoretical_loss": 3.445692273020759,
      "tokens_seen": 2094858240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036886659979939817,
      "loss": 1.3613,
      "theoretical_loss": 3.4456834406756283,
      "tokens_seen": 2094923776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688465396188566,
      "loss": 1.3898,
      "theoretical_loss": 3.44567460868416,
      "tokens_seen": 2094989312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036882647943831494,
      "loss": 1.3197,
      "theoretical_loss": 3.4456657770463286,
      "tokens_seen": 2095054848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688064192577733,
      "loss": 1.4066,
      "theoretical_loss": 3.44565694576211,
      "tokens_seen": 2095120384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036878635907723167,
      "loss": 1.3436,
      "theoretical_loss": 3.4456481148314775,
      "tokens_seen": 2095185920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036876629889669003,
      "loss": 1.3218,
      "theoretical_loss": 3.445639284254407,
      "tokens_seen": 2095251456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036874623871614845,
      "loss": 1.4083,
      "theoretical_loss": 3.445630454030873,
      "tokens_seen": 2095316992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003687261785356068,
      "loss": 1.3455,
      "theoretical_loss": 3.44562162416085,
      "tokens_seen": 2095382528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036870611835506517,
      "loss": 1.3063,
      "theoretical_loss": 3.4456127946443127,
      "tokens_seen": 2095448064
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036868605817452353,
      "loss": 1.2633,
      "theoretical_loss": 3.4456039654812365,
      "tokens_seen": 2095513600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036866599799398195,
      "loss": 1.3725,
      "theoretical_loss": 3.445595136671596,
      "tokens_seen": 2095579136
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 764034,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1738439798355103,
      "objective/train/theoretical_loss": 3.4455863082153657,
      "objective/train/tokens_used": 466103776,
      "theoretical_loss": 3.4455863082153657,
      "tokens_seen": 2095644672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686459378134403,
      "loss": 1.2526,
      "theoretical_loss": 3.4455863082153657,
      "tokens_seen": 2095644672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686258776328987,
      "loss": 1.3044,
      "theoretical_loss": 3.4455774801125205,
      "tokens_seen": 2095710208
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686058174523571,
      "loss": 1.5151,
      "theoretical_loss": 3.4455686523630358,
      "tokens_seen": 2095775744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036858575727181545,
      "loss": 1.3669,
      "theoretical_loss": 3.4455598249668853,
      "tokens_seen": 2095841280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036856569709127387,
      "loss": 1.2847,
      "theoretical_loss": 3.4455509979240446,
      "tokens_seen": 2095906816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036854563691073223,
      "loss": 1.3165,
      "theoretical_loss": 3.4455421712344885,
      "tokens_seen": 2095972352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003685255767301906,
      "loss": 1.3753,
      "theoretical_loss": 3.445533344898191,
      "tokens_seen": 2096037888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036850551654964896,
      "loss": 1.3569,
      "theoretical_loss": 3.4455245189151285,
      "tokens_seen": 2096103424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684854563691074,
      "loss": 1.4557,
      "theoretical_loss": 3.445515693285274,
      "tokens_seen": 2096168960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036846539618856574,
      "loss": 1.3958,
      "theoretical_loss": 3.4455068680086036,
      "tokens_seen": 2096234496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684453360080241,
      "loss": 1.3968,
      "theoretical_loss": 3.4454980430850917,
      "tokens_seen": 2096300032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036842527582748246,
      "loss": 1.3377,
      "theoretical_loss": 3.4454892185147132,
      "tokens_seen": 2096365568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684052156469409,
      "loss": 1.4239,
      "theoretical_loss": 3.4454803942974426,
      "tokens_seen": 2096431104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036838515546639924,
      "loss": 1.4622,
      "theoretical_loss": 3.4454715704332552,
      "tokens_seen": 2096496640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683650952858576,
      "loss": 1.4109,
      "theoretical_loss": 3.445462746922126,
      "tokens_seen": 2096562176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036834503510531596,
      "loss": 1.2503,
      "theoretical_loss": 3.4454539237640285,
      "tokens_seen": 2096627712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683249749247743,
      "loss": 1.403,
      "theoretical_loss": 3.4454451009589393,
      "tokens_seen": 2096693248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036830491474423274,
      "loss": 1.385,
      "theoretical_loss": 3.445436278506832,
      "tokens_seen": 2096758784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003682848545636911,
      "loss": 1.4668,
      "theoretical_loss": 3.4454274564076823,
      "tokens_seen": 2096824320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036826479438314947,
      "loss": 1.3409,
      "theoretical_loss": 3.445418634661465,
      "tokens_seen": 2096889856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036824473420260783,
      "loss": 1.3594,
      "theoretical_loss": 3.445409813268154,
      "tokens_seen": 2096955392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036822467402206625,
      "loss": 1.4659,
      "theoretical_loss": 3.4454009922277242,
      "tokens_seen": 2097020928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003682046138415246,
      "loss": 1.3343,
      "theoretical_loss": 3.445392171540152,
      "tokens_seen": 2097086464
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036818455366098297,
      "loss": 1.4127,
      "theoretical_loss": 3.445383351205411,
      "tokens_seen": 2097152000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036816449348044133,
      "loss": 1.3335,
      "theoretical_loss": 3.445374531223476,
      "tokens_seen": 2097217536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003681444332998997,
      "loss": 1.5249,
      "theoretical_loss": 3.445365711594323,
      "tokens_seen": 2097283072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003681243731193581,
      "loss": 1.3209,
      "theoretical_loss": 3.445356892317925,
      "tokens_seen": 2097348608
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003681043129388165,
      "loss": 1.3393,
      "theoretical_loss": 3.4453480733942587,
      "tokens_seen": 2097414144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036808425275827484,
      "loss": 1.311,
      "theoretical_loss": 3.445339254823298,
      "tokens_seen": 2097479680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680641925777332,
      "loss": 1.226,
      "theoretical_loss": 3.445330436605018,
      "tokens_seen": 2097545216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680441323971916,
      "loss": 1.2744,
      "theoretical_loss": 3.4453216187393934,
      "tokens_seen": 2097610752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036802407221665,
      "loss": 1.3542,
      "theoretical_loss": 3.4453128012263994,
      "tokens_seen": 2097676288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036800401203610834,
      "loss": 1.3595,
      "theoretical_loss": 3.445303984066011,
      "tokens_seen": 2097741824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003679839518555667,
      "loss": 1.3351,
      "theoretical_loss": 3.445295167258202,
      "tokens_seen": 2097807360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036796389167502506,
      "loss": 1.2429,
      "theoretical_loss": 3.4452863508029488,
      "tokens_seen": 2097872896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003679438314944835,
      "loss": 1.3506,
      "theoretical_loss": 3.4452775347002254,
      "tokens_seen": 2097938432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036792377131394184,
      "loss": 1.3581,
      "theoretical_loss": 3.4452687189500066,
      "tokens_seen": 2098003968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003679037111334002,
      "loss": 1.4119,
      "theoretical_loss": 3.4452599035522677,
      "tokens_seen": 2098069504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036788365095285857,
      "loss": 1.3218,
      "theoretical_loss": 3.445251088506984,
      "tokens_seen": 2098135040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000367863590772317,
      "loss": 1.3215,
      "theoretical_loss": 3.445242273814129,
      "tokens_seen": 2098200576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036784353059177535,
      "loss": 1.3167,
      "theoretical_loss": 3.4452334594736786,
      "tokens_seen": 2098266112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003678234704112337,
      "loss": 1.3798,
      "theoretical_loss": 3.4452246454856077,
      "tokens_seen": 2098331648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036780341023069207,
      "loss": 1.5176,
      "theoretical_loss": 3.4452158318498913,
      "tokens_seen": 2098397184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036778335005015043,
      "loss": 1.2312,
      "theoretical_loss": 3.4452070185665042,
      "tokens_seen": 2098462720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036776328986960885,
      "loss": 1.3267,
      "theoretical_loss": 3.445198205635421,
      "tokens_seen": 2098528256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003677432296890672,
      "loss": 1.3297,
      "theoretical_loss": 3.4451893930566166,
      "tokens_seen": 2098593792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036772316950852557,
      "loss": 1.3517,
      "theoretical_loss": 3.4451805808300664,
      "tokens_seen": 2098659328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036770310932798393,
      "loss": 1.3699,
      "theoretical_loss": 3.4451717689557446,
      "tokens_seen": 2098724864
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036768304914744235,
      "loss": 1.3563,
      "theoretical_loss": 3.4451629574336273,
      "tokens_seen": 2098790400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676629889669007,
      "loss": 1.2697,
      "theoretical_loss": 3.4451541462636883,
      "tokens_seen": 2098855936
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 765110,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5750041007995605,
      "objective/train/theoretical_loss": 3.445145335445903,
      "objective/train/tokens_used": 469380576,
      "theoretical_loss": 3.445145335445903,
      "tokens_seen": 2098921472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676429287863591,
      "loss": 1.4734,
      "theoretical_loss": 3.445145335445903,
      "tokens_seen": 2098921472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036762286860581744,
      "loss": 1.3044,
      "theoretical_loss": 3.4451365249802466,
      "tokens_seen": 2098987008
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676028084252758,
      "loss": 1.3037,
      "theoretical_loss": 3.4451277148666932,
      "tokens_seen": 2099052544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675827482447342,
      "loss": 1.4516,
      "theoretical_loss": 3.4451189051052182,
      "tokens_seen": 2099118080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675626880641926,
      "loss": 1.2558,
      "theoretical_loss": 3.4451100956957967,
      "tokens_seen": 2099183616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036754262788365094,
      "loss": 1.431,
      "theoretical_loss": 3.4451012866384034,
      "tokens_seen": 2099249152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675225677031093,
      "loss": 1.2659,
      "theoretical_loss": 3.445092477933014,
      "tokens_seen": 2099314688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675025075225677,
      "loss": 1.381,
      "theoretical_loss": 3.445083669579602,
      "tokens_seen": 2099380224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674824473420261,
      "loss": 1.3665,
      "theoretical_loss": 3.4450748615781435,
      "tokens_seen": 2099445760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036746238716148444,
      "loss": 1.2901,
      "theoretical_loss": 3.445066053928613,
      "tokens_seen": 2099511296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674423269809428,
      "loss": 1.3324,
      "theoretical_loss": 3.4450572466309852,
      "tokens_seen": 2099576832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036742226680040117,
      "loss": 1.2707,
      "theoretical_loss": 3.445048439685236,
      "tokens_seen": 2099642368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674022066198596,
      "loss": 1.3485,
      "theoretical_loss": 3.4450396330913398,
      "tokens_seen": 2099707904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036738214643931795,
      "loss": 1.4169,
      "theoretical_loss": 3.4450308268492713,
      "tokens_seen": 2099773440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003673620862587763,
      "loss": 1.3343,
      "theoretical_loss": 3.4450220209590054,
      "tokens_seen": 2099838976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036734202607823467,
      "loss": 1.2902,
      "theoretical_loss": 3.4450132154205177,
      "tokens_seen": 2099904512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003673219658976931,
      "loss": 1.3017,
      "theoretical_loss": 3.445004410233783,
      "tokens_seen": 2099970048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036730190571715145,
      "loss": 1.3733,
      "theoretical_loss": 3.444995605398776,
      "tokens_seen": 2100035584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003672818455366098,
      "loss": 1.2952,
      "theoretical_loss": 3.4449868009154714,
      "tokens_seen": 2100101120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003672617853560682,
      "loss": 1.2873,
      "theoretical_loss": 3.444977996783845,
      "tokens_seen": 2100166656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036724172517552654,
      "loss": 1.3068,
      "theoretical_loss": 3.444969193003871,
      "tokens_seen": 2100232192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036722166499498495,
      "loss": 1.3285,
      "theoretical_loss": 3.4449603895755248,
      "tokens_seen": 2100297728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003672016048144433,
      "loss": 1.307,
      "theoretical_loss": 3.4449515864987816,
      "tokens_seen": 2100363264
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003671815446339017,
      "loss": 1.4674,
      "theoretical_loss": 3.4449427837736155,
      "tokens_seen": 2100428800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036716148445336004,
      "loss": 1.3359,
      "theoretical_loss": 3.4449339814000024,
      "tokens_seen": 2100494336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036714142427281846,
      "loss": 1.3569,
      "theoretical_loss": 3.4449251793779174,
      "tokens_seen": 2100559872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003671213640922768,
      "loss": 1.3685,
      "theoretical_loss": 3.4449163777073344,
      "tokens_seen": 2100625408
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036710130391173524,
      "loss": 1.2736,
      "theoretical_loss": 3.4449075763882293,
      "tokens_seen": 2100690944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670812437311936,
      "loss": 1.3554,
      "theoretical_loss": 3.444898775420577,
      "tokens_seen": 2100756480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000367061183550652,
      "loss": 1.2767,
      "theoretical_loss": 3.4448899748043518,
      "tokens_seen": 2100822016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670411233701104,
      "loss": 1.3741,
      "theoretical_loss": 3.44488117453953,
      "tokens_seen": 2100887552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036702106318956874,
      "loss": 1.259,
      "theoretical_loss": 3.4448723746260854,
      "tokens_seen": 2100953088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670010030090271,
      "loss": 1.3969,
      "theoretical_loss": 3.4448635750639935,
      "tokens_seen": 2101018624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036698094282848546,
      "loss": 1.3561,
      "theoretical_loss": 3.44485477585323,
      "tokens_seen": 2101084160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003669608826479439,
      "loss": 1.257,
      "theoretical_loss": 3.4448459769937685,
      "tokens_seen": 2101149696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036694082246740224,
      "loss": 1.2743,
      "theoretical_loss": 3.4448371784855847,
      "tokens_seen": 2101215232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003669207622868606,
      "loss": 1.3324,
      "theoretical_loss": 3.4448283803286537,
      "tokens_seen": 2101280768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036690070210631897,
      "loss": 1.3854,
      "theoretical_loss": 3.4448195825229506,
      "tokens_seen": 2101346304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668806419257774,
      "loss": 1.2842,
      "theoretical_loss": 3.4448107850684506,
      "tokens_seen": 2101411840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036686058174523575,
      "loss": 1.3194,
      "theoretical_loss": 3.444801987965128,
      "tokens_seen": 2101477376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668405215646941,
      "loss": 1.3103,
      "theoretical_loss": 3.444793191212958,
      "tokens_seen": 2101542912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036682046138415247,
      "loss": 1.3467,
      "theoretical_loss": 3.444784394811917,
      "tokens_seen": 2101608448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036680040120361083,
      "loss": 1.3795,
      "theoretical_loss": 3.4447755987619777,
      "tokens_seen": 2101673984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036678034102306925,
      "loss": 1.4101,
      "theoretical_loss": 3.444766803063117,
      "tokens_seen": 2101739520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003667602808425276,
      "loss": 1.4029,
      "theoretical_loss": 3.4447580077153095,
      "tokens_seen": 2101805056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366740220661986,
      "loss": 1.3449,
      "theoretical_loss": 3.4447492127185293,
      "tokens_seen": 2101870592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036672016048144434,
      "loss": 1.3296,
      "theoretical_loss": 3.4447404180727528,
      "tokens_seen": 2101936128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036670010030090275,
      "loss": 1.262,
      "theoretical_loss": 3.444731623777954,
      "tokens_seen": 2102001664
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666800401203611,
      "loss": 1.2664,
      "theoretical_loss": 3.4447228298341086,
      "tokens_seen": 2102067200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666599799398195,
      "loss": 1.2667,
      "theoretical_loss": 3.4447140362411917,
      "tokens_seen": 2102132736
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 765966,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3137551546096802,
      "objective/train/theoretical_loss": 3.444705242999178,
      "objective/train/tokens_used": 472657376,
      "theoretical_loss": 3.444705242999178,
      "tokens_seen": 2102198272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036663991975927784,
      "loss": 1.3257,
      "theoretical_loss": 3.444705242999178,
      "tokens_seen": 2102198272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666198595787362,
      "loss": 1.3463,
      "theoretical_loss": 3.4446964501080424,
      "tokens_seen": 2102263808
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665997993981946,
      "loss": 1.2422,
      "theoretical_loss": 3.4446876575677603,
      "tokens_seen": 2102329344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366579739217653,
      "loss": 1.2923,
      "theoretical_loss": 3.444678865378307,
      "tokens_seen": 2102394880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036655967903711134,
      "loss": 1.3261,
      "theoretical_loss": 3.444670073539657,
      "tokens_seen": 2102460416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665396188565697,
      "loss": 1.3142,
      "theoretical_loss": 3.4446612820517855,
      "tokens_seen": 2102525952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665195586760281,
      "loss": 1.3094,
      "theoretical_loss": 3.4446524909146676,
      "tokens_seen": 2102591488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664994984954865,
      "loss": 1.2831,
      "theoretical_loss": 3.444643700128279,
      "tokens_seen": 2102657024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036647943831494484,
      "loss": 1.2635,
      "theoretical_loss": 3.444634909692594,
      "tokens_seen": 2102722560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664593781344032,
      "loss": 1.3214,
      "theoretical_loss": 3.444626119607588,
      "tokens_seen": 2102788096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036643931795386157,
      "loss": 1.3377,
      "theoretical_loss": 3.4446173298732354,
      "tokens_seen": 2102853632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036641925777332,
      "loss": 1.3006,
      "theoretical_loss": 3.4446085404895124,
      "tokens_seen": 2102919168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036639919759277835,
      "loss": 1.4046,
      "theoretical_loss": 3.4445997514563937,
      "tokens_seen": 2102984704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663791374122367,
      "loss": 1.3554,
      "theoretical_loss": 3.4445909627738542,
      "tokens_seen": 2103050240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036635907723169507,
      "loss": 1.3723,
      "theoretical_loss": 3.444582174441869,
      "tokens_seen": 2103115776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663390170511535,
      "loss": 1.3524,
      "theoretical_loss": 3.444573386460413,
      "tokens_seen": 2103181312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036631895687061185,
      "loss": 1.2296,
      "theoretical_loss": 3.444564598829462,
      "tokens_seen": 2103246848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662988966900702,
      "loss": 1.4416,
      "theoretical_loss": 3.4445558115489905,
      "tokens_seen": 2103312384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662788365095286,
      "loss": 1.3546,
      "theoretical_loss": 3.4445470246189736,
      "tokens_seen": 2103377920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036625877632898694,
      "loss": 1.2454,
      "theoretical_loss": 3.4445382380393865,
      "tokens_seen": 2103443456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036623871614844535,
      "loss": 1.3167,
      "theoretical_loss": 3.4445294518102045,
      "tokens_seen": 2103508992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662186559679037,
      "loss": 1.3641,
      "theoretical_loss": 3.4445206659314027,
      "tokens_seen": 2103574528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003661985957873621,
      "loss": 1.2864,
      "theoretical_loss": 3.4445118804029558,
      "tokens_seen": 2103640064
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036617853560682044,
      "loss": 1.3506,
      "theoretical_loss": 3.444503095224839,
      "tokens_seen": 2103705600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036615847542627886,
      "loss": 1.4125,
      "theoretical_loss": 3.4444943103970282,
      "tokens_seen": 2103771136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661384152457372,
      "loss": 1.3836,
      "theoretical_loss": 3.4444855259194975,
      "tokens_seen": 2103836672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661183550651956,
      "loss": 1.26,
      "theoretical_loss": 3.444476741792223,
      "tokens_seen": 2103902208
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036609829488465394,
      "loss": 1.3657,
      "theoretical_loss": 3.4444679580151787,
      "tokens_seen": 2103967744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036607823470411236,
      "loss": 1.2863,
      "theoretical_loss": 3.4444591745883404,
      "tokens_seen": 2104033280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003660581745235707,
      "loss": 1.3529,
      "theoretical_loss": 3.4444503915116833,
      "tokens_seen": 2104098816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003660381143430291,
      "loss": 1.2918,
      "theoretical_loss": 3.4444416087851826,
      "tokens_seen": 2104164352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036601805416248745,
      "loss": 1.3565,
      "theoretical_loss": 3.4444328264088124,
      "tokens_seen": 2104229888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003659979939819458,
      "loss": 1.3884,
      "theoretical_loss": 3.44442404438255,
      "tokens_seen": 2104295424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003659779338014042,
      "loss": 1.3525,
      "theoretical_loss": 3.444415262706368,
      "tokens_seen": 2104360960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003659578736208626,
      "loss": 1.2538,
      "theoretical_loss": 3.444406481380243,
      "tokens_seen": 2104426496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036593781344032095,
      "loss": 1.2556,
      "theoretical_loss": 3.44439770040415,
      "tokens_seen": 2104492032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003659177532597793,
      "loss": 1.2113,
      "theoretical_loss": 3.444388919778064,
      "tokens_seen": 2104557568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036589769307923773,
      "loss": 1.3034,
      "theoretical_loss": 3.44438013950196,
      "tokens_seen": 2104623104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658776328986961,
      "loss": 1.3635,
      "theoretical_loss": 3.4443713595758134,
      "tokens_seen": 2104688640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036585757271815445,
      "loss": 1.27,
      "theoretical_loss": 3.4443625799995994,
      "tokens_seen": 2104754176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658375125376128,
      "loss": 1.353,
      "theoretical_loss": 3.444353800773293,
      "tokens_seen": 2104819712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658174523570712,
      "loss": 1.3352,
      "theoretical_loss": 3.4443450218968694,
      "tokens_seen": 2104885248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657973921765296,
      "loss": 1.3609,
      "theoretical_loss": 3.4443362433703038,
      "tokens_seen": 2104950784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036577733199598796,
      "loss": 1.3424,
      "theoretical_loss": 3.444327465193571,
      "tokens_seen": 2105016320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657572718154463,
      "loss": 1.2589,
      "theoretical_loss": 3.444318687366647,
      "tokens_seen": 2105081856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657372116349047,
      "loss": 1.3467,
      "theoretical_loss": 3.4443099098895065,
      "tokens_seen": 2105147392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657171514543631,
      "loss": 1.2555,
      "theoretical_loss": 3.4443011327621242,
      "tokens_seen": 2105212928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036569709127382146,
      "loss": 1.3104,
      "theoretical_loss": 3.4442923559844756,
      "tokens_seen": 2105278464
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656770310932798,
      "loss": 1.347,
      "theoretical_loss": 3.4442835795565365,
      "tokens_seen": 2105344000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656569709127382,
      "loss": 1.3144,
      "theoretical_loss": 3.4442748034782813,
      "tokens_seen": 2105409536
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 767157,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3022453784942627,
      "objective/train/theoretical_loss": 3.444266027749685,
      "objective/train/tokens_used": 475934176,
      "theoretical_loss": 3.444266027749685,
      "tokens_seen": 2105475072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036563691073219655,
      "loss": 1.2919,
      "theoretical_loss": 3.444266027749685,
      "tokens_seen": 2105475072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036561685055165496,
      "loss": 1.4169,
      "theoretical_loss": 3.4442572523707238,
      "tokens_seen": 2105540608
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655967903711134,
      "loss": 1.2953,
      "theoretical_loss": 3.4442484773413726,
      "tokens_seen": 2105606144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036557673019057174,
      "loss": 1.292,
      "theoretical_loss": 3.444239702661606,
      "tokens_seen": 2105671680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655566700100301,
      "loss": 1.3519,
      "theoretical_loss": 3.4442309283313994,
      "tokens_seen": 2105737216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655366098294885,
      "loss": 1.2543,
      "theoretical_loss": 3.444222154350728,
      "tokens_seen": 2105802752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655165496489469,
      "loss": 1.3603,
      "theoretical_loss": 3.4442133807195674,
      "tokens_seen": 2105868288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036549648946840525,
      "loss": 1.3932,
      "theoretical_loss": 3.444204607437892,
      "tokens_seen": 2105933824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654764292878636,
      "loss": 1.3252,
      "theoretical_loss": 3.4441958345056785,
      "tokens_seen": 2105999360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036545636910732197,
      "loss": 1.3187,
      "theoretical_loss": 3.4441870619229,
      "tokens_seen": 2106064896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654363089267804,
      "loss": 1.3562,
      "theoretical_loss": 3.4441782896895337,
      "tokens_seen": 2106130432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036541624874623875,
      "loss": 1.4001,
      "theoretical_loss": 3.4441695178055536,
      "tokens_seen": 2106195968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653961885656971,
      "loss": 1.3228,
      "theoretical_loss": 3.444160746270935,
      "tokens_seen": 2106261504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036537612838515547,
      "loss": 1.3213,
      "theoretical_loss": 3.4441519750856537,
      "tokens_seen": 2106327040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653560682046139,
      "loss": 1.2848,
      "theoretical_loss": 3.4441432042496842,
      "tokens_seen": 2106392576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036533600802407225,
      "loss": 1.4016,
      "theoretical_loss": 3.4441344337630024,
      "tokens_seen": 2106458112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653159478435306,
      "loss": 1.3188,
      "theoretical_loss": 3.4441256636255835,
      "tokens_seen": 2106523648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365295887662989,
      "loss": 1.3766,
      "theoretical_loss": 3.444116893837402,
      "tokens_seen": 2106589184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036527582748244734,
      "loss": 1.3528,
      "theoretical_loss": 3.4441081243984337,
      "tokens_seen": 2106654720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036525576730190575,
      "loss": 1.2933,
      "theoretical_loss": 3.444099355308654,
      "tokens_seen": 2106720256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652357071213641,
      "loss": 1.284,
      "theoretical_loss": 3.444090586568038,
      "tokens_seen": 2106785792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652156469408225,
      "loss": 1.368,
      "theoretical_loss": 3.44408181817656,
      "tokens_seen": 2106851328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036519558676028084,
      "loss": 1.2779,
      "theoretical_loss": 3.4440730501341967,
      "tokens_seen": 2106916864
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036517552657973926,
      "loss": 1.3118,
      "theoretical_loss": 3.444064282440922,
      "tokens_seen": 2106982400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003651554663991976,
      "loss": 1.2364,
      "theoretical_loss": 3.4440555150967125,
      "tokens_seen": 2107047936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365135406218656,
      "loss": 1.2787,
      "theoretical_loss": 3.444046748101542,
      "tokens_seen": 2107113472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036511534603811434,
      "loss": 1.2568,
      "theoretical_loss": 3.444037981455387,
      "tokens_seen": 2107179008
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036509528585757276,
      "loss": 1.2529,
      "theoretical_loss": 3.4440292151582224,
      "tokens_seen": 2107244544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003650752256770311,
      "loss": 1.2324,
      "theoretical_loss": 3.444020449210023,
      "tokens_seen": 2107310080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003650551654964895,
      "loss": 1.3414,
      "theoretical_loss": 3.4440116836107646,
      "tokens_seen": 2107375616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036503510531594785,
      "loss": 1.201,
      "theoretical_loss": 3.444002918360422,
      "tokens_seen": 2107441152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003650150451354062,
      "loss": 1.3474,
      "theoretical_loss": 3.443994153458971,
      "tokens_seen": 2107506688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003649949849548646,
      "loss": 1.3961,
      "theoretical_loss": 3.4439853889063863,
      "tokens_seen": 2107572224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000364974924774323,
      "loss": 1.4234,
      "theoretical_loss": 3.4439766247026435,
      "tokens_seen": 2107637760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036495486459378135,
      "loss": 1.2678,
      "theoretical_loss": 3.4439678608477178,
      "tokens_seen": 2107703296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003649348044132397,
      "loss": 1.3538,
      "theoretical_loss": 3.443959097341584,
      "tokens_seen": 2107768832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036491474423269813,
      "loss": 1.2892,
      "theoretical_loss": 3.4439503341842186,
      "tokens_seen": 2107834368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648946840521565,
      "loss": 1.3419,
      "theoretical_loss": 3.443941571375596,
      "tokens_seen": 2107899904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036487462387161485,
      "loss": 1.2762,
      "theoretical_loss": 3.443932808915691,
      "tokens_seen": 2107965440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648545636910732,
      "loss": 1.3632,
      "theoretical_loss": 3.4439240468044803,
      "tokens_seen": 2108030976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648345035105316,
      "loss": 1.2554,
      "theoretical_loss": 3.4439152850419377,
      "tokens_seen": 2108096512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036481444332999,
      "loss": 1.361,
      "theoretical_loss": 3.4439065236280393,
      "tokens_seen": 2108162048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036479438314944836,
      "loss": 1.2867,
      "theoretical_loss": 3.4438977625627603,
      "tokens_seen": 2108227584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647743229689067,
      "loss": 1.3119,
      "theoretical_loss": 3.4438890018460757,
      "tokens_seen": 2108293120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647542627883651,
      "loss": 1.3828,
      "theoretical_loss": 3.4438802414779612,
      "tokens_seen": 2108358656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647342026078235,
      "loss": 1.294,
      "theoretical_loss": 3.443871481458392,
      "tokens_seen": 2108424192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036471414242728186,
      "loss": 1.2974,
      "theoretical_loss": 3.4438627217873434,
      "tokens_seen": 2108489728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646940822467402,
      "loss": 1.3563,
      "theoretical_loss": 3.4438539624647904,
      "tokens_seen": 2108555264
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646740220661986,
      "loss": 1.4333,
      "theoretical_loss": 3.4438452034907083,
      "tokens_seen": 2108620800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036465396188565695,
      "loss": 1.3752,
      "theoretical_loss": 3.443836444865073,
      "tokens_seen": 2108686336
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 768317,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.153150200843811,
      "objective/train/theoretical_loss": 3.443827686587859,
      "objective/train/tokens_used": 479210976,
      "theoretical_loss": 3.443827686587859,
      "tokens_seen": 2108751872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036463390170511536,
      "loss": 1.1949,
      "theoretical_loss": 3.443827686587859,
      "tokens_seen": 2108751872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646138415245737,
      "loss": 1.3138,
      "theoretical_loss": 3.443818928659043,
      "tokens_seen": 2108817408
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645937813440321,
      "loss": 1.3458,
      "theoretical_loss": 3.4438101710785984,
      "tokens_seen": 2108882944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036457372116349045,
      "loss": 1.3948,
      "theoretical_loss": 3.443801413846502,
      "tokens_seen": 2108948480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036455366098294887,
      "loss": 1.3119,
      "theoretical_loss": 3.4437926569627284,
      "tokens_seen": 2109014016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036453360080240723,
      "loss": 1.3953,
      "theoretical_loss": 3.4437839004272535,
      "tokens_seen": 2109079552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645135406218656,
      "loss": 1.3311,
      "theoretical_loss": 3.4437751442400515,
      "tokens_seen": 2109145088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036449348044132395,
      "loss": 1.3462,
      "theoretical_loss": 3.443766388401099,
      "tokens_seen": 2109210624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644734202607823,
      "loss": 1.2598,
      "theoretical_loss": 3.443757632910371,
      "tokens_seen": 2109276160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036445336008024073,
      "loss": 1.3711,
      "theoretical_loss": 3.4437488777678418,
      "tokens_seen": 2109341696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644332998996991,
      "loss": 1.3975,
      "theoretical_loss": 3.4437401229734883,
      "tokens_seen": 2109407232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036441323971915746,
      "loss": 1.4033,
      "theoretical_loss": 3.443731368527285,
      "tokens_seen": 2109472768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643931795386158,
      "loss": 1.3222,
      "theoretical_loss": 3.4437226144292072,
      "tokens_seen": 2109538304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036437311935807424,
      "loss": 1.3333,
      "theoretical_loss": 3.4437138606792304,
      "tokens_seen": 2109603840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643530591775326,
      "loss": 1.272,
      "theoretical_loss": 3.4437051072773297,
      "tokens_seen": 2109669376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036433299899699096,
      "loss": 1.2041,
      "theoretical_loss": 3.4436963542234813,
      "tokens_seen": 2109734912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643129388164493,
      "loss": 1.3683,
      "theoretical_loss": 3.4436876015176594,
      "tokens_seen": 2109800448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642928786359077,
      "loss": 1.393,
      "theoretical_loss": 3.4436788491598405,
      "tokens_seen": 2109865984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642728184553661,
      "loss": 1.2238,
      "theoretical_loss": 3.4436700971499983,
      "tokens_seen": 2109931520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036425275827482446,
      "loss": 1.1717,
      "theoretical_loss": 3.4436613454881098,
      "tokens_seen": 2109997056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642326980942828,
      "loss": 1.3435,
      "theoretical_loss": 3.44365259417415,
      "tokens_seen": 2110062592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642126379137412,
      "loss": 1.3071,
      "theoretical_loss": 3.443643843208094,
      "tokens_seen": 2110128128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641925777331996,
      "loss": 1.3552,
      "theoretical_loss": 3.4436350925899166,
      "tokens_seen": 2110193664
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036417251755265797,
      "loss": 1.3927,
      "theoretical_loss": 3.443626342319594,
      "tokens_seen": 2110259200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036415245737211633,
      "loss": 1.2876,
      "theoretical_loss": 3.4436175923971017,
      "tokens_seen": 2110324736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641323971915747,
      "loss": 1.2379,
      "theoretical_loss": 3.4436088428224143,
      "tokens_seen": 2110390272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036411233701103316,
      "loss": 1.3604,
      "theoretical_loss": 3.443600093595508,
      "tokens_seen": 2110455808
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640922768304915,
      "loss": 1.3995,
      "theoretical_loss": 3.4435913447163573,
      "tokens_seen": 2110521344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640722166499499,
      "loss": 1.2756,
      "theoretical_loss": 3.443582596184938,
      "tokens_seen": 2110586880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036405215646940825,
      "loss": 1.2481,
      "theoretical_loss": 3.4435738480012255,
      "tokens_seen": 2110652416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640320962888666,
      "loss": 1.291,
      "theoretical_loss": 3.4435651001651957,
      "tokens_seen": 2110717952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000364012036108325,
      "loss": 1.3307,
      "theoretical_loss": 3.443556352676823,
      "tokens_seen": 2110783488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003639919759277834,
      "loss": 1.2903,
      "theoretical_loss": 3.443547605536083,
      "tokens_seen": 2110849024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036397191574724175,
      "loss": 1.1547,
      "theoretical_loss": 3.443538858742952,
      "tokens_seen": 2110914560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003639518555667001,
      "loss": 1.3908,
      "theoretical_loss": 3.4435301122974047,
      "tokens_seen": 2110980096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036393179538615853,
      "loss": 1.2112,
      "theoretical_loss": 3.443521366199416,
      "tokens_seen": 2111045632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003639117352056169,
      "loss": 1.1677,
      "theoretical_loss": 3.4435126204489626,
      "tokens_seen": 2111111168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036389167502507525,
      "loss": 1.3342,
      "theoretical_loss": 3.4435038750460185,
      "tokens_seen": 2111176704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638716148445336,
      "loss": 1.4041,
      "theoretical_loss": 3.44349512999056,
      "tokens_seen": 2111242240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363851554663992,
      "loss": 1.2708,
      "theoretical_loss": 3.4434863852825623,
      "tokens_seen": 2111307776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638314944834504,
      "loss": 1.3325,
      "theoretical_loss": 3.443477640922001,
      "tokens_seen": 2111373312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036381143430290876,
      "loss": 1.1847,
      "theoretical_loss": 3.4434688969088514,
      "tokens_seen": 2111438848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637913741223671,
      "loss": 1.4129,
      "theoretical_loss": 3.4434601532430884,
      "tokens_seen": 2111504384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637713139418255,
      "loss": 1.2798,
      "theoretical_loss": 3.4434514099246876,
      "tokens_seen": 2111569920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637512537612839,
      "loss": 1.1966,
      "theoretical_loss": 3.443442666953625,
      "tokens_seen": 2111635456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036373119358074226,
      "loss": 1.3453,
      "theoretical_loss": 3.4434339243298755,
      "tokens_seen": 2111700992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637111334002006,
      "loss": 1.305,
      "theoretical_loss": 3.4434251820534145,
      "tokens_seen": 2111766528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363691073219659,
      "loss": 1.2496,
      "theoretical_loss": 3.443416440124218,
      "tokens_seen": 2111832064
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036367101303911735,
      "loss": 1.2774,
      "theoretical_loss": 3.443407698542261,
      "tokens_seen": 2111897600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036365095285857576,
      "loss": 1.3294,
      "theoretical_loss": 3.443398957307519,
      "tokens_seen": 2111963136
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 769436,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3213162422180176,
      "objective/train/theoretical_loss": 3.4433902164199672,
      "objective/train/tokens_used": 482487776,
      "theoretical_loss": 3.4433902164199672,
      "tokens_seen": 2112028672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636308926780341,
      "loss": 1.4059,
      "theoretical_loss": 3.4433902164199672,
      "tokens_seen": 2112028672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636108324974925,
      "loss": 1.2794,
      "theoretical_loss": 3.4433814758795815,
      "tokens_seen": 2112094208
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036359077231695085,
      "loss": 1.3745,
      "theoretical_loss": 3.443372735686337,
      "tokens_seen": 2112159744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036357071213640927,
      "loss": 1.3908,
      "theoretical_loss": 3.443363995840209,
      "tokens_seen": 2112225280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036355065195586763,
      "loss": 1.2363,
      "theoretical_loss": 3.4433552563411736,
      "tokens_seen": 2112290816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363530591775326,
      "loss": 1.2649,
      "theoretical_loss": 3.4433465171892057,
      "tokens_seen": 2112356352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036351053159478435,
      "loss": 1.3192,
      "theoretical_loss": 3.443337778384281,
      "tokens_seen": 2112421888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634904714142427,
      "loss": 1.3019,
      "theoretical_loss": 3.4433290399263745,
      "tokens_seen": 2112487424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036347041123370113,
      "loss": 1.3039,
      "theoretical_loss": 3.4433203018154623,
      "tokens_seen": 2112552960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634503510531595,
      "loss": 1.2723,
      "theoretical_loss": 3.443311564051519,
      "tokens_seen": 2112618496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036343029087261786,
      "loss": 1.3675,
      "theoretical_loss": 3.443302826634521,
      "tokens_seen": 2112684032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634102306920762,
      "loss": 1.4277,
      "theoretical_loss": 3.4432940895644433,
      "tokens_seen": 2112749568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036339017051153464,
      "loss": 1.2983,
      "theoretical_loss": 3.4432853528412615,
      "tokens_seen": 2112815104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363370110330993,
      "loss": 1.3462,
      "theoretical_loss": 3.443276616464951,
      "tokens_seen": 2112880640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036335005015045136,
      "loss": 1.4393,
      "theoretical_loss": 3.443267880435487,
      "tokens_seen": 2112946176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633299899699097,
      "loss": 1.4435,
      "theoretical_loss": 3.4432591447528456,
      "tokens_seen": 2113011712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633099297893681,
      "loss": 1.2632,
      "theoretical_loss": 3.4432504094170016,
      "tokens_seen": 2113077248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632898696088265,
      "loss": 1.2447,
      "theoretical_loss": 3.443241674427931,
      "tokens_seen": 2113142784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036326980942828486,
      "loss": 1.237,
      "theoretical_loss": 3.443232939785609,
      "tokens_seen": 2113208320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632497492477432,
      "loss": 1.2437,
      "theoretical_loss": 3.443224205490011,
      "tokens_seen": 2113273856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632296890672016,
      "loss": 1.2972,
      "theoretical_loss": 3.443215471541113,
      "tokens_seen": 2113339392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036320962888666,
      "loss": 1.3624,
      "theoretical_loss": 3.44320673793889,
      "tokens_seen": 2113404928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036318956870611837,
      "loss": 1.3551,
      "theoretical_loss": 3.4431980046833175,
      "tokens_seen": 2113470464
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036316950852557673,
      "loss": 1.3455,
      "theoretical_loss": 3.4431892717743713,
      "tokens_seen": 2113536000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631494483450351,
      "loss": 1.2969,
      "theoretical_loss": 3.443180539212026,
      "tokens_seen": 2113601536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036312938816449345,
      "loss": 1.2587,
      "theoretical_loss": 3.4431718069962587,
      "tokens_seen": 2113667072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036310932798395187,
      "loss": 1.3012,
      "theoretical_loss": 3.4431630751270434,
      "tokens_seen": 2113732608
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036308926780341023,
      "loss": 1.2223,
      "theoretical_loss": 3.4431543436043563,
      "tokens_seen": 2113798144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003630692076228686,
      "loss": 1.2994,
      "theoretical_loss": 3.4431456124281734,
      "tokens_seen": 2113863680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036304914744232696,
      "loss": 1.3054,
      "theoretical_loss": 3.443136881598469,
      "tokens_seen": 2113929216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036302908726178537,
      "loss": 1.5396,
      "theoretical_loss": 3.443128151115219,
      "tokens_seen": 2113994752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036300902708124374,
      "loss": 1.3605,
      "theoretical_loss": 3.4431194209784,
      "tokens_seen": 2114060288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629889669007021,
      "loss": 1.3526,
      "theoretical_loss": 3.443110691187986,
      "tokens_seen": 2114125824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036296890672016046,
      "loss": 1.3501,
      "theoretical_loss": 3.4431019617439533,
      "tokens_seen": 2114191360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629488465396188,
      "loss": 1.3248,
      "theoretical_loss": 3.4430932326462775,
      "tokens_seen": 2114256896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036292878635907724,
      "loss": 1.3472,
      "theoretical_loss": 3.4430845038949336,
      "tokens_seen": 2114322432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629087261785356,
      "loss": 1.3485,
      "theoretical_loss": 3.4430757754898975,
      "tokens_seen": 2114387968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036288866599799396,
      "loss": 1.3616,
      "theoretical_loss": 3.4430670474311444,
      "tokens_seen": 2114453504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628686058174523,
      "loss": 1.2865,
      "theoretical_loss": 3.4430583197186504,
      "tokens_seen": 2114519040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036284854563691074,
      "loss": 1.3653,
      "theoretical_loss": 3.443049592352391,
      "tokens_seen": 2114584576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628284854563691,
      "loss": 1.3586,
      "theoretical_loss": 3.443040865332341,
      "tokens_seen": 2114650112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036280842527582747,
      "loss": 1.2638,
      "theoretical_loss": 3.443032138658477,
      "tokens_seen": 2114715648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036278836509528583,
      "loss": 1.2767,
      "theoretical_loss": 3.4430234123307732,
      "tokens_seen": 2114781184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036276830491474424,
      "loss": 1.2595,
      "theoretical_loss": 3.443014686349206,
      "tokens_seen": 2114846720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627482447342026,
      "loss": 1.3639,
      "theoretical_loss": 3.4430059607137515,
      "tokens_seen": 2114912256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036272818455366097,
      "loss": 1.3061,
      "theoretical_loss": 3.442997235424384,
      "tokens_seen": 2114977792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036270812437311933,
      "loss": 1.2461,
      "theoretical_loss": 3.4429885104810793,
      "tokens_seen": 2115043328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626880641925777,
      "loss": 1.2092,
      "theoretical_loss": 3.442979785883814,
      "tokens_seen": 2115108864
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626680040120361,
      "loss": 1.3049,
      "theoretical_loss": 3.442971061632562,
      "tokens_seen": 2115174400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036264794383149447,
      "loss": 1.3942,
      "theoretical_loss": 3.442962337727301,
      "tokens_seen": 2115239936
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 770743,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.205157995223999,
      "objective/train/theoretical_loss": 3.4429536141680046,
      "objective/train/tokens_used": 485764576,
      "theoretical_loss": 3.4429536141680046,
      "tokens_seen": 2115305472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036262788365095283,
      "loss": 1.2541,
      "theoretical_loss": 3.4429536141680046,
      "tokens_seen": 2115305472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036260782347041125,
      "loss": 1.3759,
      "theoretical_loss": 3.442944890954649,
      "tokens_seen": 2115371008
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036258776328986967,
      "loss": 1.3069,
      "theoretical_loss": 3.4429361680872104,
      "tokens_seen": 2115436544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036256770310932803,
      "loss": 1.3426,
      "theoretical_loss": 3.442927445565663,
      "tokens_seen": 2115502080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625476429287864,
      "loss": 1.286,
      "theoretical_loss": 3.442918723389984,
      "tokens_seen": 2115567616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036252758274824475,
      "loss": 1.2622,
      "theoretical_loss": 3.442910001560148,
      "tokens_seen": 2115633152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625075225677031,
      "loss": 1.2757,
      "theoretical_loss": 3.4429012800761303,
      "tokens_seen": 2115698688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036248746238716153,
      "loss": 1.401,
      "theoretical_loss": 3.442892558937907,
      "tokens_seen": 2115764224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624674022066199,
      "loss": 1.2647,
      "theoretical_loss": 3.4428838381454545,
      "tokens_seen": 2115829760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036244734202607826,
      "loss": 1.27,
      "theoretical_loss": 3.4428751176987467,
      "tokens_seen": 2115895296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624272818455366,
      "loss": 1.328,
      "theoretical_loss": 3.4428663975977596,
      "tokens_seen": 2115960832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036240722166499504,
      "loss": 1.388,
      "theoretical_loss": 3.44285767784247,
      "tokens_seen": 2116026368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623871614844534,
      "loss": 1.273,
      "theoretical_loss": 3.4428489584328523,
      "tokens_seen": 2116091904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036236710130391176,
      "loss": 1.3418,
      "theoretical_loss": 3.4428402393688824,
      "tokens_seen": 2116157440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623470411233701,
      "loss": 1.3098,
      "theoretical_loss": 3.442831520650536,
      "tokens_seen": 2116222976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623269809428285,
      "loss": 1.3199,
      "theoretical_loss": 3.4428228022777887,
      "tokens_seen": 2116288512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623069207622869,
      "loss": 1.2965,
      "theoretical_loss": 3.442814084250616,
      "tokens_seen": 2116354048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036228686058174526,
      "loss": 1.313,
      "theoretical_loss": 3.4428053665689937,
      "tokens_seen": 2116419584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622668004012036,
      "loss": 1.4513,
      "theoretical_loss": 3.4427966492328967,
      "tokens_seen": 2116485120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362246740220662,
      "loss": 1.2546,
      "theoretical_loss": 3.4427879322423016,
      "tokens_seen": 2116550656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622266800401204,
      "loss": 1.2784,
      "theoretical_loss": 3.4427792155971835,
      "tokens_seen": 2116616192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036220661985957877,
      "loss": 1.461,
      "theoretical_loss": 3.442770499297518,
      "tokens_seen": 2116681728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036218655967903713,
      "loss": 1.2706,
      "theoretical_loss": 3.442761783343281,
      "tokens_seen": 2116747264
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621664994984955,
      "loss": 1.3393,
      "theoretical_loss": 3.442753067734447,
      "tokens_seen": 2116812800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036214643931795385,
      "loss": 1.3943,
      "theoretical_loss": 3.4427443524709935,
      "tokens_seen": 2116878336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036212637913741227,
      "loss": 1.3089,
      "theoretical_loss": 3.442735637552895,
      "tokens_seen": 2116943872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036210631895687063,
      "loss": 1.3366,
      "theoretical_loss": 3.442726922980127,
      "tokens_seen": 2117009408
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362086258776329,
      "loss": 1.2695,
      "theoretical_loss": 3.442718208752665,
      "tokens_seen": 2117074944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036206619859578736,
      "loss": 1.2996,
      "theoretical_loss": 3.4427094948704857,
      "tokens_seen": 2117140480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620461384152458,
      "loss": 1.3239,
      "theoretical_loss": 3.442700781333564,
      "tokens_seen": 2117206016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036202607823470414,
      "loss": 1.3329,
      "theoretical_loss": 3.4426920681418753,
      "tokens_seen": 2117271552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620060180541625,
      "loss": 1.2319,
      "theoretical_loss": 3.4426833552953955,
      "tokens_seen": 2117337088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036198595787362086,
      "loss": 1.3202,
      "theoretical_loss": 3.4426746427941004,
      "tokens_seen": 2117402624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003619658976930792,
      "loss": 1.2641,
      "theoretical_loss": 3.4426659306379657,
      "tokens_seen": 2117468160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036194583751253764,
      "loss": 1.384,
      "theoretical_loss": 3.4426572188269664,
      "tokens_seen": 2117533696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000361925777331996,
      "loss": 1.3536,
      "theoretical_loss": 3.4426485073610786,
      "tokens_seen": 2117599232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036190571715145436,
      "loss": 1.189,
      "theoretical_loss": 3.442639796240278,
      "tokens_seen": 2117664768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618856569709127,
      "loss": 1.2704,
      "theoretical_loss": 3.4426310854645408,
      "tokens_seen": 2117730304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036186559679037114,
      "loss": 1.3323,
      "theoretical_loss": 3.4426223750338414,
      "tokens_seen": 2117795840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618455366098295,
      "loss": 1.3183,
      "theoretical_loss": 3.442613664948156,
      "tokens_seen": 2117861376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036182547642928787,
      "loss": 1.2426,
      "theoretical_loss": 3.442604955207461,
      "tokens_seen": 2117926912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036180541624874623,
      "loss": 1.303,
      "theoretical_loss": 3.442596245811731,
      "tokens_seen": 2117992448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036178535606820464,
      "loss": 1.2611,
      "theoretical_loss": 3.442587536760942,
      "tokens_seen": 2118057984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000361765295887663,
      "loss": 1.3111,
      "theoretical_loss": 3.4425788280550695,
      "tokens_seen": 2118123520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036174523570712137,
      "loss": 1.3992,
      "theoretical_loss": 3.44257011969409,
      "tokens_seen": 2118189056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036172517552657973,
      "loss": 1.3823,
      "theoretical_loss": 3.4425614116779784,
      "tokens_seen": 2118254592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003617051153460381,
      "loss": 1.2894,
      "theoretical_loss": 3.44255270400671,
      "tokens_seen": 2118320128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003616850551654965,
      "loss": 1.3736,
      "theoretical_loss": 3.4425439966802616,
      "tokens_seen": 2118385664
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036166499498495487,
      "loss": 1.2861,
      "theoretical_loss": 3.4425352896986086,
      "tokens_seen": 2118451200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036164493480441323,
      "loss": 1.3303,
      "theoretical_loss": 3.442526583061726,
      "tokens_seen": 2118516736
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 771858,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.377901554107666,
      "objective/train/theoretical_loss": 3.44251787676959,
      "objective/train/tokens_used": 489041376,
      "theoretical_loss": 3.44251787676959,
      "tokens_seen": 2118582272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003616248746238716,
      "loss": 1.4125,
      "theoretical_loss": 3.44251787676959,
      "tokens_seen": 2118582272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036160481444333,
      "loss": 1.3668,
      "theoretical_loss": 3.442509170822176,
      "tokens_seen": 2118647808
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615847542627884,
      "loss": 1.3339,
      "theoretical_loss": 3.4425004652194593,
      "tokens_seen": 2118713344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036156469408224674,
      "loss": 1.2589,
      "theoretical_loss": 3.442491759961417,
      "tokens_seen": 2118778880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615446339017051,
      "loss": 1.372,
      "theoretical_loss": 3.4424830550480237,
      "tokens_seen": 2118844416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036152457372116346,
      "loss": 1.3164,
      "theoretical_loss": 3.442474350479255,
      "tokens_seen": 2118909952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615045135406219,
      "loss": 1.2664,
      "theoretical_loss": 3.4424656462550876,
      "tokens_seen": 2118975488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036148445336008024,
      "loss": 1.3556,
      "theoretical_loss": 3.442456942375496,
      "tokens_seen": 2119041024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614643931795386,
      "loss": 1.3522,
      "theoretical_loss": 3.4424482388404565,
      "tokens_seen": 2119106560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036144433299899697,
      "loss": 1.3515,
      "theoretical_loss": 3.442439535649945,
      "tokens_seen": 2119172096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614242728184554,
      "loss": 1.3032,
      "theoretical_loss": 3.442430832803937,
      "tokens_seen": 2119237632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036140421263791374,
      "loss": 1.3835,
      "theoretical_loss": 3.4424221303024076,
      "tokens_seen": 2119303168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613841524573721,
      "loss": 1.3999,
      "theoretical_loss": 3.4424134281453336,
      "tokens_seen": 2119368704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036136409227683047,
      "loss": 1.3163,
      "theoretical_loss": 3.44240472633269,
      "tokens_seen": 2119434240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036134403209628883,
      "loss": 1.3365,
      "theoretical_loss": 3.4423960248644523,
      "tokens_seen": 2119499776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036132397191574725,
      "loss": 1.3551,
      "theoretical_loss": 3.442387323740597,
      "tokens_seen": 2119565312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613039117352056,
      "loss": 1.4044,
      "theoretical_loss": 3.4423786229610998,
      "tokens_seen": 2119630848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036128385155466397,
      "loss": 1.3283,
      "theoretical_loss": 3.442369922525936,
      "tokens_seen": 2119696384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036126379137412233,
      "loss": 1.3172,
      "theoretical_loss": 3.442361222435081,
      "tokens_seen": 2119761920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036124373119358075,
      "loss": 1.3253,
      "theoretical_loss": 3.4423525226885108,
      "tokens_seen": 2119827456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612236710130391,
      "loss": 1.332,
      "theoretical_loss": 3.442343823286201,
      "tokens_seen": 2119892992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612036108324975,
      "loss": 1.306,
      "theoretical_loss": 3.4423351242281286,
      "tokens_seen": 2119958528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036118355065195584,
      "loss": 1.3389,
      "theoretical_loss": 3.442326425514268,
      "tokens_seen": 2120024064
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003611634904714142,
      "loss": 1.3884,
      "theoretical_loss": 3.442317727144595,
      "tokens_seen": 2120089600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003611434302908726,
      "loss": 1.3722,
      "theoretical_loss": 3.4423090291190857,
      "tokens_seen": 2120155136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000361123370110331,
      "loss": 1.3955,
      "theoretical_loss": 3.442300331437716,
      "tokens_seen": 2120220672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003611033099297894,
      "loss": 1.3419,
      "theoretical_loss": 3.442291634100461,
      "tokens_seen": 2120286208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036108324974924776,
      "loss": 1.3343,
      "theoretical_loss": 3.4422829371072967,
      "tokens_seen": 2120351744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610631895687062,
      "loss": 1.2999,
      "theoretical_loss": 3.4422742404581994,
      "tokens_seen": 2120417280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036104312938816454,
      "loss": 1.3733,
      "theoretical_loss": 3.4422655441531447,
      "tokens_seen": 2120482816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610230692076229,
      "loss": 1.1956,
      "theoretical_loss": 3.442256848192107,
      "tokens_seen": 2120548352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036100300902708126,
      "loss": 1.3165,
      "theoretical_loss": 3.442248152575065,
      "tokens_seen": 2120613888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003609829488465396,
      "loss": 1.1961,
      "theoretical_loss": 3.4422394573019908,
      "tokens_seen": 2120679424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036096288866599804,
      "loss": 1.3948,
      "theoretical_loss": 3.442230762372863,
      "tokens_seen": 2120744960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003609428284854564,
      "loss": 1.2374,
      "theoretical_loss": 3.4422220677876565,
      "tokens_seen": 2120810496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036092276830491476,
      "loss": 1.4987,
      "theoretical_loss": 3.4422133735463465,
      "tokens_seen": 2120876032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003609027081243731,
      "loss": 1.3857,
      "theoretical_loss": 3.4422046796489094,
      "tokens_seen": 2120941568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036088264794383154,
      "loss": 1.2896,
      "theoretical_loss": 3.442195986095321,
      "tokens_seen": 2121007104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608625877632899,
      "loss": 1.237,
      "theoretical_loss": 3.4421872928855564,
      "tokens_seen": 2121072640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036084252758274827,
      "loss": 1.2575,
      "theoretical_loss": 3.442178600019592,
      "tokens_seen": 2121138176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036082246740220663,
      "loss": 1.2289,
      "theoretical_loss": 3.4421699074974033,
      "tokens_seen": 2121203712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036080240722166505,
      "loss": 1.2649,
      "theoretical_loss": 3.4421612153189667,
      "tokens_seen": 2121269248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607823470411234,
      "loss": 1.3219,
      "theoretical_loss": 3.442152523484257,
      "tokens_seen": 2121334784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036076228686058177,
      "loss": 1.3498,
      "theoretical_loss": 3.442143831993251,
      "tokens_seen": 2121400320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036074222668004013,
      "loss": 1.3002,
      "theoretical_loss": 3.4421351408459238,
      "tokens_seen": 2121465856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607221664994985,
      "loss": 1.3041,
      "theoretical_loss": 3.442126450042251,
      "tokens_seen": 2121531392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607021063189569,
      "loss": 1.335,
      "theoretical_loss": 3.4421177595822092,
      "tokens_seen": 2121596928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036068204613841527,
      "loss": 1.3382,
      "theoretical_loss": 3.4421090694657734,
      "tokens_seen": 2121662464
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036066198595787364,
      "loss": 1.2393,
      "theoretical_loss": 3.4421003796929197,
      "tokens_seen": 2121728000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360641925777332,
      "loss": 1.2346,
      "theoretical_loss": 3.4420916902636245,
      "tokens_seen": 2121793536
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 773131,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.180802345275879,
      "objective/train/theoretical_loss": 3.442083001177863,
      "objective/train/tokens_used": 492318176,
      "theoretical_loss": 3.442083001177863,
      "tokens_seen": 2121859072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606218655967904,
      "loss": 1.2983,
      "theoretical_loss": 3.442083001177863,
      "tokens_seen": 2121859072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606018054162488,
      "loss": 1.4459,
      "theoretical_loss": 3.4420743124356106,
      "tokens_seen": 2121924608
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036058174523570714,
      "loss": 1.2797,
      "theoretical_loss": 3.442065624036844,
      "tokens_seen": 2121990144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605616850551655,
      "loss": 1.357,
      "theoretical_loss": 3.4420569359815385,
      "tokens_seen": 2122055680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036054162487462386,
      "loss": 1.3843,
      "theoretical_loss": 3.4420482482696704,
      "tokens_seen": 2122121216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605215646940823,
      "loss": 1.3349,
      "theoretical_loss": 3.4420395609012147,
      "tokens_seen": 2122186752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036050150451354064,
      "loss": 1.2876,
      "theoretical_loss": 3.4420308738761474,
      "tokens_seen": 2122252288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360481444332999,
      "loss": 1.2774,
      "theoretical_loss": 3.4420221871944454,
      "tokens_seen": 2122317824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036046138415245737,
      "loss": 1.3869,
      "theoretical_loss": 3.442013500856083,
      "tokens_seen": 2122383360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003604413239719158,
      "loss": 1.311,
      "theoretical_loss": 3.442004814861037,
      "tokens_seen": 2122448896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036042126379137414,
      "loss": 1.4405,
      "theoretical_loss": 3.441996129209283,
      "tokens_seen": 2122514432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003604012036108325,
      "loss": 1.3202,
      "theoretical_loss": 3.4419874439007967,
      "tokens_seen": 2122579968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036038114343029087,
      "loss": 1.414,
      "theoretical_loss": 3.441978758935554,
      "tokens_seen": 2122645504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036036108324974923,
      "loss": 1.4109,
      "theoretical_loss": 3.441970074313531,
      "tokens_seen": 2122711040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036034102306920765,
      "loss": 1.3051,
      "theoretical_loss": 3.4419613900347032,
      "tokens_seen": 2122776576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360320962888666,
      "loss": 1.3899,
      "theoretical_loss": 3.4419527060990465,
      "tokens_seen": 2122842112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036030090270812437,
      "loss": 1.3934,
      "theoretical_loss": 3.4419440225065374,
      "tokens_seen": 2122907648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036028084252758273,
      "loss": 1.39,
      "theoretical_loss": 3.4419353392571503,
      "tokens_seen": 2122973184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036026078234704115,
      "loss": 1.3739,
      "theoretical_loss": 3.441926656350862,
      "tokens_seen": 2123038720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602407221664995,
      "loss": 1.1971,
      "theoretical_loss": 3.441917973787649,
      "tokens_seen": 2123104256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602206619859579,
      "loss": 1.3694,
      "theoretical_loss": 3.441909291567486,
      "tokens_seen": 2123169792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036020060180541624,
      "loss": 1.4121,
      "theoretical_loss": 3.4419006096903493,
      "tokens_seen": 2123235328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601805416248746,
      "loss": 1.2774,
      "theoretical_loss": 3.441891928156215,
      "tokens_seen": 2123300864
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360160481444333,
      "loss": 1.3135,
      "theoretical_loss": 3.441883246965058,
      "tokens_seen": 2123366400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601404212637914,
      "loss": 1.382,
      "theoretical_loss": 3.441874566116855,
      "tokens_seen": 2123431936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036012036108324974,
      "loss": 1.3478,
      "theoretical_loss": 3.4418658856115822,
      "tokens_seen": 2123497472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601003009027081,
      "loss": 1.3478,
      "theoretical_loss": 3.4418572054492147,
      "tokens_seen": 2123563008
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600802407221665,
      "loss": 1.2753,
      "theoretical_loss": 3.441848525629729,
      "tokens_seen": 2123628544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600601805416249,
      "loss": 1.3944,
      "theoretical_loss": 3.4418398461531003,
      "tokens_seen": 2123694080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036004012036108324,
      "loss": 1.3863,
      "theoretical_loss": 3.4418311670193047,
      "tokens_seen": 2123759616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600200601805416,
      "loss": 1.2979,
      "theoretical_loss": 3.441822488228319,
      "tokens_seen": 2123825152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035999999999999997,
      "loss": 1.2667,
      "theoretical_loss": 3.4418138097801174,
      "tokens_seen": 2123890688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003599799398194584,
      "loss": 1.3518,
      "theoretical_loss": 3.441805131674677,
      "tokens_seen": 2123956224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035995987963891675,
      "loss": 1.3924,
      "theoretical_loss": 3.4417964539119734,
      "tokens_seen": 2124021760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003599398194583751,
      "loss": 1.2942,
      "theoretical_loss": 3.4417877764919824,
      "tokens_seen": 2124087296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035991975927783347,
      "loss": 1.3718,
      "theoretical_loss": 3.44177909941468,
      "tokens_seen": 2124152832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598996990972919,
      "loss": 1.2736,
      "theoretical_loss": 3.441770422680042,
      "tokens_seen": 2124218368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035987963891675025,
      "loss": 1.4003,
      "theoretical_loss": 3.4417617462880443,
      "tokens_seen": 2124283904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598595787362086,
      "loss": 1.2612,
      "theoretical_loss": 3.4417530702386627,
      "tokens_seen": 2124349440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359839518555667,
      "loss": 1.3643,
      "theoretical_loss": 3.4417443945318738,
      "tokens_seen": 2124414976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035981945837512534,
      "loss": 1.5319,
      "theoretical_loss": 3.441735719167652,
      "tokens_seen": 2124480512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035979939819458375,
      "loss": 1.2494,
      "theoretical_loss": 3.441727044145975,
      "tokens_seen": 2124546048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003597793380140421,
      "loss": 1.2923,
      "theoretical_loss": 3.441718369466817,
      "tokens_seen": 2124611584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003597592778335005,
      "loss": 1.2791,
      "theoretical_loss": 3.4417096951301556,
      "tokens_seen": 2124677120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035973921765295884,
      "loss": 1.3473,
      "theoretical_loss": 3.4417010211359655,
      "tokens_seen": 2124742656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035971915747241726,
      "loss": 1.3139,
      "theoretical_loss": 3.4416923474842234,
      "tokens_seen": 2124808192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596990972918756,
      "loss": 1.3076,
      "theoretical_loss": 3.441683674174904,
      "tokens_seen": 2124873728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359679037111334,
      "loss": 1.2904,
      "theoretical_loss": 3.441675001207985,
      "tokens_seen": 2124939264
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035965897693079234,
      "loss": 1.3804,
      "theoretical_loss": 3.44166632858344,
      "tokens_seen": 2125004800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035963891675025076,
      "loss": 1.3686,
      "theoretical_loss": 3.4416576563012473,
      "tokens_seen": 2125070336
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 774338,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4190452098846436,
      "objective/train/theoretical_loss": 3.4416489843613816,
      "objective/train/tokens_used": 495594976,
      "theoretical_loss": 3.4416489843613816,
      "tokens_seen": 2125135872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596188565697091,
      "loss": 1.376,
      "theoretical_loss": 3.4416489843613816,
      "tokens_seen": 2125135872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035959879638916754,
      "loss": 1.3828,
      "theoretical_loss": 3.4416403127638193,
      "tokens_seen": 2125201408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003595787362086259,
      "loss": 1.3238,
      "theoretical_loss": 3.441631641508536,
      "tokens_seen": 2125266944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035955867602808426,
      "loss": 1.4609,
      "theoretical_loss": 3.4416229705955077,
      "tokens_seen": 2125332480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003595386158475427,
      "loss": 1.3779,
      "theoretical_loss": 3.44161430002471,
      "tokens_seen": 2125398016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035951855566700104,
      "loss": 1.2614,
      "theoretical_loss": 3.4416056297961193,
      "tokens_seen": 2125463552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594984954864594,
      "loss": 1.2702,
      "theoretical_loss": 3.4415969599097114,
      "tokens_seen": 2125529088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035947843530591777,
      "loss": 1.3084,
      "theoretical_loss": 3.4415882903654627,
      "tokens_seen": 2125594624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594583751253762,
      "loss": 1.3342,
      "theoretical_loss": 3.4415796211633483,
      "tokens_seen": 2125660160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035943831494483454,
      "loss": 1.3428,
      "theoretical_loss": 3.4415709523033446,
      "tokens_seen": 2125725696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594182547642929,
      "loss": 1.3003,
      "theoretical_loss": 3.441562283785428,
      "tokens_seen": 2125791232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035939819458375127,
      "loss": 1.4025,
      "theoretical_loss": 3.4415536156095734,
      "tokens_seen": 2125856768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035937813440320963,
      "loss": 1.4132,
      "theoretical_loss": 3.4415449477757574,
      "tokens_seen": 2125922304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035935807422266805,
      "loss": 1.3226,
      "theoretical_loss": 3.441536280283956,
      "tokens_seen": 2125987840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593380140421264,
      "loss": 1.34,
      "theoretical_loss": 3.441527613134145,
      "tokens_seen": 2126053376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035931795386158477,
      "loss": 1.3548,
      "theoretical_loss": 3.4415189463263007,
      "tokens_seen": 2126118912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035929789368104313,
      "loss": 1.277,
      "theoretical_loss": 3.441510279860399,
      "tokens_seen": 2126184448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035927783350050155,
      "loss": 1.3075,
      "theoretical_loss": 3.441501613736415,
      "tokens_seen": 2126249984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592577733199599,
      "loss": 1.3694,
      "theoretical_loss": 3.441492947954326,
      "tokens_seen": 2126315520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592377131394183,
      "loss": 1.2697,
      "theoretical_loss": 3.4414842825141063,
      "tokens_seen": 2126381056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035921765295887664,
      "loss": 1.304,
      "theoretical_loss": 3.4414756174157337,
      "tokens_seen": 2126446592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359197592778335,
      "loss": 1.3831,
      "theoretical_loss": 3.4414669526591832,
      "tokens_seen": 2126512128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591775325977934,
      "loss": 1.3481,
      "theoretical_loss": 3.441458288244431,
      "tokens_seen": 2126577664
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591574724172518,
      "loss": 1.3012,
      "theoretical_loss": 3.441449624171453,
      "tokens_seen": 2126643200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035913741223671014,
      "loss": 1.314,
      "theoretical_loss": 3.441440960440225,
      "tokens_seen": 2126708736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591173520561685,
      "loss": 1.3919,
      "theoretical_loss": 3.4414322970507234,
      "tokens_seen": 2126774272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590972918756269,
      "loss": 1.3416,
      "theoretical_loss": 3.4414236340029243,
      "tokens_seen": 2126839808
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590772316950853,
      "loss": 1.3243,
      "theoretical_loss": 3.441414971296803,
      "tokens_seen": 2126905344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035905717151454364,
      "loss": 1.3324,
      "theoretical_loss": 3.4414063089323363,
      "tokens_seen": 2126970880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359037111334002,
      "loss": 1.187,
      "theoretical_loss": 3.4413976469094996,
      "tokens_seen": 2127036416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035901705115346037,
      "loss": 1.3337,
      "theoretical_loss": 3.4413889852282686,
      "tokens_seen": 2127101952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589969909729188,
      "loss": 1.4321,
      "theoretical_loss": 3.4413803238886205,
      "tokens_seen": 2127167488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035897693079237715,
      "loss": 1.3863,
      "theoretical_loss": 3.44137166289053,
      "tokens_seen": 2127233024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589568706118355,
      "loss": 1.3384,
      "theoretical_loss": 3.4413630022339743,
      "tokens_seen": 2127298560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035893681043129387,
      "loss": 1.4916,
      "theoretical_loss": 3.4413543419189283,
      "tokens_seen": 2127364096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589167502507523,
      "loss": 1.3072,
      "theoretical_loss": 3.4413456819453687,
      "tokens_seen": 2127429632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035889669007021065,
      "loss": 1.3133,
      "theoretical_loss": 3.441337022313272,
      "tokens_seen": 2127495168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358876629889669,
      "loss": 1.2943,
      "theoretical_loss": 3.441328363022613,
      "tokens_seen": 2127560704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588565697091274,
      "loss": 1.3519,
      "theoretical_loss": 3.441319704073368,
      "tokens_seen": 2127626240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035883650952858574,
      "loss": 1.3547,
      "theoretical_loss": 3.441311045465514,
      "tokens_seen": 2127691776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035881644934804415,
      "loss": 1.4391,
      "theoretical_loss": 3.4413023871990256,
      "tokens_seen": 2127757312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587963891675025,
      "loss": 1.2724,
      "theoretical_loss": 3.44129372927388,
      "tokens_seen": 2127822848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587763289869609,
      "loss": 1.205,
      "theoretical_loss": 3.4412850716900527,
      "tokens_seen": 2127888384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035875626880641924,
      "loss": 1.3634,
      "theoretical_loss": 3.4412764144475196,
      "tokens_seen": 2127953920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035873620862587766,
      "loss": 1.2051,
      "theoretical_loss": 3.4412677575462576,
      "tokens_seen": 2128019456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358716148445336,
      "loss": 1.3038,
      "theoretical_loss": 3.441259100986241,
      "tokens_seen": 2128084992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586960882647944,
      "loss": 1.2973,
      "theoretical_loss": 3.441250444767448,
      "tokens_seen": 2128150528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035867602808425274,
      "loss": 1.4109,
      "theoretical_loss": 3.441241788889853,
      "tokens_seen": 2128216064
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035865596790371116,
      "loss": 1.2411,
      "theoretical_loss": 3.441233133353433,
      "tokens_seen": 2128281600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586359077231695,
      "loss": 1.2656,
      "theoretical_loss": 3.441224478158164,
      "tokens_seen": 2128347136
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 775219,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2397558689117432,
      "objective/train/theoretical_loss": 3.441215823304021,
      "objective/train/tokens_used": 498871776,
      "theoretical_loss": 3.441215823304021,
      "tokens_seen": 2128412672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586158475426279,
      "loss": 1.3053,
      "theoretical_loss": 3.441215823304021,
      "tokens_seen": 2128412672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035859578736208625,
      "loss": 1.4534,
      "theoretical_loss": 3.441207168790981,
      "tokens_seen": 2128478208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585757271815446,
      "loss": 1.3226,
      "theoretical_loss": 3.4411985146190203,
      "tokens_seen": 2128543744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358555667001003,
      "loss": 1.2199,
      "theoretical_loss": 3.4411898607881137,
      "tokens_seen": 2128609280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585356068204614,
      "loss": 1.4343,
      "theoretical_loss": 3.4411812072982384,
      "tokens_seen": 2128674816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035851554663991975,
      "loss": 1.321,
      "theoretical_loss": 3.4411725541493703,
      "tokens_seen": 2128740352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584954864593781,
      "loss": 1.3199,
      "theoretical_loss": 3.441163901341485,
      "tokens_seen": 2128805888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035847542627883653,
      "loss": 1.324,
      "theoretical_loss": 3.4411552488745594,
      "tokens_seen": 2128871424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584553660982949,
      "loss": 1.2769,
      "theoretical_loss": 3.441146596748568,
      "tokens_seen": 2128936960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035843530591775325,
      "loss": 1.2738,
      "theoretical_loss": 3.4411379449634887,
      "tokens_seen": 2129002496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584152457372116,
      "loss": 1.1917,
      "theoretical_loss": 3.4411292935192965,
      "tokens_seen": 2129068032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035839518555667,
      "loss": 1.2898,
      "theoretical_loss": 3.4411206424159677,
      "tokens_seen": 2129133568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583751253761284,
      "loss": 1.2502,
      "theoretical_loss": 3.4411119916534787,
      "tokens_seen": 2129199104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035835506519558676,
      "loss": 1.3959,
      "theoretical_loss": 3.4411033412318046,
      "tokens_seen": 2129264640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583350050150451,
      "loss": 1.4017,
      "theoretical_loss": 3.441094691150923,
      "tokens_seen": 2129330176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583149448345035,
      "loss": 1.3461,
      "theoretical_loss": 3.441086041410809,
      "tokens_seen": 2129395712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582948846539619,
      "loss": 1.4017,
      "theoretical_loss": 3.4410773920114384,
      "tokens_seen": 2129461248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035827482447342026,
      "loss": 1.2654,
      "theoretical_loss": 3.441068742952788,
      "tokens_seen": 2129526784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582547642928786,
      "loss": 1.4366,
      "theoretical_loss": 3.4410600942348335,
      "tokens_seen": 2129592320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358234704112337,
      "loss": 1.3301,
      "theoretical_loss": 3.4410514458575516,
      "tokens_seen": 2129657856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035821464393179535,
      "loss": 1.2691,
      "theoretical_loss": 3.4410427978209173,
      "tokens_seen": 2129723392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035819458375125376,
      "loss": 1.3385,
      "theoretical_loss": 3.4410341501249078,
      "tokens_seen": 2129788928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581745235707121,
      "loss": 1.3358,
      "theoretical_loss": 3.4410255027694987,
      "tokens_seen": 2129854464
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581544633901705,
      "loss": 1.3992,
      "theoretical_loss": 3.441016855754666,
      "tokens_seen": 2129920000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035813440320962885,
      "loss": 1.4394,
      "theoretical_loss": 3.441008209080386,
      "tokens_seen": 2129985536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035811434302908727,
      "loss": 1.4123,
      "theoretical_loss": 3.4409995627466348,
      "tokens_seen": 2130051072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580942828485457,
      "loss": 1.4891,
      "theoretical_loss": 3.440990916753388,
      "tokens_seen": 2130116608
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035807422266800404,
      "loss": 1.3302,
      "theoretical_loss": 3.440982271100623,
      "tokens_seen": 2130182144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580541624874624,
      "loss": 1.2888,
      "theoretical_loss": 3.440973625788314,
      "tokens_seen": 2130247680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035803410230692077,
      "loss": 1.2651,
      "theoretical_loss": 3.4409649808164393,
      "tokens_seen": 2130313216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580140421263792,
      "loss": 1.3167,
      "theoretical_loss": 3.440956336184974,
      "tokens_seen": 2130378752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035799398194583755,
      "loss": 1.3723,
      "theoretical_loss": 3.4409476918938937,
      "tokens_seen": 2130444288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003579739217652959,
      "loss": 1.423,
      "theoretical_loss": 3.440939047943175,
      "tokens_seen": 2130509824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035795386158475427,
      "loss": 1.3125,
      "theoretical_loss": 3.4409304043327937,
      "tokens_seen": 2130575360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003579338014042127,
      "loss": 1.422,
      "theoretical_loss": 3.4409217610627265,
      "tokens_seen": 2130640896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035791374122367105,
      "loss": 1.3242,
      "theoretical_loss": 3.4409131181329498,
      "tokens_seen": 2130706432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578936810431294,
      "loss": 1.4027,
      "theoretical_loss": 3.4409044755434386,
      "tokens_seen": 2130771968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578736208625878,
      "loss": 1.3327,
      "theoretical_loss": 3.44089583329417,
      "tokens_seen": 2130837504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035785356068204614,
      "loss": 1.2811,
      "theoretical_loss": 3.4408871913851193,
      "tokens_seen": 2130903040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035783350050150455,
      "loss": 1.3479,
      "theoretical_loss": 3.4408785498162637,
      "tokens_seen": 2130968576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578134403209629,
      "loss": 1.197,
      "theoretical_loss": 3.4408699085875787,
      "tokens_seen": 2131034112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577933801404213,
      "loss": 1.3662,
      "theoretical_loss": 3.44086126769904,
      "tokens_seen": 2131099648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035777331995987964,
      "loss": 1.272,
      "theoretical_loss": 3.4408526271506252,
      "tokens_seen": 2131165184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035775325977933806,
      "loss": 1.2617,
      "theoretical_loss": 3.4408439869423093,
      "tokens_seen": 2131230720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577331995987964,
      "loss": 1.3017,
      "theoretical_loss": 3.440835347074068,
      "tokens_seen": 2131296256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577131394182548,
      "loss": 1.2578,
      "theoretical_loss": 3.440826707545879,
      "tokens_seen": 2131361792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035769307923771314,
      "loss": 1.2312,
      "theoretical_loss": 3.4408180683577174,
      "tokens_seen": 2131427328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035767301905717156,
      "loss": 1.3423,
      "theoretical_loss": 3.4408094295095593,
      "tokens_seen": 2131492864
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576529588766299,
      "loss": 1.2767,
      "theoretical_loss": 3.4408007910013816,
      "tokens_seen": 2131558400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576328986960883,
      "loss": 1.3621,
      "theoretical_loss": 3.4407921528331595,
      "tokens_seen": 2131623936
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 776312,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.494091510772705,
      "objective/train/theoretical_loss": 3.4407835150048705,
      "objective/train/tokens_used": 502148576,
      "theoretical_loss": 3.4407835150048705,
      "tokens_seen": 2131689472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035761283851554665,
      "loss": 1.4551,
      "theoretical_loss": 3.4407835150048705,
      "tokens_seen": 2131689472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357592778335005,
      "loss": 1.2621,
      "theoretical_loss": 3.440774877516489,
      "tokens_seen": 2131755008
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575727181544634,
      "loss": 1.4304,
      "theoretical_loss": 3.440766240367993,
      "tokens_seen": 2131820544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575526579739218,
      "loss": 1.2241,
      "theoretical_loss": 3.4407576035593572,
      "tokens_seen": 2131886080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035753259779338015,
      "loss": 1.3136,
      "theoretical_loss": 3.440748967090559,
      "tokens_seen": 2131951616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575125376128385,
      "loss": 1.3389,
      "theoretical_loss": 3.440740330961573,
      "tokens_seen": 2132017152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035749247743229693,
      "loss": 1.3099,
      "theoretical_loss": 3.440731695172377,
      "tokens_seen": 2132082688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003574724172517553,
      "loss": 1.3085,
      "theoretical_loss": 3.440723059722947,
      "tokens_seen": 2132148224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035745235707121365,
      "loss": 1.3017,
      "theoretical_loss": 3.440714424613258,
      "tokens_seen": 2132213760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357432296890672,
      "loss": 1.3108,
      "theoretical_loss": 3.440705789843287,
      "tokens_seen": 2132279296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003574122367101304,
      "loss": 1.4492,
      "theoretical_loss": 3.4406971554130106,
      "tokens_seen": 2132344832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573921765295888,
      "loss": 1.3548,
      "theoretical_loss": 3.440688521322404,
      "tokens_seen": 2132410368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035737211634904716,
      "loss": 1.5501,
      "theoretical_loss": 3.440679887571444,
      "tokens_seen": 2132475904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573520561685055,
      "loss": 1.451,
      "theoretical_loss": 3.440671254160107,
      "tokens_seen": 2132541440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573319959879639,
      "loss": 1.3102,
      "theoretical_loss": 3.440662621088369,
      "tokens_seen": 2132606976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573119358074223,
      "loss": 1.3613,
      "theoretical_loss": 3.440653988356206,
      "tokens_seen": 2132672512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035729187562688066,
      "loss": 1.3527,
      "theoretical_loss": 3.440645355963594,
      "tokens_seen": 2132738048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357271815446339,
      "loss": 1.352,
      "theoretical_loss": 3.44063672391051,
      "tokens_seen": 2132803584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572517552657974,
      "loss": 1.3763,
      "theoretical_loss": 3.44062809219693,
      "tokens_seen": 2132869120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035723169508525575,
      "loss": 1.2878,
      "theoretical_loss": 3.4406194608228295,
      "tokens_seen": 2132934656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035721163490471416,
      "loss": 1.324,
      "theoretical_loss": 3.440610829788185,
      "tokens_seen": 2133000192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571915747241725,
      "loss": 1.1916,
      "theoretical_loss": 3.4406021990929734,
      "tokens_seen": 2133065728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571715145436309,
      "loss": 1.3533,
      "theoretical_loss": 3.44059356873717,
      "tokens_seen": 2133131264
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035715145436308925,
      "loss": 1.4111,
      "theoretical_loss": 3.440584938720752,
      "tokens_seen": 2133196800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035713139418254767,
      "loss": 1.3214,
      "theoretical_loss": 3.4405763090436947,
      "tokens_seen": 2133262336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035711133400200603,
      "loss": 1.294,
      "theoretical_loss": 3.4405676797059748,
      "tokens_seen": 2133327872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003570912738214644,
      "loss": 1.3682,
      "theoretical_loss": 3.440559050707569,
      "tokens_seen": 2133393408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035707121364092275,
      "loss": 1.3797,
      "theoretical_loss": 3.440550422048452,
      "tokens_seen": 2133458944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003570511534603811,
      "loss": 1.3337,
      "theoretical_loss": 3.440541793728602,
      "tokens_seen": 2133524480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035703109327983953,
      "loss": 1.3732,
      "theoretical_loss": 3.440533165747994,
      "tokens_seen": 2133590016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003570110330992979,
      "loss": 1.2371,
      "theoretical_loss": 3.4405245381066045,
      "tokens_seen": 2133655552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035699097291875626,
      "loss": 1.3211,
      "theoretical_loss": 3.440515910804409,
      "tokens_seen": 2133721088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003569709127382146,
      "loss": 1.353,
      "theoretical_loss": 3.4405072838413853,
      "tokens_seen": 2133786624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035695085255767303,
      "loss": 1.351,
      "theoretical_loss": 3.4404986572175087,
      "tokens_seen": 2133852160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003569307923771314,
      "loss": 1.4066,
      "theoretical_loss": 3.440490030932756,
      "tokens_seen": 2133917696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035691073219658976,
      "loss": 1.2886,
      "theoretical_loss": 3.4404814049871026,
      "tokens_seen": 2133983232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568906720160481,
      "loss": 1.3413,
      "theoretical_loss": 3.440472779380525,
      "tokens_seen": 2134048768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568706118355065,
      "loss": 1.4214,
      "theoretical_loss": 3.440464154113,
      "tokens_seen": 2134114304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568505516549649,
      "loss": 1.3022,
      "theoretical_loss": 3.440455529184504,
      "tokens_seen": 2134179840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035683049147442326,
      "loss": 1.3155,
      "theoretical_loss": 3.440446904595012,
      "tokens_seen": 2134245376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568104312938816,
      "loss": 1.259,
      "theoretical_loss": 3.4404382803445017,
      "tokens_seen": 2134310912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035679037111334,
      "loss": 1.2681,
      "theoretical_loss": 3.4404296564329484,
      "tokens_seen": 2134376448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567703109327984,
      "loss": 1.2632,
      "theoretical_loss": 3.440421032860329,
      "tokens_seen": 2134441984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035675025075225677,
      "loss": 1.4561,
      "theoretical_loss": 3.440412409626619,
      "tokens_seen": 2134507520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035673019057171513,
      "loss": 1.4086,
      "theoretical_loss": 3.440403786731795,
      "tokens_seen": 2134573056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567101303911735,
      "loss": 1.369,
      "theoretical_loss": 3.4403951641758344,
      "tokens_seen": 2134638592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035669007021063185,
      "loss": 1.2481,
      "theoretical_loss": 3.440386541958712,
      "tokens_seen": 2134704128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035667001003009027,
      "loss": 1.3134,
      "theoretical_loss": 3.4403779200804046,
      "tokens_seen": 2134769664
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035664994984954863,
      "loss": 1.3466,
      "theoretical_loss": 3.4403692985408885,
      "tokens_seen": 2134835200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000356629889669007,
      "loss": 1.3554,
      "theoretical_loss": 3.4403606773401396,
      "tokens_seen": 2134900736
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 777595,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3070225715637207,
      "objective/train/theoretical_loss": 3.440352056478135,
      "objective/train/tokens_used": 505425376,
      "theoretical_loss": 3.440352056478135,
      "tokens_seen": 2134966272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035660982948846536,
      "loss": 1.3461,
      "theoretical_loss": 3.440352056478135,
      "tokens_seen": 2134966272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565897693079238,
      "loss": 1.3112,
      "theoretical_loss": 3.4403434359548504,
      "tokens_seen": 2135031808
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565697091273822,
      "loss": 1.2702,
      "theoretical_loss": 3.4403348157702625,
      "tokens_seen": 2135097344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035654964894684055,
      "loss": 1.4314,
      "theoretical_loss": 3.440326195924347,
      "tokens_seen": 2135162880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565295887662989,
      "loss": 1.3201,
      "theoretical_loss": 3.4403175764170806,
      "tokens_seen": 2135228416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035650952858575733,
      "loss": 1.2803,
      "theoretical_loss": 3.4403089572484395,
      "tokens_seen": 2135293952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564894684052157,
      "loss": 1.3105,
      "theoretical_loss": 3.4403003384184005,
      "tokens_seen": 2135359488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035646940822467405,
      "loss": 1.2803,
      "theoretical_loss": 3.4402917199269396,
      "tokens_seen": 2135425024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564493480441324,
      "loss": 1.2174,
      "theoretical_loss": 3.4402831017740323,
      "tokens_seen": 2135490560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564292878635908,
      "loss": 1.3404,
      "theoretical_loss": 3.440274483959656,
      "tokens_seen": 2135556096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564092276830492,
      "loss": 1.3267,
      "theoretical_loss": 3.440265866483786,
      "tokens_seen": 2135621632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035638916750250756,
      "loss": 1.4911,
      "theoretical_loss": 3.4402572493464,
      "tokens_seen": 2135687168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563691073219659,
      "loss": 1.1375,
      "theoretical_loss": 3.440248632547473,
      "tokens_seen": 2135752704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563490471414243,
      "loss": 1.2635,
      "theoretical_loss": 3.440240016086982,
      "tokens_seen": 2135818240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563289869608827,
      "loss": 1.2706,
      "theoretical_loss": 3.4402313999649037,
      "tokens_seen": 2135883776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035630892678034106,
      "loss": 1.2511,
      "theoretical_loss": 3.440222784181213,
      "tokens_seen": 2135949312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562888665997994,
      "loss": 1.3204,
      "theoretical_loss": 3.4402141687358876,
      "tokens_seen": 2136014848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562688064192578,
      "loss": 1.4885,
      "theoretical_loss": 3.4402055536289033,
      "tokens_seen": 2136080384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035624874623871615,
      "loss": 1.3531,
      "theoretical_loss": 3.4401969388602365,
      "tokens_seen": 2136145920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035622868605817456,
      "loss": 1.1891,
      "theoretical_loss": 3.4401883244298634,
      "tokens_seen": 2136211456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562086258776329,
      "loss": 1.3732,
      "theoretical_loss": 3.4401797103377607,
      "tokens_seen": 2136276992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003561885656970913,
      "loss": 1.2878,
      "theoretical_loss": 3.4401710965839047,
      "tokens_seen": 2136342528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035616850551654965,
      "loss": 1.296,
      "theoretical_loss": 3.440162483168271,
      "tokens_seen": 2136408064
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035614844533600807,
      "loss": 1.2962,
      "theoretical_loss": 3.440153870090837,
      "tokens_seen": 2136473600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035612838515546643,
      "loss": 1.3415,
      "theoretical_loss": 3.4401452573515785,
      "tokens_seen": 2136539136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003561083249749248,
      "loss": 1.3779,
      "theoretical_loss": 3.4401366449504716,
      "tokens_seen": 2136604672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035608826479438315,
      "loss": 1.2601,
      "theoretical_loss": 3.440128032887493,
      "tokens_seen": 2136670208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003560682046138415,
      "loss": 1.2638,
      "theoretical_loss": 3.4401194211626187,
      "tokens_seen": 2136735744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035604814443329993,
      "loss": 1.3308,
      "theoretical_loss": 3.4401108097758257,
      "tokens_seen": 2136801280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003560280842527583,
      "loss": 1.2922,
      "theoretical_loss": 3.4401021987270903,
      "tokens_seen": 2136866816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035600802407221666,
      "loss": 1.3145,
      "theoretical_loss": 3.440093588016388,
      "tokens_seen": 2136932352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355987963891675,
      "loss": 1.4025,
      "theoretical_loss": 3.4400849776436964,
      "tokens_seen": 2136997888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035596790371113344,
      "loss": 1.2907,
      "theoretical_loss": 3.4400763676089907,
      "tokens_seen": 2137063424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003559478435305918,
      "loss": 1.2305,
      "theoretical_loss": 3.4400677579122476,
      "tokens_seen": 2137128960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035592778335005016,
      "loss": 1.3397,
      "theoretical_loss": 3.440059148553444,
      "tokens_seen": 2137194496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003559077231695085,
      "loss": 1.2745,
      "theoretical_loss": 3.440050539532556,
      "tokens_seen": 2137260032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558876629889669,
      "loss": 1.3253,
      "theoretical_loss": 3.44004193084956,
      "tokens_seen": 2137325568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558676028084253,
      "loss": 1.3965,
      "theoretical_loss": 3.4400333225044317,
      "tokens_seen": 2137391104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035584754262788366,
      "loss": 1.4965,
      "theoretical_loss": 3.4400247144971488,
      "tokens_seen": 2137456640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355827482447342,
      "loss": 1.3132,
      "theoretical_loss": 3.440016106827686,
      "tokens_seen": 2137522176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558074222668004,
      "loss": 1.3446,
      "theoretical_loss": 3.440007499496021,
      "tokens_seen": 2137587712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003557873620862588,
      "loss": 1.2975,
      "theoretical_loss": 3.43999889250213,
      "tokens_seen": 2137653248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035576730190571717,
      "loss": 1.3259,
      "theoretical_loss": 3.4399902858459894,
      "tokens_seen": 2137718784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035574724172517553,
      "loss": 1.4234,
      "theoretical_loss": 3.439981679527575,
      "tokens_seen": 2137784320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003557271815446339,
      "loss": 1.2532,
      "theoretical_loss": 3.439973073546864,
      "tokens_seen": 2137849856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035570712136409225,
      "loss": 1.3639,
      "theoretical_loss": 3.439964467903832,
      "tokens_seen": 2137915392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035568706118355067,
      "loss": 1.3525,
      "theoretical_loss": 3.439955862598456,
      "tokens_seen": 2137980928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035566700100300903,
      "loss": 1.291,
      "theoretical_loss": 3.4399472576307124,
      "tokens_seen": 2138046464
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556469408224674,
      "loss": 1.356,
      "theoretical_loss": 3.439938653000577,
      "tokens_seen": 2138112000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035562688064192576,
      "loss": 1.4528,
      "theoretical_loss": 3.4399300487080264,
      "tokens_seen": 2138177536
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 778668,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3252495527267456,
      "objective/train/theoretical_loss": 3.4399214447530375,
      "objective/train/tokens_used": 508702176,
      "theoretical_loss": 3.4399214447530375,
      "tokens_seen": 2138243072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035560682046138417,
      "loss": 1.3317,
      "theoretical_loss": 3.4399214447530375,
      "tokens_seen": 2138243072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035558676028084253,
      "loss": 1.3404,
      "theoretical_loss": 3.4399128411355866,
      "tokens_seen": 2138308608
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003555667001003009,
      "loss": 1.2442,
      "theoretical_loss": 3.4399042378556497,
      "tokens_seen": 2138374144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035554663991975926,
      "loss": 1.3616,
      "theoretical_loss": 3.4398956349132037,
      "tokens_seen": 2138439680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003555265797392176,
      "loss": 1.2863,
      "theoretical_loss": 3.4398870323082242,
      "tokens_seen": 2138505216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035550651955867604,
      "loss": 1.3091,
      "theoretical_loss": 3.4398784300406886,
      "tokens_seen": 2138570752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554864593781344,
      "loss": 1.1826,
      "theoretical_loss": 3.439869828110573,
      "tokens_seen": 2138636288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035546639919759276,
      "loss": 1.3143,
      "theoretical_loss": 3.4398612265178534,
      "tokens_seen": 2138701824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554463390170511,
      "loss": 1.3747,
      "theoretical_loss": 3.439852625262507,
      "tokens_seen": 2138767360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035542627883650954,
      "loss": 1.1323,
      "theoretical_loss": 3.4398440243445094,
      "tokens_seen": 2138832896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554062186559679,
      "loss": 1.2308,
      "theoretical_loss": 3.4398354237638378,
      "tokens_seen": 2138898432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035538615847542627,
      "loss": 1.2043,
      "theoretical_loss": 3.439826823520468,
      "tokens_seen": 2138963968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035536609829488463,
      "loss": 1.3601,
      "theoretical_loss": 3.4398182236143766,
      "tokens_seen": 2139029504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035534603811434304,
      "loss": 1.3758,
      "theoretical_loss": 3.43980962404554,
      "tokens_seen": 2139095040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553259779338014,
      "loss": 1.4445,
      "theoretical_loss": 3.439801024813935,
      "tokens_seen": 2139160576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035530591775325977,
      "loss": 1.2661,
      "theoretical_loss": 3.439792425919538,
      "tokens_seen": 2139226112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035528585757271813,
      "loss": 1.3666,
      "theoretical_loss": 3.439783827362325,
      "tokens_seen": 2139291648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552657973921765,
      "loss": 1.3466,
      "theoretical_loss": 3.439775229142273,
      "tokens_seen": 2139357184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552457372116349,
      "loss": 1.3266,
      "theoretical_loss": 3.439766631259358,
      "tokens_seen": 2139422720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035522567703109327,
      "loss": 1.3053,
      "theoretical_loss": 3.4397580337135567,
      "tokens_seen": 2139488256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035520561685055163,
      "loss": 1.256,
      "theoretical_loss": 3.4397494365048455,
      "tokens_seen": 2139553792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035518555667001,
      "loss": 1.3325,
      "theoretical_loss": 3.4397408396332008,
      "tokens_seen": 2139619328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551654964894684,
      "loss": 1.322,
      "theoretical_loss": 3.4397322430985993,
      "tokens_seen": 2139684864
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551454363089268,
      "loss": 1.331,
      "theoretical_loss": 3.439723646901017,
      "tokens_seen": 2139750400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035512537612838514,
      "loss": 1.2893,
      "theoretical_loss": 3.43971505104043,
      "tokens_seen": 2139815936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551053159478435,
      "loss": 1.2141,
      "theoretical_loss": 3.4397064555168164,
      "tokens_seen": 2139881472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550852557673019,
      "loss": 1.348,
      "theoretical_loss": 3.439697860330152,
      "tokens_seen": 2139947008
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035506519558676033,
      "loss": 1.3245,
      "theoretical_loss": 3.439689265480412,
      "tokens_seen": 2140012544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550451354062187,
      "loss": 1.3246,
      "theoretical_loss": 3.439680670967574,
      "tokens_seen": 2140078080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035502507522567706,
      "loss": 1.3225,
      "theoretical_loss": 3.4396720767916147,
      "tokens_seen": 2140143616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550050150451354,
      "loss": 1.3036,
      "theoretical_loss": 3.4396634829525095,
      "tokens_seen": 2140209152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035498495486459384,
      "loss": 1.2314,
      "theoretical_loss": 3.439654889450236,
      "tokens_seen": 2140274688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549648946840522,
      "loss": 1.3085,
      "theoretical_loss": 3.4396462962847707,
      "tokens_seen": 2140340224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035494483450351056,
      "loss": 1.3708,
      "theoretical_loss": 3.4396377034560888,
      "tokens_seen": 2140405760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549247743229689,
      "loss": 1.2907,
      "theoretical_loss": 3.4396291109641677,
      "tokens_seen": 2140471296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549047141424273,
      "loss": 1.3083,
      "theoretical_loss": 3.4396205188089843,
      "tokens_seen": 2140536832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548846539618857,
      "loss": 1.2757,
      "theoretical_loss": 3.4396119269905143,
      "tokens_seen": 2140602368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035486459378134406,
      "loss": 1.2771,
      "theoretical_loss": 3.4396033355087345,
      "tokens_seen": 2140667904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548445336008024,
      "loss": 1.3275,
      "theoretical_loss": 3.439594744363622,
      "tokens_seen": 2140733440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548244734202608,
      "loss": 1.3031,
      "theoretical_loss": 3.439586153555152,
      "tokens_seen": 2140798976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548044132397192,
      "loss": 1.3226,
      "theoretical_loss": 3.4395775630833016,
      "tokens_seen": 2140864512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035478435305917757,
      "loss": 1.258,
      "theoretical_loss": 3.439568972948048,
      "tokens_seen": 2140930048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035476429287863593,
      "loss": 1.3212,
      "theoretical_loss": 3.439560383149367,
      "tokens_seen": 2140995584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003547442326980943,
      "loss": 1.2982,
      "theoretical_loss": 3.439551793687235,
      "tokens_seen": 2141061120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035472417251755265,
      "loss": 1.2696,
      "theoretical_loss": 3.439543204561629,
      "tokens_seen": 2141126656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035470411233701107,
      "loss": 1.3041,
      "theoretical_loss": 3.439534615772525,
      "tokens_seen": 2141192192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035468405215646943,
      "loss": 1.3282,
      "theoretical_loss": 3.4395260273199,
      "tokens_seen": 2141257728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546639919759278,
      "loss": 1.3522,
      "theoretical_loss": 3.4395174392037298,
      "tokens_seen": 2141323264
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035464393179538616,
      "loss": 1.3441,
      "theoretical_loss": 3.439508851423992,
      "tokens_seen": 2141388800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035462387161484457,
      "loss": 1.3356,
      "theoretical_loss": 3.4395002639806624,
      "tokens_seen": 2141454336
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 779818,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.398399829864502,
      "objective/train/theoretical_loss": 3.4394916768737174,
      "objective/train/tokens_used": 511978976,
      "theoretical_loss": 3.4394916768737174,
      "tokens_seen": 2141519872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035460381143430293,
      "loss": 1.3089,
      "theoretical_loss": 3.4394916768737174,
      "tokens_seen": 2141519872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545837512537613,
      "loss": 1.287,
      "theoretical_loss": 3.4394830901031344,
      "tokens_seen": 2141585408
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035456369107321966,
      "loss": 1.3604,
      "theoretical_loss": 3.4394745036688885,
      "tokens_seen": 2141650944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354543630892678,
      "loss": 1.2881,
      "theoretical_loss": 3.439465917570958,
      "tokens_seen": 2141716480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035452357071213644,
      "loss": 1.3275,
      "theoretical_loss": 3.439457331809318,
      "tokens_seen": 2141782016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545035105315948,
      "loss": 1.3962,
      "theoretical_loss": 3.4394487463839454,
      "tokens_seen": 2141847552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035448345035105316,
      "loss": 1.2574,
      "theoretical_loss": 3.4394401612948173,
      "tokens_seen": 2141913088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544633901705115,
      "loss": 1.4146,
      "theoretical_loss": 3.4394315765419097,
      "tokens_seen": 2141978624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035444332998996994,
      "loss": 1.1949,
      "theoretical_loss": 3.4394229921251993,
      "tokens_seen": 2142044160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544232698094283,
      "loss": 1.4292,
      "theoretical_loss": 3.4394144080446623,
      "tokens_seen": 2142109696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035440320962888667,
      "loss": 1.2901,
      "theoretical_loss": 3.439405824300276,
      "tokens_seen": 2142175232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035438314944834503,
      "loss": 1.318,
      "theoretical_loss": 3.439397240892016,
      "tokens_seen": 2142240768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035436308926780344,
      "loss": 1.3866,
      "theoretical_loss": 3.43938865781986,
      "tokens_seen": 2142306304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003543430290872618,
      "loss": 1.4974,
      "theoretical_loss": 3.4393800750837835,
      "tokens_seen": 2142371840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035432296890672017,
      "loss": 1.2849,
      "theoretical_loss": 3.4393714926837635,
      "tokens_seen": 2142437376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035430290872617853,
      "loss": 1.2741,
      "theoretical_loss": 3.439362910619777,
      "tokens_seen": 2142502912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542828485456369,
      "loss": 1.2831,
      "theoretical_loss": 3.4393543288918,
      "tokens_seen": 2142568448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542627883650953,
      "loss": 1.2862,
      "theoretical_loss": 3.439345747499809,
      "tokens_seen": 2142633984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035424272818455367,
      "loss": 1.3534,
      "theoretical_loss": 3.4393371664437806,
      "tokens_seen": 2142699520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035422266800401203,
      "loss": 1.2924,
      "theoretical_loss": 3.4393285857236915,
      "tokens_seen": 2142765056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542026078234704,
      "loss": 1.3513,
      "theoretical_loss": 3.4393200053395185,
      "tokens_seen": 2142830592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541825476429288,
      "loss": 1.3135,
      "theoretical_loss": 3.4393114252912387,
      "tokens_seen": 2142896128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541624874623872,
      "loss": 1.2822,
      "theoretical_loss": 3.439302845578827,
      "tokens_seen": 2142961664
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035414242728184554,
      "loss": 1.3528,
      "theoretical_loss": 3.4392942662022614,
      "tokens_seen": 2143027200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541223671013039,
      "loss": 1.2919,
      "theoretical_loss": 3.439285687161518,
      "tokens_seen": 2143092736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035410230692076226,
      "loss": 1.2921,
      "theoretical_loss": 3.439277108456573,
      "tokens_seen": 2143158272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003540822467402207,
      "loss": 1.3208,
      "theoretical_loss": 3.439268530087404,
      "tokens_seen": 2143223808
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035406218655967904,
      "loss": 1.4625,
      "theoretical_loss": 3.4392599520539866,
      "tokens_seen": 2143289344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003540421263791374,
      "loss": 1.3268,
      "theoretical_loss": 3.439251374356298,
      "tokens_seen": 2143354880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035402206619859577,
      "loss": 1.1718,
      "theoretical_loss": 3.4392427969943147,
      "tokens_seen": 2143420416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003540020060180542,
      "loss": 1.3166,
      "theoretical_loss": 3.4392342199680126,
      "tokens_seen": 2143485952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035398194583751254,
      "loss": 1.2865,
      "theoretical_loss": 3.4392256432773696,
      "tokens_seen": 2143551488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003539618856569709,
      "loss": 1.3057,
      "theoretical_loss": 3.439217066922361,
      "tokens_seen": 2143617024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035394182547642927,
      "loss": 1.4133,
      "theoretical_loss": 3.4392084909029643,
      "tokens_seen": 2143682560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035392176529588763,
      "loss": 1.3409,
      "theoretical_loss": 3.439199915219156,
      "tokens_seen": 2143748096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035390170511534605,
      "loss": 1.3532,
      "theoretical_loss": 3.4391913398709124,
      "tokens_seen": 2143813632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538816449348044,
      "loss": 1.2113,
      "theoretical_loss": 3.4391827648582103,
      "tokens_seen": 2143879168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035386158475426277,
      "loss": 1.3304,
      "theoretical_loss": 3.439174190181026,
      "tokens_seen": 2143944704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035384152457372113,
      "loss": 1.2251,
      "theoretical_loss": 3.439165615839336,
      "tokens_seen": 2144010240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035382146439317955,
      "loss": 1.3451,
      "theoretical_loss": 3.439157041833118,
      "tokens_seen": 2144075776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538014042126379,
      "loss": 1.3919,
      "theoretical_loss": 3.4391484681623474,
      "tokens_seen": 2144141312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537813440320963,
      "loss": 1.3531,
      "theoretical_loss": 3.4391398948270018,
      "tokens_seen": 2144206848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035376128385155464,
      "loss": 1.3001,
      "theoretical_loss": 3.439131321827057,
      "tokens_seen": 2144272384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353741223671013,
      "loss": 1.2921,
      "theoretical_loss": 3.4391227491624896,
      "tokens_seen": 2144337920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537211634904714,
      "loss": 1.1948,
      "theoretical_loss": 3.4391141768332774,
      "tokens_seen": 2144403456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537011033099298,
      "loss": 1.3473,
      "theoretical_loss": 3.439105604839396,
      "tokens_seen": 2144468992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035368104312938814,
      "loss": 1.3074,
      "theoretical_loss": 3.439097033180822,
      "tokens_seen": 2144534528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536609829488465,
      "loss": 1.2737,
      "theoretical_loss": 3.4390884618575326,
      "tokens_seen": 2144600064
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536409227683049,
      "loss": 1.2875,
      "theoretical_loss": 3.4390798908695035,
      "tokens_seen": 2144665600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536208625877633,
      "loss": 1.3816,
      "theoretical_loss": 3.439071320216713,
      "tokens_seen": 2144731136
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 780961,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3549442291259766,
      "objective/train/theoretical_loss": 3.4390627498991355,
      "objective/train/tokens_used": 515255776,
      "theoretical_loss": 3.4390627498991355,
      "tokens_seen": 2144796672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536008024072217,
      "loss": 1.344,
      "theoretical_loss": 3.4390627498991355,
      "tokens_seen": 2144796672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035358074222668006,
      "loss": 1.2742,
      "theoretical_loss": 3.43905417991675,
      "tokens_seen": 2144862208
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535606820461384,
      "loss": 1.3611,
      "theoretical_loss": 3.4390456102695315,
      "tokens_seen": 2144927744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035354062186559684,
      "loss": 1.3172,
      "theoretical_loss": 3.4390370409574573,
      "tokens_seen": 2144993280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535205616850552,
      "loss": 1.331,
      "theoretical_loss": 3.439028471980504,
      "tokens_seen": 2145058816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035350050150451356,
      "loss": 1.3048,
      "theoretical_loss": 3.439019903338648,
      "tokens_seen": 2145124352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534804413239719,
      "loss": 1.4079,
      "theoretical_loss": 3.439011335031866,
      "tokens_seen": 2145189888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035346038114343034,
      "loss": 1.2809,
      "theoretical_loss": 3.4390027670601353,
      "tokens_seen": 2145255424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534403209628887,
      "loss": 1.2763,
      "theoretical_loss": 3.4389941994234317,
      "tokens_seen": 2145320960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035342026078234707,
      "loss": 1.2864,
      "theoretical_loss": 3.4389856321217325,
      "tokens_seen": 2145386496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035340020060180543,
      "loss": 1.2932,
      "theoretical_loss": 3.4389770651550142,
      "tokens_seen": 2145452032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035338014042126384,
      "loss": 1.3216,
      "theoretical_loss": 3.438968498523253,
      "tokens_seen": 2145517568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003533600802407222,
      "loss": 1.3375,
      "theoretical_loss": 3.4389599322264264,
      "tokens_seen": 2145583104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035334002006018057,
      "loss": 1.3485,
      "theoretical_loss": 3.43895136626451,
      "tokens_seen": 2145648640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035331995987963893,
      "loss": 1.3139,
      "theoretical_loss": 3.4389428006374816,
      "tokens_seen": 2145714176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532998996990973,
      "loss": 1.4077,
      "theoretical_loss": 3.438934235345317,
      "tokens_seen": 2145779712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532798395185557,
      "loss": 1.3511,
      "theoretical_loss": 3.4389256703879942,
      "tokens_seen": 2145845248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035325977933801407,
      "loss": 1.4069,
      "theoretical_loss": 3.438917105765488,
      "tokens_seen": 2145910784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035323971915747243,
      "loss": 1.4001,
      "theoretical_loss": 3.438908541477776,
      "tokens_seen": 2145976320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532196589769308,
      "loss": 1.2601,
      "theoretical_loss": 3.4388999775248354,
      "tokens_seen": 2146041856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531995987963892,
      "loss": 1.2556,
      "theoretical_loss": 3.438891413906642,
      "tokens_seen": 2146107392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531795386158476,
      "loss": 1.3329,
      "theoretical_loss": 3.4388828506231737,
      "tokens_seen": 2146172928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035315947843530594,
      "loss": 1.3569,
      "theoretical_loss": 3.438874287674406,
      "tokens_seen": 2146238464
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531394182547643,
      "loss": 1.2781,
      "theoretical_loss": 3.4388657250603156,
      "tokens_seen": 2146304000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035311935807422266,
      "loss": 1.3,
      "theoretical_loss": 3.43885716278088,
      "tokens_seen": 2146369536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530992978936811,
      "loss": 1.339,
      "theoretical_loss": 3.4388486008360752,
      "tokens_seen": 2146435072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035307923771313944,
      "loss": 1.3521,
      "theoretical_loss": 3.4388400392258784,
      "tokens_seen": 2146500608
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530591775325978,
      "loss": 1.282,
      "theoretical_loss": 3.4388314779502664,
      "tokens_seen": 2146566144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035303911735205617,
      "loss": 1.3233,
      "theoretical_loss": 3.4388229170092153,
      "tokens_seen": 2146631680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530190571715146,
      "loss": 1.3275,
      "theoretical_loss": 3.4388143564027023,
      "tokens_seen": 2146697216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035299899699097294,
      "loss": 1.3534,
      "theoretical_loss": 3.438805796130704,
      "tokens_seen": 2146762752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003529789368104313,
      "loss": 1.3265,
      "theoretical_loss": 3.4387972361931967,
      "tokens_seen": 2146828288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035295887662988967,
      "loss": 1.2809,
      "theoretical_loss": 3.438788676590158,
      "tokens_seen": 2146893824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035293881644934803,
      "loss": 1.3095,
      "theoretical_loss": 3.4387801173215635,
      "tokens_seen": 2146959360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035291875626880645,
      "loss": 1.3546,
      "theoretical_loss": 3.4387715583873915,
      "tokens_seen": 2147024896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003528986960882648,
      "loss": 1.3254,
      "theoretical_loss": 3.438762999787617,
      "tokens_seen": 2147090432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035287863590772317,
      "loss": 1.3375,
      "theoretical_loss": 3.4387544415222178,
      "tokens_seen": 2147155968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035285857572718153,
      "loss": 1.2386,
      "theoretical_loss": 3.43874588359117,
      "tokens_seen": 2147221504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035283851554663995,
      "loss": 1.2849,
      "theoretical_loss": 3.4387373259944507,
      "tokens_seen": 2147287040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003528184553660983,
      "loss": 1.2986,
      "theoretical_loss": 3.4387287687320374,
      "tokens_seen": 2147352576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527983951855567,
      "loss": 1.3696,
      "theoretical_loss": 3.438720211803905,
      "tokens_seen": 2147418112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035277833500501504,
      "loss": 1.3382,
      "theoretical_loss": 3.4387116552100316,
      "tokens_seen": 2147483648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527582748244734,
      "loss": 1.2614,
      "theoretical_loss": 3.438703098950394,
      "tokens_seen": 2147549184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527382146439318,
      "loss": 1.3767,
      "theoretical_loss": 3.4386945430249685,
      "tokens_seen": 2147614720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527181544633902,
      "loss": 1.3414,
      "theoretical_loss": 3.4386859874337317,
      "tokens_seen": 2147680256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035269809428284854,
      "loss": 1.3447,
      "theoretical_loss": 3.4386774321766604,
      "tokens_seen": 2147745792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526780341023069,
      "loss": 1.2385,
      "theoretical_loss": 3.4386688772537317,
      "tokens_seen": 2147811328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526579739217653,
      "loss": 1.3352,
      "theoretical_loss": 3.438660322664922,
      "tokens_seen": 2147876864
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526379137412237,
      "loss": 1.3521,
      "theoretical_loss": 3.4386517684102085,
      "tokens_seen": 2147942400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035261785356068204,
      "loss": 1.3489,
      "theoretical_loss": 3.4386432144895673,
      "tokens_seen": 2148007936
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 782164,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3615994453430176,
      "objective/train/theoretical_loss": 3.4386346609029763,
      "objective/train/tokens_used": 518532576,
      "theoretical_loss": 3.4386346609029763,
      "tokens_seen": 2148073472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525977933801404,
      "loss": 1.3905,
      "theoretical_loss": 3.4386346609029763,
      "tokens_seen": 2148073472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035257773319959877,
      "loss": 1.3062,
      "theoretical_loss": 3.4386261076504105,
      "tokens_seen": 2148139008
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525576730190572,
      "loss": 1.3979,
      "theoretical_loss": 3.4386175547318487,
      "tokens_seen": 2148204544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035253761283851555,
      "loss": 1.345,
      "theoretical_loss": 3.438609002147266,
      "tokens_seen": 2148270080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525175526579739,
      "loss": 1.2995,
      "theoretical_loss": 3.43860044989664,
      "tokens_seen": 2148335616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035249749247743227,
      "loss": 1.3768,
      "theoretical_loss": 3.438591897979947,
      "tokens_seen": 2148401152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524774322968907,
      "loss": 1.3092,
      "theoretical_loss": 3.4385833463971647,
      "tokens_seen": 2148466688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035245737211634905,
      "loss": 1.2958,
      "theoretical_loss": 3.4385747951482686,
      "tokens_seen": 2148532224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524373119358074,
      "loss": 1.3365,
      "theoretical_loss": 3.438566244233236,
      "tokens_seen": 2148597760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524172517552658,
      "loss": 1.373,
      "theoretical_loss": 3.4385576936520446,
      "tokens_seen": 2148663296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035239719157472414,
      "loss": 1.2744,
      "theoretical_loss": 3.4385491434046696,
      "tokens_seen": 2148728832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035237713139418255,
      "loss": 1.2524,
      "theoretical_loss": 3.438540593491089,
      "tokens_seen": 2148794368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523570712136409,
      "loss": 1.3847,
      "theoretical_loss": 3.4385320439112794,
      "tokens_seen": 2148859904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523370110330993,
      "loss": 1.294,
      "theoretical_loss": 3.438523494665217,
      "tokens_seen": 2148925440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035231695085255764,
      "loss": 1.2763,
      "theoretical_loss": 3.438514945752879,
      "tokens_seen": 2148990976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035229689067201606,
      "loss": 1.3273,
      "theoretical_loss": 3.4385063971742422,
      "tokens_seen": 2149056512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522768304914744,
      "loss": 1.3004,
      "theoretical_loss": 3.4384978489292832,
      "tokens_seen": 2149122048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522567703109328,
      "loss": 1.3441,
      "theoretical_loss": 3.438489301017979,
      "tokens_seen": 2149187584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035223671013039114,
      "loss": 1.3159,
      "theoretical_loss": 3.4384807534403063,
      "tokens_seen": 2149253120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035221664994984956,
      "loss": 1.447,
      "theoretical_loss": 3.4384722061962423,
      "tokens_seen": 2149318656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521965897693079,
      "loss": 1.1938,
      "theoretical_loss": 3.438463659285763,
      "tokens_seen": 2149384192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521765295887663,
      "loss": 1.3454,
      "theoretical_loss": 3.4384551127088456,
      "tokens_seen": 2149449728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035215646940822465,
      "loss": 1.322,
      "theoretical_loss": 3.4384465664654673,
      "tokens_seen": 2149515264
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352136409227683,
      "loss": 1.3285,
      "theoretical_loss": 3.4384380205556044,
      "tokens_seen": 2149580800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521163490471414,
      "loss": 1.3322,
      "theoretical_loss": 3.4384294749792343,
      "tokens_seen": 2149646336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035209628886659984,
      "loss": 1.2728,
      "theoretical_loss": 3.438420929736333,
      "tokens_seen": 2149711872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003520762286860582,
      "loss": 1.296,
      "theoretical_loss": 3.4384123848268784,
      "tokens_seen": 2149777408
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035205616850551657,
      "loss": 1.346,
      "theoretical_loss": 3.438403840250846,
      "tokens_seen": 2149842944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352036108324975,
      "loss": 1.3266,
      "theoretical_loss": 3.4383952960082134,
      "tokens_seen": 2149908480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035201604814443334,
      "loss": 1.4807,
      "theoretical_loss": 3.4383867520989577,
      "tokens_seen": 2149974016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003519959879638917,
      "loss": 1.2718,
      "theoretical_loss": 3.438378208523055,
      "tokens_seen": 2150039552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035197592778335007,
      "loss": 1.3319,
      "theoretical_loss": 3.4383696652804825,
      "tokens_seen": 2150105088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035195586760280843,
      "loss": 1.2501,
      "theoretical_loss": 3.438361122371217,
      "tokens_seen": 2150170624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035193580742226685,
      "loss": 1.361,
      "theoretical_loss": 3.438352579795236,
      "tokens_seen": 2150236160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003519157472417252,
      "loss": 1.2137,
      "theoretical_loss": 3.4383440375525147,
      "tokens_seen": 2150301696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035189568706118357,
      "loss": 1.2792,
      "theoretical_loss": 3.438335495643032,
      "tokens_seen": 2150367232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035187562688064193,
      "loss": 1.3469,
      "theoretical_loss": 3.4383269540667625,
      "tokens_seen": 2150432768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035185556670010035,
      "loss": 1.3378,
      "theoretical_loss": 3.438318412823685,
      "tokens_seen": 2150498304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518355065195587,
      "loss": 1.394,
      "theoretical_loss": 3.4383098719137757,
      "tokens_seen": 2150563840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518154463390171,
      "loss": 1.3268,
      "theoretical_loss": 3.438301331337011,
      "tokens_seen": 2150629376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035179538615847544,
      "loss": 1.3358,
      "theoretical_loss": 3.4382927910933683,
      "tokens_seen": 2150694912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517753259779338,
      "loss": 1.3563,
      "theoretical_loss": 3.438284251182824,
      "tokens_seen": 2150760448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517552657973922,
      "loss": 1.3519,
      "theoretical_loss": 3.438275711605356,
      "tokens_seen": 2150825984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517352056168506,
      "loss": 1.3287,
      "theoretical_loss": 3.4382671723609395,
      "tokens_seen": 2150891520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035171514543630894,
      "loss": 1.4074,
      "theoretical_loss": 3.4382586334495526,
      "tokens_seen": 2150957056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516950852557673,
      "loss": 1.4037,
      "theoretical_loss": 3.4382500948711723,
      "tokens_seen": 2151022592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516750250752257,
      "loss": 1.2955,
      "theoretical_loss": 3.4382415566257745,
      "tokens_seen": 2151088128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516549648946841,
      "loss": 1.4355,
      "theoretical_loss": 3.438233018713336,
      "tokens_seen": 2151153664
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035163490471414244,
      "loss": 1.45,
      "theoretical_loss": 3.438224481133835,
      "tokens_seen": 2151219200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516148445336008,
      "loss": 1.2596,
      "theoretical_loss": 3.438215943887247,
      "tokens_seen": 2151284736
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 783310,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.378399133682251,
      "objective/train/theoretical_loss": 3.43820740697355,
      "objective/train/tokens_used": 521809376,
      "theoretical_loss": 3.43820740697355,
      "tokens_seen": 2151350272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035159478435305917,
      "loss": 1.3267,
      "theoretical_loss": 3.43820740697355,
      "tokens_seen": 2151350272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515747241725176,
      "loss": 1.2686,
      "theoretical_loss": 3.4381988703927204,
      "tokens_seen": 2151415808
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035155466399197595,
      "loss": 1.3023,
      "theoretical_loss": 3.4381903341447346,
      "tokens_seen": 2151481344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515346038114343,
      "loss": 1.2003,
      "theoretical_loss": 3.4381817982295697,
      "tokens_seen": 2151546880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035151454363089267,
      "loss": 1.301,
      "theoretical_loss": 3.4381732626472035,
      "tokens_seen": 2151612416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514944834503511,
      "loss": 1.3024,
      "theoretical_loss": 3.438164727397612,
      "tokens_seen": 2151677952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035147442326980945,
      "loss": 1.3875,
      "theoretical_loss": 3.438156192480772,
      "tokens_seen": 2151743488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514543630892678,
      "loss": 1.4739,
      "theoretical_loss": 3.438147657896661,
      "tokens_seen": 2151809024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514343029087262,
      "loss": 1.3442,
      "theoretical_loss": 3.4381391236452554,
      "tokens_seen": 2151874560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035141424272818454,
      "loss": 1.2211,
      "theoretical_loss": 3.438130589726532,
      "tokens_seen": 2151940096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035139418254764295,
      "loss": 1.3373,
      "theoretical_loss": 3.4381220561404686,
      "tokens_seen": 2152005632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513741223671013,
      "loss": 1.3237,
      "theoretical_loss": 3.4381135228870408,
      "tokens_seen": 2152071168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513540621865597,
      "loss": 1.4257,
      "theoretical_loss": 3.438104989966227,
      "tokens_seen": 2152136704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035133400200601804,
      "loss": 1.3333,
      "theoretical_loss": 3.4380964573780024,
      "tokens_seen": 2152202240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035131394182547646,
      "loss": 1.2998,
      "theoretical_loss": 3.4380879251223453,
      "tokens_seen": 2152267776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512938816449348,
      "loss": 1.2912,
      "theoretical_loss": 3.438079393199232,
      "tokens_seen": 2152333312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512738214643932,
      "loss": 1.2987,
      "theoretical_loss": 3.4380708616086393,
      "tokens_seen": 2152398848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035125376128385154,
      "loss": 1.2368,
      "theoretical_loss": 3.4380623303505446,
      "tokens_seen": 2152464384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035123370110330996,
      "loss": 1.3106,
      "theoretical_loss": 3.4380537994249245,
      "tokens_seen": 2152529920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512136409227683,
      "loss": 1.3073,
      "theoretical_loss": 3.4380452688317558,
      "tokens_seen": 2152595456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511935807422267,
      "loss": 1.3273,
      "theoretical_loss": 3.438036738571016,
      "tokens_seen": 2152660992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035117352056168505,
      "loss": 1.3947,
      "theoretical_loss": 3.4380282086426814,
      "tokens_seen": 2152726528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511534603811434,
      "loss": 1.273,
      "theoretical_loss": 3.4380196790467292,
      "tokens_seen": 2152792064
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511334002006018,
      "loss": 1.2551,
      "theoretical_loss": 3.4380111497831365,
      "tokens_seen": 2152857600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511133400200602,
      "loss": 1.4666,
      "theoretical_loss": 3.4380026208518792,
      "tokens_seen": 2152923136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035109327983951855,
      "loss": 1.2774,
      "theoretical_loss": 3.4379940922529357,
      "tokens_seen": 2152988672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003510732196589769,
      "loss": 1.3835,
      "theoretical_loss": 3.4379855639862824,
      "tokens_seen": 2153054208
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035105315947843533,
      "loss": 1.2825,
      "theoretical_loss": 3.4379770360518958,
      "tokens_seen": 2153119744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003510330992978937,
      "loss": 1.2703,
      "theoretical_loss": 3.437968508449753,
      "tokens_seen": 2153185280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035101303911735205,
      "loss": 1.321,
      "theoretical_loss": 3.437959981179832,
      "tokens_seen": 2153250816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509929789368104,
      "loss": 1.3321,
      "theoretical_loss": 3.437951454242108,
      "tokens_seen": 2153316352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509729187562688,
      "loss": 1.3462,
      "theoretical_loss": 3.437942927636559,
      "tokens_seen": 2153381888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509528585757272,
      "loss": 1.3044,
      "theoretical_loss": 3.4379344013631616,
      "tokens_seen": 2153447424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035093279839518556,
      "loss": 1.267,
      "theoretical_loss": 3.4379258754218935,
      "tokens_seen": 2153512960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509127382146439,
      "loss": 1.3815,
      "theoretical_loss": 3.4379173498127304,
      "tokens_seen": 2153578496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508926780341023,
      "loss": 1.338,
      "theoretical_loss": 3.4379088245356506,
      "tokens_seen": 2153644032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508726178535607,
      "loss": 1.2662,
      "theoretical_loss": 3.43790029959063,
      "tokens_seen": 2153709568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035085255767301906,
      "loss": 1.3286,
      "theoretical_loss": 3.437891774977646,
      "tokens_seen": 2153775104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508324974924774,
      "loss": 1.3535,
      "theoretical_loss": 3.437883250696675,
      "tokens_seen": 2153840640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508124373119358,
      "loss": 1.3665,
      "theoretical_loss": 3.4378747267476957,
      "tokens_seen": 2153906176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035079237713139415,
      "loss": 1.3857,
      "theoretical_loss": 3.4378662031306826,
      "tokens_seen": 2153971712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035077231695085256,
      "loss": 1.2865,
      "theoretical_loss": 3.4378576798456146,
      "tokens_seen": 2154037248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507522567703109,
      "loss": 1.3488,
      "theoretical_loss": 3.437849156892468,
      "tokens_seen": 2154102784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507321965897693,
      "loss": 1.3197,
      "theoretical_loss": 3.4378406342712196,
      "tokens_seen": 2154168320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035071213640922765,
      "loss": 1.2979,
      "theoretical_loss": 3.4378321119818462,
      "tokens_seen": 2154233856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035069207622868607,
      "loss": 1.3715,
      "theoretical_loss": 3.4378235900243257,
      "tokens_seen": 2154299392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035067201604814443,
      "loss": 1.3941,
      "theoretical_loss": 3.4378150683986344,
      "tokens_seen": 2154364928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506519558676028,
      "loss": 1.2239,
      "theoretical_loss": 3.437806547104749,
      "tokens_seen": 2154430464
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035063189568706115,
      "loss": 1.4198,
      "theoretical_loss": 3.4377980261426475,
      "tokens_seen": 2154496000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506118355065195,
      "loss": 1.2291,
      "theoretical_loss": 3.437789505512306,
      "tokens_seen": 2154561536
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 784476,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3676598072052002,
      "objective/train/theoretical_loss": 3.437780985213702,
      "objective/train/tokens_used": 525086176,
      "theoretical_loss": 3.437780985213702,
      "tokens_seen": 2154627072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000350591775325978,
      "loss": 1.3837,
      "theoretical_loss": 3.437780985213702,
      "tokens_seen": 2154627072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035057171514543635,
      "loss": 1.2263,
      "theoretical_loss": 3.4377724652468116,
      "tokens_seen": 2154692608
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505516549648947,
      "loss": 1.3726,
      "theoretical_loss": 3.4377639456116134,
      "tokens_seen": 2154758144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035053159478435307,
      "loss": 1.3508,
      "theoretical_loss": 3.4377554263080827,
      "tokens_seen": 2154823680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505115346038115,
      "loss": 1.3855,
      "theoretical_loss": 3.4377469073361975,
      "tokens_seen": 2154889216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035049147442326985,
      "loss": 1.3991,
      "theoretical_loss": 3.4377383886959345,
      "tokens_seen": 2154954752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504714142427282,
      "loss": 1.3087,
      "theoretical_loss": 3.4377298703872707,
      "tokens_seen": 2155020288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504513540621866,
      "loss": 1.2403,
      "theoretical_loss": 3.4377213524101835,
      "tokens_seen": 2155085824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035043129388164494,
      "loss": 1.2824,
      "theoretical_loss": 3.4377128347646497,
      "tokens_seen": 2155151360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035041123370110335,
      "loss": 1.2955,
      "theoretical_loss": 3.4377043174506454,
      "tokens_seen": 2155216896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503911735205617,
      "loss": 1.3288,
      "theoretical_loss": 3.4376958004681493,
      "tokens_seen": 2155282432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503711133400201,
      "loss": 1.3505,
      "theoretical_loss": 3.4376872838171373,
      "tokens_seen": 2155347968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035035105315947844,
      "loss": 1.2374,
      "theoretical_loss": 3.4376787674975864,
      "tokens_seen": 2155413504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035033099297893686,
      "loss": 1.4408,
      "theoretical_loss": 3.437670251509474,
      "tokens_seen": 2155479040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503109327983952,
      "loss": 1.3667,
      "theoretical_loss": 3.4376617358527772,
      "tokens_seen": 2155544576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502908726178536,
      "loss": 1.3007,
      "theoretical_loss": 3.4376532205274724,
      "tokens_seen": 2155610112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035027081243731194,
      "loss": 1.3497,
      "theoretical_loss": 3.437644705533537,
      "tokens_seen": 2155675648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035025075225677036,
      "loss": 1.2769,
      "theoretical_loss": 3.4376361908709487,
      "tokens_seen": 2155741184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502306920762287,
      "loss": 1.3185,
      "theoretical_loss": 3.437627676539684,
      "tokens_seen": 2155806720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502106318956871,
      "loss": 1.3205,
      "theoretical_loss": 3.437619162539719,
      "tokens_seen": 2155872256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035019057171514545,
      "loss": 1.3376,
      "theoretical_loss": 3.4376106488710327,
      "tokens_seen": 2155937792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501705115346038,
      "loss": 1.3628,
      "theoretical_loss": 3.4376021355336004,
      "tokens_seen": 2156003328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501504513540622,
      "loss": 1.3153,
      "theoretical_loss": 3.4375936225273995,
      "tokens_seen": 2156068864
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501303911735206,
      "loss": 1.2844,
      "theoretical_loss": 3.4375851098524075,
      "tokens_seen": 2156134400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035011033099297895,
      "loss": 1.2758,
      "theoretical_loss": 3.4375765975086017,
      "tokens_seen": 2156199936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003500902708124373,
      "loss": 1.2258,
      "theoretical_loss": 3.4375680854959585,
      "tokens_seen": 2156265472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035007021063189573,
      "loss": 1.2413,
      "theoretical_loss": 3.437559573814455,
      "tokens_seen": 2156331008
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003500501504513541,
      "loss": 1.2511,
      "theoretical_loss": 3.4375510624640686,
      "tokens_seen": 2156396544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035003009027081245,
      "loss": 1.3138,
      "theoretical_loss": 3.437542551444776,
      "tokens_seen": 2156462080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003500100300902708,
      "loss": 1.3907,
      "theoretical_loss": 3.437534040756555,
      "tokens_seen": 2156527616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499899699097292,
      "loss": 1.272,
      "theoretical_loss": 3.4375255303993812,
      "tokens_seen": 2156593152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499699097291876,
      "loss": 1.2678,
      "theoretical_loss": 3.4375170203732335,
      "tokens_seen": 2156658688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034994984954864596,
      "loss": 1.3548,
      "theoretical_loss": 3.4375085106780876,
      "tokens_seen": 2156724224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499297893681043,
      "loss": 1.4314,
      "theoretical_loss": 3.437500001313921,
      "tokens_seen": 2156789760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499097291875627,
      "loss": 1.2993,
      "theoretical_loss": 3.437491492280711,
      "tokens_seen": 2156855296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498896690070211,
      "loss": 1.2253,
      "theoretical_loss": 3.437482983578434,
      "tokens_seen": 2156920832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034986960882647946,
      "loss": 1.2981,
      "theoretical_loss": 3.437474475207068,
      "tokens_seen": 2156986368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498495486459378,
      "loss": 1.337,
      "theoretical_loss": 3.437465967166589,
      "tokens_seen": 2157051904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498294884653962,
      "loss": 1.2337,
      "theoretical_loss": 3.4374574594569753,
      "tokens_seen": 2157117440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034980942828485455,
      "loss": 1.3061,
      "theoretical_loss": 3.437448952078203,
      "tokens_seen": 2157182976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034978936810431296,
      "loss": 1.4652,
      "theoretical_loss": 3.4374404450302496,
      "tokens_seen": 2157248512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497693079237713,
      "loss": 1.2987,
      "theoretical_loss": 3.4374319383130922,
      "tokens_seen": 2157314048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497492477432297,
      "loss": 1.4275,
      "theoretical_loss": 3.4374234319267076,
      "tokens_seen": 2157379584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034972918756268805,
      "loss": 1.4213,
      "theoretical_loss": 3.4374149258710736,
      "tokens_seen": 2157445120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034970912738214647,
      "loss": 1.3919,
      "theoretical_loss": 3.437406420146167,
      "tokens_seen": 2157510656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034968906720160483,
      "loss": 1.3731,
      "theoretical_loss": 3.4373979147519638,
      "tokens_seen": 2157576192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496690070210632,
      "loss": 1.226,
      "theoretical_loss": 3.4373894096884423,
      "tokens_seen": 2157641728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034964894684052155,
      "loss": 1.3293,
      "theoretical_loss": 3.4373809049555795,
      "tokens_seen": 2157707264
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496288866599799,
      "loss": 1.3446,
      "theoretical_loss": 3.437372400553352,
      "tokens_seen": 2157772800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034960882647943833,
      "loss": 1.445,
      "theoretical_loss": 3.4373638964817377,
      "tokens_seen": 2157838336
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 785665,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1829450130462646,
      "objective/train/theoretical_loss": 3.437355392740713,
      "objective/train/tokens_used": 528362976,
      "theoretical_loss": 3.437355392740713,
      "tokens_seen": 2157903872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495887662988967,
      "loss": 1.2758,
      "theoretical_loss": 3.437355392740713,
      "tokens_seen": 2157903872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034956870611835506,
      "loss": 1.3502,
      "theoretical_loss": 3.437346889330255,
      "tokens_seen": 2157969408
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495486459378134,
      "loss": 1.2818,
      "theoretical_loss": 3.4373383862503415,
      "tokens_seen": 2158034944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034952858575727183,
      "loss": 1.269,
      "theoretical_loss": 3.4373298835009485,
      "tokens_seen": 2158100480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495085255767302,
      "loss": 1.3553,
      "theoretical_loss": 3.437321381082054,
      "tokens_seen": 2158166016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034948846539618856,
      "loss": 1.2537,
      "theoretical_loss": 3.437312878993635,
      "tokens_seen": 2158231552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494684052156469,
      "loss": 1.3779,
      "theoretical_loss": 3.4373043772356686,
      "tokens_seen": 2158297088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494483450351053,
      "loss": 1.4004,
      "theoretical_loss": 3.4372958758081316,
      "tokens_seen": 2158362624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494282848545637,
      "loss": 1.3365,
      "theoretical_loss": 3.4372873747110013,
      "tokens_seen": 2158428160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034940822467402206,
      "loss": 1.289,
      "theoretical_loss": 3.4372788739442552,
      "tokens_seen": 2158493696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493881644934804,
      "loss": 1.4093,
      "theoretical_loss": 3.43727037350787,
      "tokens_seen": 2158559232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493681043129388,
      "loss": 1.3102,
      "theoretical_loss": 3.4372618734018228,
      "tokens_seen": 2158624768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493480441323972,
      "loss": 1.3522,
      "theoretical_loss": 3.437253373626091,
      "tokens_seen": 2158690304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034932798395185557,
      "loss": 1.2482,
      "theoretical_loss": 3.437244874180651,
      "tokens_seen": 2158755840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034930792377131393,
      "loss": 1.2846,
      "theoretical_loss": 3.4372363750654813,
      "tokens_seen": 2158821376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492878635907723,
      "loss": 1.2888,
      "theoretical_loss": 3.4372278762805584,
      "tokens_seen": 2158886912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034926780341023065,
      "loss": 1.3336,
      "theoretical_loss": 3.437219377825859,
      "tokens_seen": 2158952448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034924774322968907,
      "loss": 1.3491,
      "theoretical_loss": 3.43721087970136,
      "tokens_seen": 2159017984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034922768304914743,
      "loss": 1.3504,
      "theoretical_loss": 3.43720238190704,
      "tokens_seen": 2159083520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492076228686058,
      "loss": 1.2928,
      "theoretical_loss": 3.437193884442875,
      "tokens_seen": 2159149056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034918756268806416,
      "loss": 1.336,
      "theoretical_loss": 3.4371853873088423,
      "tokens_seen": 2159214592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034916750250752257,
      "loss": 1.2966,
      "theoretical_loss": 3.4371768905049196,
      "tokens_seen": 2159280128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034914744232698093,
      "loss": 1.3895,
      "theoretical_loss": 3.437168394031083,
      "tokens_seen": 2159345664
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003491273821464393,
      "loss": 1.4201,
      "theoretical_loss": 3.4371598978873106,
      "tokens_seen": 2159411200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034910732196589766,
      "loss": 1.3285,
      "theoretical_loss": 3.4371514020735794,
      "tokens_seen": 2159476736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034908726178535613,
      "loss": 1.3024,
      "theoretical_loss": 3.437142906589866,
      "tokens_seen": 2159542272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490672016048145,
      "loss": 1.3293,
      "theoretical_loss": 3.437134411436148,
      "tokens_seen": 2159607808
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034904714142427285,
      "loss": 1.3073,
      "theoretical_loss": 3.437125916612403,
      "tokens_seen": 2159673344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490270812437312,
      "loss": 1.3164,
      "theoretical_loss": 3.4371174221186074,
      "tokens_seen": 2159738880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490070210631896,
      "loss": 1.3424,
      "theoretical_loss": 3.437108927954739,
      "tokens_seen": 2159804416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348986960882648,
      "loss": 1.371,
      "theoretical_loss": 3.4371004341207745,
      "tokens_seen": 2159869952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034896690070210636,
      "loss": 1.3036,
      "theoretical_loss": 3.4370919406166913,
      "tokens_seen": 2159935488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489468405215647,
      "loss": 1.3252,
      "theoretical_loss": 3.4370834474424665,
      "tokens_seen": 2160001024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489267803410231,
      "loss": 1.3924,
      "theoretical_loss": 3.437074954598077,
      "tokens_seen": 2160066560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489067201604815,
      "loss": 1.3545,
      "theoretical_loss": 3.437066462083501,
      "tokens_seen": 2160132096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034888665997993986,
      "loss": 1.4518,
      "theoretical_loss": 3.4370579698987145,
      "tokens_seen": 2160197632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488665997993982,
      "loss": 1.3248,
      "theoretical_loss": 3.4370494780436953,
      "tokens_seen": 2160263168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488465396188566,
      "loss": 1.3962,
      "theoretical_loss": 3.4370409865184204,
      "tokens_seen": 2160328704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034882647943831495,
      "loss": 1.2966,
      "theoretical_loss": 3.437032495322867,
      "tokens_seen": 2160394240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034880641925777336,
      "loss": 1.3209,
      "theoretical_loss": 3.437024004457012,
      "tokens_seen": 2160459776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487863590772317,
      "loss": 1.3036,
      "theoretical_loss": 3.4370155139208336,
      "tokens_seen": 2160525312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487662988966901,
      "loss": 1.4217,
      "theoretical_loss": 3.437007023714308,
      "tokens_seen": 2160590848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034874623871614845,
      "loss": 1.4746,
      "theoretical_loss": 3.436998533837413,
      "tokens_seen": 2160656384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034872617853560687,
      "loss": 1.2652,
      "theoretical_loss": 3.436990044290125,
      "tokens_seen": 2160721920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034870611835506523,
      "loss": 1.4176,
      "theoretical_loss": 3.4369815550724225,
      "tokens_seen": 2160787456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486860581745236,
      "loss": 1.2972,
      "theoretical_loss": 3.4369730661842817,
      "tokens_seen": 2160852992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034866599799398195,
      "loss": 1.4011,
      "theoretical_loss": 3.43696457762568,
      "tokens_seen": 2160918528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486459378134403,
      "loss": 1.2829,
      "theoretical_loss": 3.4369560893965945,
      "tokens_seen": 2160984064
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034862587763289873,
      "loss": 1.2168,
      "theoretical_loss": 3.436947601497003,
      "tokens_seen": 2161049600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486058174523571,
      "loss": 1.3849,
      "theoretical_loss": 3.436939113926882,
      "tokens_seen": 2161115136
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 786923,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1785483360290527,
      "objective/train/theoretical_loss": 3.436930626686209,
      "objective/train/tokens_used": 531639776,
      "theoretical_loss": 3.436930626686209,
      "tokens_seen": 2161180672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034858575727181546,
      "loss": 1.1877,
      "theoretical_loss": 3.436930626686209,
      "tokens_seen": 2161180672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003485656970912738,
      "loss": 1.4342,
      "theoretical_loss": 3.436922139774962,
      "tokens_seen": 2161246208
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034854563691073223,
      "loss": 1.2839,
      "theoretical_loss": 3.436913653193117,
      "tokens_seen": 2161311744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003485255767301906,
      "loss": 1.229,
      "theoretical_loss": 3.4369051669406514,
      "tokens_seen": 2161377280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034850551654964896,
      "loss": 1.2479,
      "theoretical_loss": 3.436896681017543,
      "tokens_seen": 2161442816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484854563691073,
      "loss": 1.2655,
      "theoretical_loss": 3.436888195423769,
      "tokens_seen": 2161508352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484653961885657,
      "loss": 1.2562,
      "theoretical_loss": 3.4368797101593063,
      "tokens_seen": 2161573888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484453360080241,
      "loss": 1.4088,
      "theoretical_loss": 3.436871225224132,
      "tokens_seen": 2161639424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034842527582748246,
      "loss": 1.2679,
      "theoretical_loss": 3.4368627406182237,
      "tokens_seen": 2161704960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484052156469408,
      "loss": 1.2422,
      "theoretical_loss": 3.436854256341559,
      "tokens_seen": 2161770496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483851554663992,
      "loss": 1.2552,
      "theoretical_loss": 3.436845772394114,
      "tokens_seen": 2161836032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483650952858576,
      "loss": 1.3491,
      "theoretical_loss": 3.4368372887758674,
      "tokens_seen": 2161901568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034834503510531597,
      "loss": 1.4595,
      "theoretical_loss": 3.436828805486795,
      "tokens_seen": 2161967104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034832497492477433,
      "loss": 1.4305,
      "theoretical_loss": 3.436820322526875,
      "tokens_seen": 2162032640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483049147442327,
      "loss": 1.2658,
      "theoretical_loss": 3.4368118398960843,
      "tokens_seen": 2162098176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034828485456369105,
      "loss": 1.2528,
      "theoretical_loss": 3.4368033575944,
      "tokens_seen": 2162163712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034826479438314947,
      "loss": 1.4206,
      "theoretical_loss": 3.4367948756218,
      "tokens_seen": 2162229248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034824473420260783,
      "loss": 1.3028,
      "theoretical_loss": 3.4367863939782612,
      "tokens_seen": 2162294784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003482246740220662,
      "loss": 1.2927,
      "theoretical_loss": 3.4367779126637603,
      "tokens_seen": 2162360320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034820461384152456,
      "loss": 1.3685,
      "theoretical_loss": 3.4367694316782753,
      "tokens_seen": 2162425856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034818455366098297,
      "loss": 1.3137,
      "theoretical_loss": 3.4367609510217836,
      "tokens_seen": 2162491392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034816449348044133,
      "loss": 1.2694,
      "theoretical_loss": 3.4367524706942616,
      "tokens_seen": 2162556928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481444332998997,
      "loss": 1.3111,
      "theoretical_loss": 3.436743990695687,
      "tokens_seen": 2162622464
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034812437311935806,
      "loss": 1.2871,
      "theoretical_loss": 3.4367355110260376,
      "tokens_seen": 2162688000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481043129388164,
      "loss": 1.3616,
      "theoretical_loss": 3.4367270316852903,
      "tokens_seen": 2162753536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034808425275827484,
      "loss": 1.3638,
      "theoretical_loss": 3.4367185526734216,
      "tokens_seen": 2162819072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003480641925777332,
      "loss": 1.4093,
      "theoretical_loss": 3.43671007399041,
      "tokens_seen": 2162884608
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034804413239719156,
      "loss": 1.3259,
      "theoretical_loss": 3.436701595636232,
      "tokens_seen": 2162950144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003480240722166499,
      "loss": 1.3571,
      "theoretical_loss": 3.4366931176108655,
      "tokens_seen": 2163015680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034800401203610834,
      "loss": 1.2948,
      "theoretical_loss": 3.4366846399142874,
      "tokens_seen": 2163081216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003479839518555667,
      "loss": 1.2727,
      "theoretical_loss": 3.436676162546475,
      "tokens_seen": 2163146752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034796389167502506,
      "loss": 1.3393,
      "theoretical_loss": 3.4366676855074054,
      "tokens_seen": 2163212288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034794383149448343,
      "loss": 1.3181,
      "theoretical_loss": 3.4366592087970558,
      "tokens_seen": 2163277824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034792377131394184,
      "loss": 1.3202,
      "theoretical_loss": 3.436650732415404,
      "tokens_seen": 2163343360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003479037111334002,
      "loss": 1.328,
      "theoretical_loss": 3.4366422563624277,
      "tokens_seen": 2163408896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034788365095285857,
      "loss": 1.3365,
      "theoretical_loss": 3.436633780638103,
      "tokens_seen": 2163474432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034786359077231693,
      "loss": 1.2109,
      "theoretical_loss": 3.436625305242408,
      "tokens_seen": 2163539968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478435305917753,
      "loss": 1.3071,
      "theoretical_loss": 3.4366168301753195,
      "tokens_seen": 2163605504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478234704112337,
      "loss": 1.3247,
      "theoretical_loss": 3.4366083554368156,
      "tokens_seen": 2163671040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034780341023069207,
      "loss": 1.2939,
      "theoretical_loss": 3.4365998810268725,
      "tokens_seen": 2163736576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034778335005015043,
      "loss": 1.337,
      "theoretical_loss": 3.4365914069454684,
      "tokens_seen": 2163802112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477632898696088,
      "loss": 1.3165,
      "theoretical_loss": 3.4365829331925806,
      "tokens_seen": 2163867648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477432296890672,
      "loss": 1.387,
      "theoretical_loss": 3.436574459768186,
      "tokens_seen": 2163933184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477231695085256,
      "loss": 1.4402,
      "theoretical_loss": 3.436565986672262,
      "tokens_seen": 2163998720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034770310932798394,
      "loss": 1.4596,
      "theoretical_loss": 3.436557513904786,
      "tokens_seen": 2164064256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476830491474423,
      "loss": 1.336,
      "theoretical_loss": 3.436549041465735,
      "tokens_seen": 2164129792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034766298896690066,
      "loss": 1.336,
      "theoretical_loss": 3.436540569355087,
      "tokens_seen": 2164195328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476429287863591,
      "loss": 1.2667,
      "theoretical_loss": 3.436532097572819,
      "tokens_seen": 2164260864
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034762286860581744,
      "loss": 1.4381,
      "theoretical_loss": 3.436523626118908,
      "tokens_seen": 2164326400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476028084252758,
      "loss": 1.3469,
      "theoretical_loss": 3.436515154993332,
      "tokens_seen": 2164391936
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 788153,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5332597494125366,
      "objective/train/theoretical_loss": 3.436506684196068,
      "objective/train/tokens_used": 534916576,
      "theoretical_loss": 3.436506684196068,
      "tokens_seen": 2164457472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475827482447342,
      "loss": 1.4794,
      "theoretical_loss": 3.436506684196068,
      "tokens_seen": 2164457472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034756268806419263,
      "loss": 1.3252,
      "theoretical_loss": 3.4364982137270927,
      "tokens_seen": 2164523008
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347542627883651,
      "loss": 1.294,
      "theoretical_loss": 3.4364897435863844,
      "tokens_seen": 2164588544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034752256770310936,
      "loss": 1.2614,
      "theoretical_loss": 3.43648127377392,
      "tokens_seen": 2164654080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475025075225677,
      "loss": 1.3492,
      "theoretical_loss": 3.436472804289677,
      "tokens_seen": 2164719616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474824473420261,
      "loss": 1.275,
      "theoretical_loss": 3.436464335133633,
      "tokens_seen": 2164785152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474623871614845,
      "loss": 1.3473,
      "theoretical_loss": 3.4364558663057645,
      "tokens_seen": 2164850688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034744232698094286,
      "loss": 1.3333,
      "theoretical_loss": 3.4364473978060492,
      "tokens_seen": 2164916224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474222668004012,
      "loss": 1.235,
      "theoretical_loss": 3.4364389296344653,
      "tokens_seen": 2164981760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474022066198596,
      "loss": 1.3302,
      "theoretical_loss": 3.4364304617909887,
      "tokens_seen": 2165047296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347382146439318,
      "loss": 1.3896,
      "theoretical_loss": 3.436421994275598,
      "tokens_seen": 2165112832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034736208625877637,
      "loss": 1.2921,
      "theoretical_loss": 3.43641352708827,
      "tokens_seen": 2165178368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034734202607823473,
      "loss": 1.4465,
      "theoretical_loss": 3.436405060228982,
      "tokens_seen": 2165243904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473219658976931,
      "loss": 1.3738,
      "theoretical_loss": 3.4363965936977117,
      "tokens_seen": 2165309440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034730190571715145,
      "loss": 1.4406,
      "theoretical_loss": 3.436388127494436,
      "tokens_seen": 2165374976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034728184553660987,
      "loss": 1.3367,
      "theoretical_loss": 3.4363796616191324,
      "tokens_seen": 2165440512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034726178535606823,
      "loss": 1.3057,
      "theoretical_loss": 3.4363711960717787,
      "tokens_seen": 2165506048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472417251755266,
      "loss": 1.2725,
      "theoretical_loss": 3.436362730852352,
      "tokens_seen": 2165571584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034722166499498496,
      "loss": 1.2551,
      "theoretical_loss": 3.4363542659608295,
      "tokens_seen": 2165637120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034720160481444337,
      "loss": 1.3699,
      "theoretical_loss": 3.436345801397189,
      "tokens_seen": 2165702656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034718154463390173,
      "loss": 1.4416,
      "theoretical_loss": 3.4363373371614068,
      "tokens_seen": 2165768192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471614844533601,
      "loss": 1.3418,
      "theoretical_loss": 3.4363288732534616,
      "tokens_seen": 2165833728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034714142427281846,
      "loss": 1.2469,
      "theoretical_loss": 3.43632040967333,
      "tokens_seen": 2165899264
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471213640922768,
      "loss": 1.3655,
      "theoretical_loss": 3.4363119464209904,
      "tokens_seen": 2165964800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034710130391173524,
      "loss": 1.3427,
      "theoretical_loss": 3.4363034834964186,
      "tokens_seen": 2166030336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003470812437311936,
      "loss": 1.3184,
      "theoretical_loss": 3.4362950208995935,
      "tokens_seen": 2166095872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034706118355065196,
      "loss": 1.2642,
      "theoretical_loss": 3.436286558630491,
      "tokens_seen": 2166161408
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003470411233701103,
      "loss": 1.3457,
      "theoretical_loss": 3.4362780966890902,
      "tokens_seen": 2166226944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034702106318956874,
      "loss": 1.3559,
      "theoretical_loss": 3.436269635075367,
      "tokens_seen": 2166292480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003470010030090271,
      "loss": 1.3273,
      "theoretical_loss": 3.4362611737893,
      "tokens_seen": 2166358016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034698094282848547,
      "loss": 1.3022,
      "theoretical_loss": 3.436252712830865,
      "tokens_seen": 2166423552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034696088264794383,
      "loss": 1.2717,
      "theoretical_loss": 3.436244252200041,
      "tokens_seen": 2166489088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034694082246740224,
      "loss": 1.2509,
      "theoretical_loss": 3.436235791896805,
      "tokens_seen": 2166554624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003469207622868606,
      "loss": 1.2517,
      "theoretical_loss": 3.436227331921134,
      "tokens_seen": 2166620160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034690070210631897,
      "loss": 1.3783,
      "theoretical_loss": 3.4362188722730056,
      "tokens_seen": 2166685696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034688064192577733,
      "loss": 1.4356,
      "theoretical_loss": 3.4362104129523976,
      "tokens_seen": 2166751232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468605817452357,
      "loss": 1.3083,
      "theoretical_loss": 3.4362019539592863,
      "tokens_seen": 2166816768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468405215646941,
      "loss": 1.3557,
      "theoretical_loss": 3.436193495293651,
      "tokens_seen": 2166882304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034682046138415247,
      "loss": 1.3302,
      "theoretical_loss": 3.4361850369554667,
      "tokens_seen": 2166947840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034680040120361083,
      "loss": 1.3662,
      "theoretical_loss": 3.4361765789447127,
      "tokens_seen": 2167013376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467803410230692,
      "loss": 1.3596,
      "theoretical_loss": 3.4361681212613657,
      "tokens_seen": 2167078912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467602808425276,
      "loss": 1.3058,
      "theoretical_loss": 3.4361596639054035,
      "tokens_seen": 2167144448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346740220661986,
      "loss": 1.3088,
      "theoretical_loss": 3.4361512068768034,
      "tokens_seen": 2167209984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034672016048144434,
      "loss": 1.1991,
      "theoretical_loss": 3.4361427501755424,
      "tokens_seen": 2167275520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467001003009027,
      "loss": 1.3397,
      "theoretical_loss": 3.436134293801598,
      "tokens_seen": 2167341056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034668004012036106,
      "loss": 1.3777,
      "theoretical_loss": 3.4361258377549486,
      "tokens_seen": 2167406592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003466599799398195,
      "loss": 1.3044,
      "theoretical_loss": 3.4361173820355706,
      "tokens_seen": 2167472128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034663991975927784,
      "loss": 1.4952,
      "theoretical_loss": 3.4361089266434415,
      "tokens_seen": 2167537664
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003466198595787362,
      "loss": 1.2997,
      "theoretical_loss": 3.4361004715785395,
      "tokens_seen": 2167603200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034659979939819456,
      "loss": 1.33,
      "theoretical_loss": 3.436092016840841,
      "tokens_seen": 2167668736
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 789244,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3871870040893555,
      "objective/train/theoretical_loss": 3.436083562430324,
      "objective/train/tokens_used": 538193376,
      "theoretical_loss": 3.436083562430324,
      "tokens_seen": 2167734272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346579739217653,
      "loss": 1.3178,
      "theoretical_loss": 3.436083562430324,
      "tokens_seen": 2167734272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034655967903711134,
      "loss": 1.2407,
      "theoretical_loss": 3.4360751083469663,
      "tokens_seen": 2167799808
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465396188565697,
      "loss": 1.287,
      "theoretical_loss": 3.4360666545907446,
      "tokens_seen": 2167865344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034651955867602807,
      "loss": 1.3855,
      "theoretical_loss": 3.436058201161637,
      "tokens_seen": 2167930880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034649949849548643,
      "loss": 1.435,
      "theoretical_loss": 3.4360497480596206,
      "tokens_seen": 2167996416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034647943831494485,
      "loss": 1.4191,
      "theoretical_loss": 3.4360412952846735,
      "tokens_seen": 2168061952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003464593781344032,
      "loss": 1.3908,
      "theoretical_loss": 3.436032842836772,
      "tokens_seen": 2168127488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034643931795386157,
      "loss": 1.3212,
      "theoretical_loss": 3.436024390715894,
      "tokens_seen": 2168193024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034641925777331993,
      "loss": 1.3645,
      "theoretical_loss": 3.4360159389220177,
      "tokens_seen": 2168258560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034639919759277835,
      "loss": 1.2613,
      "theoretical_loss": 3.4360074874551194,
      "tokens_seen": 2168324096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463791374122367,
      "loss": 1.3637,
      "theoretical_loss": 3.4359990363151773,
      "tokens_seen": 2168389632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463590772316951,
      "loss": 1.2942,
      "theoretical_loss": 3.4359905855021693,
      "tokens_seen": 2168455168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034633901705115344,
      "loss": 1.2805,
      "theoretical_loss": 3.4359821350160717,
      "tokens_seen": 2168520704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463189568706118,
      "loss": 1.3219,
      "theoretical_loss": 3.435973684856863,
      "tokens_seen": 2168586240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462988966900702,
      "loss": 1.4089,
      "theoretical_loss": 3.4359652350245202,
      "tokens_seen": 2168651776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462788365095286,
      "loss": 1.2908,
      "theoretical_loss": 3.4359567855190205,
      "tokens_seen": 2168717312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034625877632898694,
      "loss": 1.3103,
      "theoretical_loss": 3.435948336340342,
      "tokens_seen": 2168782848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462387161484453,
      "loss": 1.3133,
      "theoretical_loss": 3.4359398874884617,
      "tokens_seen": 2168848384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462186559679037,
      "loss": 1.3377,
      "theoretical_loss": 3.4359314389633573,
      "tokens_seen": 2168913920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461985957873621,
      "loss": 1.3764,
      "theoretical_loss": 3.4359229907650066,
      "tokens_seen": 2168979456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034617853560682044,
      "loss": 1.3872,
      "theoretical_loss": 3.4359145428933866,
      "tokens_seen": 2169044992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461584754262788,
      "loss": 1.2725,
      "theoretical_loss": 3.435906095348475,
      "tokens_seen": 2169110528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034613841524573717,
      "loss": 1.2686,
      "theoretical_loss": 3.435897648130249,
      "tokens_seen": 2169176064
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461183550651956,
      "loss": 1.414,
      "theoretical_loss": 3.4358892012386866,
      "tokens_seen": 2169241600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034609829488465395,
      "loss": 1.2807,
      "theoretical_loss": 3.435880754673765,
      "tokens_seen": 2169307136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034607823470411236,
      "loss": 1.2975,
      "theoretical_loss": 3.435872308435462,
      "tokens_seen": 2169372672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003460581745235707,
      "loss": 1.3427,
      "theoretical_loss": 3.435863862523754,
      "tokens_seen": 2169438208
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034603811434302914,
      "loss": 1.41,
      "theoretical_loss": 3.43585541693862,
      "tokens_seen": 2169503744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003460180541624875,
      "loss": 1.3423,
      "theoretical_loss": 3.435846971680037,
      "tokens_seen": 2169569280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034599799398194587,
      "loss": 1.2122,
      "theoretical_loss": 3.435838526747982,
      "tokens_seen": 2169634816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034597793380140423,
      "loss": 1.2868,
      "theoretical_loss": 3.4358300821424335,
      "tokens_seen": 2169700352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034595787362086264,
      "loss": 1.2497,
      "theoretical_loss": 3.4358216378633677,
      "tokens_seen": 2169765888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345937813440321,
      "loss": 1.3116,
      "theoretical_loss": 3.435813193910763,
      "tokens_seen": 2169831424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034591775325977937,
      "loss": 1.3055,
      "theoretical_loss": 3.435804750284597,
      "tokens_seen": 2169896960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034589769307923773,
      "loss": 1.2995,
      "theoretical_loss": 3.4357963069848467,
      "tokens_seen": 2169962496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458776328986961,
      "loss": 1.3358,
      "theoretical_loss": 3.4357878640114903,
      "tokens_seen": 2170028032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458575727181545,
      "loss": 1.2948,
      "theoretical_loss": 3.4357794213645043,
      "tokens_seen": 2170093568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034583751253761287,
      "loss": 1.3773,
      "theoretical_loss": 3.435770979043867,
      "tokens_seen": 2170159104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034581745235707123,
      "loss": 1.3403,
      "theoretical_loss": 3.4357625370495564,
      "tokens_seen": 2170224640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457973921765296,
      "loss": 1.3853,
      "theoretical_loss": 3.4357540953815486,
      "tokens_seen": 2170290176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345777331995988,
      "loss": 1.3735,
      "theoretical_loss": 3.4357456540398226,
      "tokens_seen": 2170355712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457572718154464,
      "loss": 1.3183,
      "theoretical_loss": 3.435737213024355,
      "tokens_seen": 2170421248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034573721163490474,
      "loss": 1.3896,
      "theoretical_loss": 3.4357287723351235,
      "tokens_seen": 2170486784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457171514543631,
      "loss": 1.3508,
      "theoretical_loss": 3.4357203319721057,
      "tokens_seen": 2170552320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034569709127382146,
      "loss": 1.3997,
      "theoretical_loss": 3.4357118919352794,
      "tokens_seen": 2170617856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456770310932799,
      "loss": 1.4799,
      "theoretical_loss": 3.435703452224622,
      "tokens_seen": 2170683392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034565697091273824,
      "loss": 1.3092,
      "theoretical_loss": 3.4356950128401107,
      "tokens_seen": 2170748928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456369107321966,
      "loss": 1.3922,
      "theoretical_loss": 3.4356865737817235,
      "tokens_seen": 2170814464
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034561685055165496,
      "loss": 1.3961,
      "theoretical_loss": 3.435678135049438,
      "tokens_seen": 2170880000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003455967903711134,
      "loss": 1.2927,
      "theoretical_loss": 3.4356696966432314,
      "tokens_seen": 2170945536
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 790517,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.431609869003296,
      "objective/train/theoretical_loss": 3.4356612585630812,
      "objective/train/tokens_used": 541470176,
      "theoretical_loss": 3.4356612585630812,
      "tokens_seen": 2171011072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034557673019057174,
      "loss": 1.4045,
      "theoretical_loss": 3.4356612585630812,
      "tokens_seen": 2171011072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003455566700100301,
      "loss": 1.3023,
      "theoretical_loss": 3.4356528208089654,
      "tokens_seen": 2171076608
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034553660982948847,
      "loss": 1.4038,
      "theoretical_loss": 3.435644383380861,
      "tokens_seen": 2171142144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034551654964894683,
      "loss": 1.3352,
      "theoretical_loss": 3.4356359462787465,
      "tokens_seen": 2171207680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034549648946840525,
      "loss": 1.3481,
      "theoretical_loss": 3.4356275095025985,
      "tokens_seen": 2171273216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454764292878636,
      "loss": 1.2947,
      "theoretical_loss": 3.4356190730523952,
      "tokens_seen": 2171338752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034545636910732197,
      "loss": 1.3783,
      "theoretical_loss": 3.4356106369281134,
      "tokens_seen": 2171404288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034543630892678033,
      "loss": 1.4503,
      "theoretical_loss": 3.435602201129732,
      "tokens_seen": 2171469824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034541624874623875,
      "loss": 1.3653,
      "theoretical_loss": 3.4355937656572264,
      "tokens_seen": 2171535360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453961885656971,
      "loss": 1.3756,
      "theoretical_loss": 3.435585330510577,
      "tokens_seen": 2171600896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453761283851555,
      "loss": 1.3216,
      "theoretical_loss": 3.435576895689759,
      "tokens_seen": 2171666432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034535606820461384,
      "loss": 1.3816,
      "theoretical_loss": 3.435568461194751,
      "tokens_seen": 2171731968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453360080240722,
      "loss": 1.3697,
      "theoretical_loss": 3.4355600270255304,
      "tokens_seen": 2171797504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453159478435306,
      "loss": 1.4883,
      "theoretical_loss": 3.4355515931820753,
      "tokens_seen": 2171863040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345295887662989,
      "loss": 1.3121,
      "theoretical_loss": 3.4355431596643626,
      "tokens_seen": 2171928576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034527582748244734,
      "loss": 1.3712,
      "theoretical_loss": 3.43553472647237,
      "tokens_seen": 2171994112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452557673019057,
      "loss": 1.3335,
      "theoretical_loss": 3.4355262936060753,
      "tokens_seen": 2172059648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452357071213641,
      "loss": 1.3112,
      "theoretical_loss": 3.435517861065456,
      "tokens_seen": 2172125184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452156469408225,
      "loss": 1.3316,
      "theoretical_loss": 3.43550942885049,
      "tokens_seen": 2172190720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034519558676028084,
      "loss": 1.3152,
      "theoretical_loss": 3.435500996961154,
      "tokens_seen": 2172256256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451755265797392,
      "loss": 1.3333,
      "theoretical_loss": 3.4354925653974266,
      "tokens_seen": 2172321792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034515546639919757,
      "loss": 1.2781,
      "theoretical_loss": 3.435484134159285,
      "tokens_seen": 2172387328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345135406218656,
      "loss": 1.2835,
      "theoretical_loss": 3.435475703246707,
      "tokens_seen": 2172452864
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034511534603811435,
      "loss": 1.4316,
      "theoretical_loss": 3.4354672726596696,
      "tokens_seen": 2172518400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450952858575727,
      "loss": 1.2724,
      "theoretical_loss": 3.4354588423981514,
      "tokens_seen": 2172583936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034507522567703107,
      "loss": 1.3588,
      "theoretical_loss": 3.4354504124621292,
      "tokens_seen": 2172649472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450551654964895,
      "loss": 1.1856,
      "theoretical_loss": 3.4354419828515805,
      "tokens_seen": 2172715008
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034503510531594785,
      "loss": 1.3569,
      "theoretical_loss": 3.435433553566484,
      "tokens_seen": 2172780544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450150451354062,
      "loss": 1.3674,
      "theoretical_loss": 3.435425124606816,
      "tokens_seen": 2172846080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003449949849548646,
      "loss": 1.2325,
      "theoretical_loss": 3.435416695972555,
      "tokens_seen": 2172911616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034497492477432294,
      "loss": 1.3763,
      "theoretical_loss": 3.4354082676636786,
      "tokens_seen": 2172977152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034495486459378135,
      "loss": 1.2315,
      "theoretical_loss": 3.4353998396801635,
      "tokens_seen": 2173042688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003449348044132397,
      "loss": 1.1306,
      "theoretical_loss": 3.4353914120219886,
      "tokens_seen": 2173108224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003449147442326981,
      "loss": 1.3038,
      "theoretical_loss": 3.435382984689131,
      "tokens_seen": 2173173760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034489468405215644,
      "loss": 1.194,
      "theoretical_loss": 3.4353745576815675,
      "tokens_seen": 2173239296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034487462387161486,
      "loss": 1.3488,
      "theoretical_loss": 3.4353661309992773,
      "tokens_seen": 2173304832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448545636910732,
      "loss": 1.4117,
      "theoretical_loss": 3.4353577046422368,
      "tokens_seen": 2173370368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448345035105316,
      "loss": 1.4061,
      "theoretical_loss": 3.435349278610424,
      "tokens_seen": 2173435904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034481444332998994,
      "loss": 1.1974,
      "theoretical_loss": 3.435340852903817,
      "tokens_seen": 2173501440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034479438314944836,
      "loss": 1.35,
      "theoretical_loss": 3.4353324275223933,
      "tokens_seen": 2173566976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447743229689067,
      "loss": 1.2944,
      "theoretical_loss": 3.43532400246613,
      "tokens_seen": 2173632512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447542627883651,
      "loss": 1.3432,
      "theoretical_loss": 3.435315577735005,
      "tokens_seen": 2173698048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034473420260782345,
      "loss": 1.3156,
      "theoretical_loss": 3.4353071533289956,
      "tokens_seen": 2173763584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447141424272818,
      "loss": 1.3341,
      "theoretical_loss": 3.43529872924808,
      "tokens_seen": 2173829120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446940822467402,
      "loss": 1.244,
      "theoretical_loss": 3.435290305492236,
      "tokens_seen": 2173894656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446740220661986,
      "loss": 1.3551,
      "theoretical_loss": 3.4352818820614415,
      "tokens_seen": 2173960192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034465396188565695,
      "loss": 1.2405,
      "theoretical_loss": 3.435273458955673,
      "tokens_seen": 2174025728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446339017051153,
      "loss": 1.2419,
      "theoretical_loss": 3.4352650361749086,
      "tokens_seen": 2174091264
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034461384152457373,
      "loss": 1.381,
      "theoretical_loss": 3.4352566137191265,
      "tokens_seen": 2174156800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034459378134403214,
      "loss": 1.3899,
      "theoretical_loss": 3.435248191588304,
      "tokens_seen": 2174222336
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 791682,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.361458659172058,
      "objective/train/theoretical_loss": 3.435239769782419,
      "objective/train/tokens_used": 544746976,
      "theoretical_loss": 3.435239769782419,
      "tokens_seen": 2174287872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003445737211634905,
      "loss": 1.356,
      "theoretical_loss": 3.435239769782419,
      "tokens_seen": 2174287872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034455366098294887,
      "loss": 1.378,
      "theoretical_loss": 3.4352313483014485,
      "tokens_seen": 2174353408
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034453360080240723,
      "loss": 1.2636,
      "theoretical_loss": 3.4352229271453707,
      "tokens_seen": 2174418944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034451354062186565,
      "loss": 1.3618,
      "theoretical_loss": 3.4352145063141633,
      "tokens_seen": 2174484480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344493480441324,
      "loss": 1.3451,
      "theoretical_loss": 3.435206085807804,
      "tokens_seen": 2174550016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034447342026078237,
      "loss": 1.4435,
      "theoretical_loss": 3.43519766562627,
      "tokens_seen": 2174615552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034445336008024073,
      "loss": 1.3167,
      "theoretical_loss": 3.43518924576954,
      "tokens_seen": 2174681088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034443329989969915,
      "loss": 1.3294,
      "theoretical_loss": 3.4351808262375907,
      "tokens_seen": 2174746624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003444132397191575,
      "loss": 1.2333,
      "theoretical_loss": 3.4351724070304,
      "tokens_seen": 2174812160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443931795386159,
      "loss": 1.3757,
      "theoretical_loss": 3.435163988147946,
      "tokens_seen": 2174877696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034437311935807424,
      "loss": 1.3469,
      "theoretical_loss": 3.4351555695902056,
      "tokens_seen": 2174943232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443530591775326,
      "loss": 1.3429,
      "theoretical_loss": 3.435147151357157,
      "tokens_seen": 2175008768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344332998996991,
      "loss": 1.3447,
      "theoretical_loss": 3.4351387334487784,
      "tokens_seen": 2175074304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443129388164494,
      "loss": 1.3634,
      "theoretical_loss": 3.4351303158650466,
      "tokens_seen": 2175139840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034429287863590774,
      "loss": 1.2966,
      "theoretical_loss": 3.4351218986059395,
      "tokens_seen": 2175205376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442728184553661,
      "loss": 1.2259,
      "theoretical_loss": 3.4351134816714355,
      "tokens_seen": 2175270912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442527582748245,
      "loss": 1.2976,
      "theoretical_loss": 3.4351050650615114,
      "tokens_seen": 2175336448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442326980942829,
      "loss": 1.2581,
      "theoretical_loss": 3.4350966487761454,
      "tokens_seen": 2175401984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034421263791374124,
      "loss": 1.3174,
      "theoretical_loss": 3.435088232815315,
      "tokens_seen": 2175467520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441925777331996,
      "loss": 1.168,
      "theoretical_loss": 3.4350798171789982,
      "tokens_seen": 2175533056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034417251755265797,
      "loss": 1.2699,
      "theoretical_loss": 3.435071401867172,
      "tokens_seen": 2175598592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441524573721164,
      "loss": 1.3749,
      "theoretical_loss": 3.435062986879815,
      "tokens_seen": 2175664128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034413239719157475,
      "loss": 1.3102,
      "theoretical_loss": 3.4350545722169046,
      "tokens_seen": 2175729664
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441123370110331,
      "loss": 1.2661,
      "theoretical_loss": 3.4350461578784186,
      "tokens_seen": 2175795200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034409227683049147,
      "loss": 1.299,
      "theoretical_loss": 3.435037743864334,
      "tokens_seen": 2175860736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003440722166499499,
      "loss": 1.287,
      "theoretical_loss": 3.4350293301746295,
      "tokens_seen": 2175926272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034405215646940825,
      "loss": 1.3887,
      "theoretical_loss": 3.4350209168092825,
      "tokens_seen": 2175991808
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003440320962888666,
      "loss": 1.1957,
      "theoretical_loss": 3.43501250376827,
      "tokens_seen": 2176057344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344012036108325,
      "loss": 1.3083,
      "theoretical_loss": 3.4350040910515713,
      "tokens_seen": 2176122880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034399197592778334,
      "loss": 1.2747,
      "theoretical_loss": 3.434995678659163,
      "tokens_seen": 2176188416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034397191574724175,
      "loss": 1.1947,
      "theoretical_loss": 3.4349872665910226,
      "tokens_seen": 2176253952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003439518555667001,
      "loss": 1.2829,
      "theoretical_loss": 3.4349788548471287,
      "tokens_seen": 2176319488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003439317953861585,
      "loss": 1.3004,
      "theoretical_loss": 3.4349704434274586,
      "tokens_seen": 2176385024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034391173520561684,
      "loss": 1.403,
      "theoretical_loss": 3.43496203233199,
      "tokens_seen": 2176450560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034389167502507526,
      "loss": 1.3556,
      "theoretical_loss": 3.4349536215607004,
      "tokens_seen": 2176516096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438716148445336,
      "loss": 1.2716,
      "theoretical_loss": 3.4349452111135683,
      "tokens_seen": 2176581632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343851554663992,
      "loss": 1.3452,
      "theoretical_loss": 3.4349368009905707,
      "tokens_seen": 2176647168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034383149448345034,
      "loss": 1.2354,
      "theoretical_loss": 3.4349283911916855,
      "tokens_seen": 2176712704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034381143430290876,
      "loss": 1.2828,
      "theoretical_loss": 3.4349199817168907,
      "tokens_seen": 2176778240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437913741223671,
      "loss": 1.3771,
      "theoretical_loss": 3.4349115725661643,
      "tokens_seen": 2176843776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437713139418255,
      "loss": 1.2748,
      "theoretical_loss": 3.4349031637394836,
      "tokens_seen": 2176909312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034375125376128385,
      "loss": 1.4442,
      "theoretical_loss": 3.4348947552368263,
      "tokens_seen": 2176974848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437311935807422,
      "loss": 1.3574,
      "theoretical_loss": 3.4348863470581703,
      "tokens_seen": 2177040384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437111334002006,
      "loss": 1.364,
      "theoretical_loss": 3.434877939203494,
      "tokens_seen": 2177105920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343691073219659,
      "loss": 1.3288,
      "theoretical_loss": 3.4348695316727733,
      "tokens_seen": 2177171456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034367101303911735,
      "loss": 1.3275,
      "theoretical_loss": 3.4348611244659883,
      "tokens_seen": 2177236992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003436509528585757,
      "loss": 1.433,
      "theoretical_loss": 3.4348527175831154,
      "tokens_seen": 2177302528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034363089267803413,
      "loss": 1.3813,
      "theoretical_loss": 3.4348443110241327,
      "tokens_seen": 2177368064
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003436108324974925,
      "loss": 1.395,
      "theoretical_loss": 3.4348359047890176,
      "tokens_seen": 2177433600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034359077231695085,
      "loss": 1.2783,
      "theoretical_loss": 3.4348274988777483,
      "tokens_seen": 2177499136
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 792987,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.399841070175171,
      "objective/train/theoretical_loss": 3.4348190932903027,
      "objective/train/tokens_used": 548023776,
      "theoretical_loss": 3.4348190932903027,
      "tokens_seen": 2177564672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435707121364092,
      "loss": 1.3861,
      "theoretical_loss": 3.4348190932903027,
      "tokens_seen": 2177564672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435506519558676,
      "loss": 1.3846,
      "theoretical_loss": 3.4348106880266585,
      "tokens_seen": 2177630208
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343530591775326,
      "loss": 1.3781,
      "theoretical_loss": 3.434802283086793,
      "tokens_seen": 2177695744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034351053159478436,
      "loss": 1.3168,
      "theoretical_loss": 3.4347938784706846,
      "tokens_seen": 2177761280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434904714142427,
      "loss": 1.2561,
      "theoretical_loss": 3.4347854741783106,
      "tokens_seen": 2177826816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434704112337011,
      "loss": 1.3239,
      "theoretical_loss": 3.434777070209649,
      "tokens_seen": 2177892352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434503510531595,
      "loss": 1.3563,
      "theoretical_loss": 3.4347686665646777,
      "tokens_seen": 2177957888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034343029087261786,
      "loss": 1.2932,
      "theoretical_loss": 3.4347602632433745,
      "tokens_seen": 2178023424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434102306920762,
      "loss": 1.3324,
      "theoretical_loss": 3.4347518602457168,
      "tokens_seen": 2178088960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433901705115346,
      "loss": 1.2846,
      "theoretical_loss": 3.434743457571683,
      "tokens_seen": 2178154496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034337011033099295,
      "loss": 1.3624,
      "theoretical_loss": 3.4347350552212506,
      "tokens_seen": 2178220032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034335005015045136,
      "loss": 1.2685,
      "theoretical_loss": 3.434726653194397,
      "tokens_seen": 2178285568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433299899699097,
      "loss": 1.3453,
      "theoretical_loss": 3.4347182514911005,
      "tokens_seen": 2178351104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433099297893681,
      "loss": 1.2145,
      "theoretical_loss": 3.434709850111339,
      "tokens_seen": 2178416640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034328986960882645,
      "loss": 1.2773,
      "theoretical_loss": 3.43470144905509,
      "tokens_seen": 2178482176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034326980942828487,
      "loss": 1.3635,
      "theoretical_loss": 3.4346930483223312,
      "tokens_seen": 2178547712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034324974924774323,
      "loss": 1.3572,
      "theoretical_loss": 3.4346846479130413,
      "tokens_seen": 2178613248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432296890672016,
      "loss": 1.3026,
      "theoretical_loss": 3.434676247827197,
      "tokens_seen": 2178678784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034320962888665995,
      "loss": 1.3476,
      "theoretical_loss": 3.4346678480647768,
      "tokens_seen": 2178744320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431895687061183,
      "loss": 1.2264,
      "theoretical_loss": 3.434659448625758,
      "tokens_seen": 2178809856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034316950852557673,
      "loss": 1.2787,
      "theoretical_loss": 3.4346510495101183,
      "tokens_seen": 2178875392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431494483450351,
      "loss": 1.2984,
      "theoretical_loss": 3.4346426507178367,
      "tokens_seen": 2178940928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034312938816449345,
      "loss": 1.3013,
      "theoretical_loss": 3.4346342522488897,
      "tokens_seen": 2179006464
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431093279839518,
      "loss": 1.3285,
      "theoretical_loss": 3.434625854103256,
      "tokens_seen": 2179072000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430892678034103,
      "loss": 1.3527,
      "theoretical_loss": 3.4346174562809133,
      "tokens_seen": 2179137536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034306920762286865,
      "loss": 1.3568,
      "theoretical_loss": 3.434609058781839,
      "tokens_seen": 2179203072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343049147442327,
      "loss": 1.1975,
      "theoretical_loss": 3.434600661606011,
      "tokens_seen": 2179268608
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430290872617854,
      "loss": 1.323,
      "theoretical_loss": 3.4345922647534075,
      "tokens_seen": 2179334144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034300902708124374,
      "loss": 1.3763,
      "theoretical_loss": 3.4345838682240064,
      "tokens_seen": 2179399680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034298896690070215,
      "loss": 1.2692,
      "theoretical_loss": 3.434575472017785,
      "tokens_seen": 2179465216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429689067201605,
      "loss": 1.3164,
      "theoretical_loss": 3.4345670761347216,
      "tokens_seen": 2179530752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429488465396189,
      "loss": 1.3078,
      "theoretical_loss": 3.434558680574794,
      "tokens_seen": 2179596288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034292878635907724,
      "loss": 1.3579,
      "theoretical_loss": 3.4345502853379792,
      "tokens_seen": 2179661824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034290872617853566,
      "loss": 1.28,
      "theoretical_loss": 3.4345418904242564,
      "tokens_seen": 2179727360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342888665997994,
      "loss": 1.3472,
      "theoretical_loss": 3.4345334958336027,
      "tokens_seen": 2179792896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428686058174524,
      "loss": 1.2463,
      "theoretical_loss": 3.4345251015659963,
      "tokens_seen": 2179858432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034284854563691074,
      "loss": 1.3217,
      "theoretical_loss": 3.4345167076214143,
      "tokens_seen": 2179923968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428284854563691,
      "loss": 1.3139,
      "theoretical_loss": 3.4345083139998356,
      "tokens_seen": 2179989504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428084252758275,
      "loss": 1.3851,
      "theoretical_loss": 3.4344999207012377,
      "tokens_seen": 2180055040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427883650952859,
      "loss": 1.427,
      "theoretical_loss": 3.434491527725598,
      "tokens_seen": 2180120576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034276830491474425,
      "loss": 1.4454,
      "theoretical_loss": 3.4344831350728944,
      "tokens_seen": 2180186112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427482447342026,
      "loss": 1.4002,
      "theoretical_loss": 3.434474742743106,
      "tokens_seen": 2180251648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342728184553661,
      "loss": 1.2849,
      "theoretical_loss": 3.4344663507362085,
      "tokens_seen": 2180317184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427081243731194,
      "loss": 1.3261,
      "theoretical_loss": 3.4344579590521818,
      "tokens_seen": 2180382720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034268806419257775,
      "loss": 1.4017,
      "theoretical_loss": 3.434449567691003,
      "tokens_seen": 2180448256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426680040120361,
      "loss": 1.3601,
      "theoretical_loss": 3.43444117665265,
      "tokens_seen": 2180513792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034264794383149453,
      "loss": 1.3857,
      "theoretical_loss": 3.4344327859371004,
      "tokens_seen": 2180579328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426278836509529,
      "loss": 1.282,
      "theoretical_loss": 3.4344243955443323,
      "tokens_seen": 2180644864
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034260782347041125,
      "loss": 1.4175,
      "theoretical_loss": 3.4344160054743234,
      "tokens_seen": 2180710400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425877632898696,
      "loss": 1.3413,
      "theoretical_loss": 3.434407615727052,
      "tokens_seen": 2180775936
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 794081,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3060002326965332,
      "objective/train/theoretical_loss": 3.434399226302496,
      "objective/train/tokens_used": 551300576,
      "theoretical_loss": 3.434399226302496,
      "tokens_seen": 2180841472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342567703109328,
      "loss": 1.3029,
      "theoretical_loss": 3.434399226302496,
      "tokens_seen": 2180841472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425476429287864,
      "loss": 1.3117,
      "theoretical_loss": 3.434390837200633,
      "tokens_seen": 2180907008
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034252758274824476,
      "loss": 1.3131,
      "theoretical_loss": 3.4343824484214407,
      "tokens_seen": 2180972544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425075225677031,
      "loss": 1.329,
      "theoretical_loss": 3.434374059964897,
      "tokens_seen": 2181038080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424874623871615,
      "loss": 1.351,
      "theoretical_loss": 3.4343656718309807,
      "tokens_seen": 2181103616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424674022066199,
      "loss": 1.2812,
      "theoretical_loss": 3.4343572840196686,
      "tokens_seen": 2181169152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034244734202607826,
      "loss": 1.3166,
      "theoretical_loss": 3.434348896530939,
      "tokens_seen": 2181234688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424272818455366,
      "loss": 1.265,
      "theoretical_loss": 3.4343405093647705,
      "tokens_seen": 2181300224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342407221664995,
      "loss": 1.2505,
      "theoretical_loss": 3.4343321225211394,
      "tokens_seen": 2181365760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034238716148445335,
      "loss": 1.389,
      "theoretical_loss": 3.4343237360000254,
      "tokens_seen": 2181431296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034236710130391176,
      "loss": 1.3569,
      "theoretical_loss": 3.434315349801405,
      "tokens_seen": 2181496832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423470411233701,
      "loss": 1.3176,
      "theoretical_loss": 3.434306963925257,
      "tokens_seen": 2181562368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423269809428285,
      "loss": 1.2503,
      "theoretical_loss": 3.4342985783715587,
      "tokens_seen": 2181627904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034230692076228685,
      "loss": 1.4076,
      "theoretical_loss": 3.4342901931402885,
      "tokens_seen": 2181693440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034228686058174527,
      "loss": 1.3911,
      "theoretical_loss": 3.434281808231424,
      "tokens_seen": 2181758976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034226680040120363,
      "loss": 1.3134,
      "theoretical_loss": 3.4342734236449433,
      "tokens_seen": 2181824512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342246740220662,
      "loss": 1.247,
      "theoretical_loss": 3.434265039380824,
      "tokens_seen": 2181890048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034222668004012035,
      "loss": 1.3527,
      "theoretical_loss": 3.434256655439045,
      "tokens_seen": 2181955584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003422066198595787,
      "loss": 1.3526,
      "theoretical_loss": 3.4342482718195826,
      "tokens_seen": 2182021120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034218655967903713,
      "loss": 1.27,
      "theoretical_loss": 3.4342398885224164,
      "tokens_seen": 2182086656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421664994984955,
      "loss": 1.255,
      "theoretical_loss": 3.434231505547523,
      "tokens_seen": 2182152192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034214643931795386,
      "loss": 1.2233,
      "theoretical_loss": 3.434223122894881,
      "tokens_seen": 2182217728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421263791374122,
      "loss": 1.3433,
      "theoretical_loss": 3.4342147405644683,
      "tokens_seen": 2182283264
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034210631895687063,
      "loss": 1.3015,
      "theoretical_loss": 3.434206358556263,
      "tokens_seen": 2182348800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342086258776329,
      "loss": 1.3572,
      "theoretical_loss": 3.4341979768702426,
      "tokens_seen": 2182414336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034206619859578736,
      "loss": 1.3685,
      "theoretical_loss": 3.4341895955063855,
      "tokens_seen": 2182479872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420461384152457,
      "loss": 1.34,
      "theoretical_loss": 3.434181214464669,
      "tokens_seen": 2182545408
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420260782347041,
      "loss": 1.3865,
      "theoretical_loss": 3.434172833745072,
      "tokens_seen": 2182610944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420060180541625,
      "loss": 1.3439,
      "theoretical_loss": 3.4341644533475715,
      "tokens_seen": 2182676480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034198595787362086,
      "loss": 1.3034,
      "theoretical_loss": 3.434156073272146,
      "tokens_seen": 2182742016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003419658976930792,
      "loss": 1.2918,
      "theoretical_loss": 3.434147693518773,
      "tokens_seen": 2182807552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003419458375125376,
      "loss": 1.3659,
      "theoretical_loss": 3.434139314087431,
      "tokens_seen": 2182873088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341925777331996,
      "loss": 1.3824,
      "theoretical_loss": 3.4341309349780977,
      "tokens_seen": 2182938624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034190571715145436,
      "loss": 1.344,
      "theoretical_loss": 3.434122556190751,
      "tokens_seen": 2183004160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418856569709127,
      "loss": 1.2866,
      "theoretical_loss": 3.4341141777253688,
      "tokens_seen": 2183069696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418655967903711,
      "loss": 1.3307,
      "theoretical_loss": 3.4341057995819293,
      "tokens_seen": 2183135232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034184553660982945,
      "loss": 1.2127,
      "theoretical_loss": 3.4340974217604106,
      "tokens_seen": 2183200768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034182547642928787,
      "loss": 1.3412,
      "theoretical_loss": 3.4340890442607903,
      "tokens_seen": 2183266304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034180541624874623,
      "loss": 1.2727,
      "theoretical_loss": 3.4340806670830464,
      "tokens_seen": 2183331840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417853560682046,
      "loss": 1.3589,
      "theoretical_loss": 3.434072290227157,
      "tokens_seen": 2183397376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034176529588766295,
      "loss": 1.3892,
      "theoretical_loss": 3.4340639136931,
      "tokens_seen": 2183462912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034174523570712137,
      "loss": 1.2902,
      "theoretical_loss": 3.4340555374808535,
      "tokens_seen": 2183528448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034172517552657973,
      "loss": 1.3418,
      "theoretical_loss": 3.4340471615903954,
      "tokens_seen": 2183593984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417051153460381,
      "loss": 1.3557,
      "theoretical_loss": 3.4340387860217034,
      "tokens_seen": 2183659520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034168505516549646,
      "loss": 1.3078,
      "theoretical_loss": 3.434030410774756,
      "tokens_seen": 2183725056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416649949849548,
      "loss": 1.3491,
      "theoretical_loss": 3.434022035849531,
      "tokens_seen": 2183790592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034164493480441324,
      "loss": 1.2932,
      "theoretical_loss": 3.434013661246006,
      "tokens_seen": 2183856128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416248746238716,
      "loss": 1.359,
      "theoretical_loss": 3.43400528696416,
      "tokens_seen": 2183921664
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034160481444332996,
      "loss": 1.2571,
      "theoretical_loss": 3.43399691300397,
      "tokens_seen": 2183987200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415847542627884,
      "loss": 1.4629,
      "theoretical_loss": 3.433988539365414,
      "tokens_seen": 2184052736
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 795264,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.33791184425354,
      "objective/train/theoretical_loss": 3.43398016604847,
      "objective/train/tokens_used": 554577376,
      "theoretical_loss": 3.43398016604847,
      "tokens_seen": 2184118272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415646940822468,
      "loss": 1.3418,
      "theoretical_loss": 3.43398016604847,
      "tokens_seen": 2184118272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034154463390170516,
      "loss": 1.3694,
      "theoretical_loss": 3.433971793053117,
      "tokens_seen": 2184183808
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415245737211635,
      "loss": 1.3126,
      "theoretical_loss": 3.433963420379332,
      "tokens_seen": 2184249344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415045135406219,
      "loss": 1.4299,
      "theoretical_loss": 3.4339550480270935,
      "tokens_seen": 2184314880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414844533600803,
      "loss": 1.3505,
      "theoretical_loss": 3.433946675996379,
      "tokens_seen": 2184380416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034146439317953866,
      "loss": 1.4095,
      "theoretical_loss": 3.4339383042871674,
      "tokens_seen": 2184445952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341444332998997,
      "loss": 1.295,
      "theoretical_loss": 3.4339299328994355,
      "tokens_seen": 2184511488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414242728184554,
      "loss": 1.3499,
      "theoretical_loss": 3.433921561833162,
      "tokens_seen": 2184577024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034140421263791375,
      "loss": 1.3926,
      "theoretical_loss": 3.433913191088325,
      "tokens_seen": 2184642560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034138415245737216,
      "loss": 1.3065,
      "theoretical_loss": 3.4339048206649023,
      "tokens_seen": 2184708096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003413640922768305,
      "loss": 1.3218,
      "theoretical_loss": 3.433896450562872,
      "tokens_seen": 2184773632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003413440320962889,
      "loss": 1.2751,
      "theoretical_loss": 3.4338880807822125,
      "tokens_seen": 2184839168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034132397191574725,
      "loss": 1.3726,
      "theoretical_loss": 3.4338797113229007,
      "tokens_seen": 2184904704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034130391173520567,
      "loss": 1.2598,
      "theoretical_loss": 3.433871342184916,
      "tokens_seen": 2184970240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034128385155466403,
      "loss": 1.3376,
      "theoretical_loss": 3.4338629733682353,
      "tokens_seen": 2185035776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412637913741224,
      "loss": 1.284,
      "theoretical_loss": 3.4338546048728373,
      "tokens_seen": 2185101312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034124373119358075,
      "loss": 1.3145,
      "theoretical_loss": 3.4338462366986997,
      "tokens_seen": 2185166848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412236710130391,
      "loss": 1.3778,
      "theoretical_loss": 3.433837868845801,
      "tokens_seen": 2185232384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034120361083249753,
      "loss": 1.3896,
      "theoretical_loss": 3.4338295013141185,
      "tokens_seen": 2185297920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411835506519559,
      "loss": 1.333,
      "theoretical_loss": 3.433821134103631,
      "tokens_seen": 2185363456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034116349047141426,
      "loss": 1.3688,
      "theoretical_loss": 3.4338127672143157,
      "tokens_seen": 2185428992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411434302908726,
      "loss": 1.279,
      "theoretical_loss": 3.4338044006461512,
      "tokens_seen": 2185494528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034112337011033103,
      "loss": 1.3386,
      "theoretical_loss": 3.4337960343991156,
      "tokens_seen": 2185560064
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411033099297894,
      "loss": 1.4406,
      "theoretical_loss": 3.4337876684731867,
      "tokens_seen": 2185625600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034108324974924776,
      "loss": 1.3079,
      "theoretical_loss": 3.4337793028683428,
      "tokens_seen": 2185691136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410631895687061,
      "loss": 1.3677,
      "theoretical_loss": 3.433770937584562,
      "tokens_seen": 2185756672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410431293881645,
      "loss": 1.3359,
      "theoretical_loss": 3.4337625726218217,
      "tokens_seen": 2185822208
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410230692076229,
      "loss": 1.3223,
      "theoretical_loss": 3.4337542079801007,
      "tokens_seen": 2185887744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034100300902708126,
      "loss": 1.4324,
      "theoretical_loss": 3.4337458436593766,
      "tokens_seen": 2185953280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003409829488465396,
      "loss": 1.4117,
      "theoretical_loss": 3.433737479659628,
      "tokens_seen": 2186018816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000340962888665998,
      "loss": 1.286,
      "theoretical_loss": 3.433729115980832,
      "tokens_seen": 2186084352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003409428284854564,
      "loss": 1.4227,
      "theoretical_loss": 3.4337207526229676,
      "tokens_seen": 2186149888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034092276830491477,
      "loss": 1.3973,
      "theoretical_loss": 3.433712389586012,
      "tokens_seen": 2186215424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034090270812437313,
      "loss": 1.267,
      "theoretical_loss": 3.4337040268699446,
      "tokens_seen": 2186280960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408826479438315,
      "loss": 1.4434,
      "theoretical_loss": 3.433695664474742,
      "tokens_seen": 2186346496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034086258776328985,
      "loss": 1.3399,
      "theoretical_loss": 3.4336873024003833,
      "tokens_seen": 2186412032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034084252758274827,
      "loss": 1.4267,
      "theoretical_loss": 3.433678940646846,
      "tokens_seen": 2186477568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034082246740220663,
      "loss": 1.2892,
      "theoretical_loss": 3.4336705792141085,
      "tokens_seen": 2186543104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000340802407221665,
      "loss": 1.2968,
      "theoretical_loss": 3.4336622181021488,
      "tokens_seen": 2186608640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034078234704112335,
      "loss": 1.292,
      "theoretical_loss": 3.4336538573109445,
      "tokens_seen": 2186674176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034076228686058177,
      "loss": 1.3133,
      "theoretical_loss": 3.4336454968404744,
      "tokens_seen": 2186739712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034074222668004013,
      "loss": 1.3695,
      "theoretical_loss": 3.4336371366907166,
      "tokens_seen": 2186805248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003407221664994985,
      "loss": 1.4023,
      "theoretical_loss": 3.4336287768616485,
      "tokens_seen": 2186870784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034070210631895686,
      "loss": 1.3003,
      "theoretical_loss": 3.4336204173532483,
      "tokens_seen": 2186936320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406820461384152,
      "loss": 1.3085,
      "theoretical_loss": 3.4336120581654948,
      "tokens_seen": 2187001856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034066198595787364,
      "loss": 1.376,
      "theoretical_loss": 3.4336036992983656,
      "tokens_seen": 2187067392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000340641925777332,
      "loss": 1.4446,
      "theoretical_loss": 3.4335953407518387,
      "tokens_seen": 2187132928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034062186559679036,
      "loss": 1.3059,
      "theoretical_loss": 3.4335869825258922,
      "tokens_seen": 2187198464
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406018054162487,
      "loss": 1.3361,
      "theoretical_loss": 3.4335786246205044,
      "tokens_seen": 2187264000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034058174523570714,
      "loss": 1.3258,
      "theoretical_loss": 3.4335702670356536,
      "tokens_seen": 2187329536
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 796380,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4696393013000488,
      "objective/train/theoretical_loss": 3.4335619097713175,
      "objective/train/tokens_used": 557854176,
      "theoretical_loss": 3.4335619097713175,
      "tokens_seen": 2187395072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405616850551655,
      "loss": 1.4422,
      "theoretical_loss": 3.4335619097713175,
      "tokens_seen": 2187395072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034054162487462386,
      "loss": 1.3469,
      "theoretical_loss": 3.4335535528274743,
      "tokens_seen": 2187460608
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405215646940822,
      "loss": 1.2971,
      "theoretical_loss": 3.4335451962041024,
      "tokens_seen": 2187526144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034050150451354064,
      "loss": 1.3619,
      "theoretical_loss": 3.4335368399011794,
      "tokens_seen": 2187591680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000340481444332999,
      "loss": 1.3903,
      "theoretical_loss": 3.4335284839186837,
      "tokens_seen": 2187657216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034046138415245737,
      "loss": 1.3446,
      "theoretical_loss": 3.4335201282565935,
      "tokens_seen": 2187722752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034044132397191573,
      "loss": 1.2827,
      "theoretical_loss": 3.433511772914887,
      "tokens_seen": 2187788288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404212637913741,
      "loss": 1.3151,
      "theoretical_loss": 3.433503417893542,
      "tokens_seen": 2187853824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404012036108325,
      "loss": 1.3519,
      "theoretical_loss": 3.4334950631925363,
      "tokens_seen": 2187919360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034038114343029087,
      "loss": 1.3695,
      "theoretical_loss": 3.433486708811849,
      "tokens_seen": 2187984896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034036108324974923,
      "loss": 1.3254,
      "theoretical_loss": 3.4334783547514576,
      "tokens_seen": 2188050432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003403410230692076,
      "loss": 1.3506,
      "theoretical_loss": 3.43347000101134,
      "tokens_seen": 2188115968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000340320962888666,
      "loss": 1.3177,
      "theoretical_loss": 3.4334616475914754,
      "tokens_seen": 2188181504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003403009027081244,
      "loss": 1.3301,
      "theoretical_loss": 3.433453294491841,
      "tokens_seen": 2188247040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034028084252758274,
      "loss": 1.2995,
      "theoretical_loss": 3.4334449417124144,
      "tokens_seen": 2188312576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003402607823470411,
      "loss": 1.4326,
      "theoretical_loss": 3.433436589253175,
      "tokens_seen": 2188378112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034024072216649946,
      "loss": 1.3466,
      "theoretical_loss": 3.4334282371141,
      "tokens_seen": 2188443648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003402206619859579,
      "loss": 1.292,
      "theoretical_loss": 3.4334198852951685,
      "tokens_seen": 2188509184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034020060180541624,
      "loss": 1.4296,
      "theoretical_loss": 3.4334115337963578,
      "tokens_seen": 2188574720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401805416248746,
      "loss": 1.3808,
      "theoretical_loss": 3.4334031826176465,
      "tokens_seen": 2188640256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034016048144433296,
      "loss": 1.3303,
      "theoretical_loss": 3.433394831759012,
      "tokens_seen": 2188705792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401404212637914,
      "loss": 1.3042,
      "theoretical_loss": 3.4333864812204338,
      "tokens_seen": 2188771328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034012036108324974,
      "loss": 1.3298,
      "theoretical_loss": 3.4333781310018887,
      "tokens_seen": 2188836864
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401003009027081,
      "loss": 1.3884,
      "theoretical_loss": 3.4333697811033557,
      "tokens_seen": 2188902400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003400802407221665,
      "loss": 1.3479,
      "theoretical_loss": 3.433361431524813,
      "tokens_seen": 2188967936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003400601805416249,
      "loss": 1.3489,
      "theoretical_loss": 3.4333530822662377,
      "tokens_seen": 2189033472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003400401203610833,
      "loss": 1.3704,
      "theoretical_loss": 3.4333447333276093,
      "tokens_seen": 2189099008
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034002006018054166,
      "loss": 1.3014,
      "theoretical_loss": 3.433336384708905,
      "tokens_seen": 2189164544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034,
      "loss": 1.3024,
      "theoretical_loss": 3.4333280364101038,
      "tokens_seen": 2189230080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003399799398194584,
      "loss": 1.3214,
      "theoretical_loss": 3.4333196884311827,
      "tokens_seen": 2189295616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003399598796389168,
      "loss": 1.186,
      "theoretical_loss": 3.433311340772121,
      "tokens_seen": 2189361152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033993981945837517,
      "loss": 1.3232,
      "theoretical_loss": 3.4333029934328962,
      "tokens_seen": 2189426688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033991975927783353,
      "loss": 1.2205,
      "theoretical_loss": 3.433294646413487,
      "tokens_seen": 2189492224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398996990972919,
      "loss": 1.3335,
      "theoretical_loss": 3.433286299713871,
      "tokens_seen": 2189557760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033987963891675025,
      "loss": 1.3321,
      "theoretical_loss": 3.4332779533340263,
      "tokens_seen": 2189623296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033985957873620867,
      "loss": 1.4256,
      "theoretical_loss": 3.433269607273932,
      "tokens_seen": 2189688832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033983951855566703,
      "loss": 1.2873,
      "theoretical_loss": 3.433261261533566,
      "tokens_seen": 2189754368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398194583751254,
      "loss": 1.3906,
      "theoretical_loss": 3.4332529161129055,
      "tokens_seen": 2189819904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033979939819458376,
      "loss": 1.311,
      "theoretical_loss": 3.4332445710119295,
      "tokens_seen": 2189885440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033977933801404217,
      "loss": 1.3967,
      "theoretical_loss": 3.4332362262306164,
      "tokens_seen": 2189950976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033975927783350053,
      "loss": 1.2852,
      "theoretical_loss": 3.4332278817689437,
      "tokens_seen": 2190016512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397392176529589,
      "loss": 1.3093,
      "theoretical_loss": 3.43321953762689,
      "tokens_seen": 2190082048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033971915747241726,
      "loss": 1.4804,
      "theoretical_loss": 3.433211193804434,
      "tokens_seen": 2190147584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396990972918756,
      "loss": 1.3167,
      "theoretical_loss": 3.4332028503015524,
      "tokens_seen": 2190213120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033967903711133404,
      "loss": 1.3219,
      "theoretical_loss": 3.433194507118225,
      "tokens_seen": 2190278656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396589769307924,
      "loss": 1.3668,
      "theoretical_loss": 3.4331861642544292,
      "tokens_seen": 2190344192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033963891675025076,
      "loss": 1.3353,
      "theoretical_loss": 3.433177821710143,
      "tokens_seen": 2190409728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396188565697091,
      "loss": 1.2263,
      "theoretical_loss": 3.4331694794853456,
      "tokens_seen": 2190475264
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033959879638916754,
      "loss": 1.3763,
      "theoretical_loss": 3.433161137580014,
      "tokens_seen": 2190540800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395787362086259,
      "loss": 1.2612,
      "theoretical_loss": 3.4331527959941273,
      "tokens_seen": 2190606336
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 797544,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4318948984146118,
      "objective/train/theoretical_loss": 3.4331444547276635,
      "objective/train/tokens_used": 561130976,
      "theoretical_loss": 3.4331444547276635,
      "tokens_seen": 2190671872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033955867602808426,
      "loss": 1.3687,
      "theoretical_loss": 3.4331444547276635,
      "tokens_seen": 2190671872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395386158475426,
      "loss": 1.2755,
      "theoretical_loss": 3.4331361137806002,
      "tokens_seen": 2190737408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033951855566700104,
      "loss": 1.3143,
      "theoretical_loss": 3.433127773152916,
      "tokens_seen": 2190802944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394984954864594,
      "loss": 1.3477,
      "theoretical_loss": 3.43311943284459,
      "tokens_seen": 2190868480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033947843530591777,
      "loss": 1.3112,
      "theoretical_loss": 3.433111092855599,
      "tokens_seen": 2190934016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033945837512537613,
      "loss": 1.3379,
      "theoretical_loss": 3.433102753185922,
      "tokens_seen": 2190999552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394383149448345,
      "loss": 1.3878,
      "theoretical_loss": 3.433094413835537,
      "tokens_seen": 2191065088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394182547642929,
      "loss": 1.1772,
      "theoretical_loss": 3.4330860748044225,
      "tokens_seen": 2191130624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033939819458375127,
      "loss": 1.4139,
      "theoretical_loss": 3.4330777360925566,
      "tokens_seen": 2191196160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033937813440320963,
      "loss": 1.3543,
      "theoretical_loss": 3.433069397699917,
      "tokens_seen": 2191261696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339358074222668,
      "loss": 1.2785,
      "theoretical_loss": 3.433061059626483,
      "tokens_seen": 2191327232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393380140421264,
      "loss": 1.2888,
      "theoretical_loss": 3.433052721872232,
      "tokens_seen": 2191392768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393179538615848,
      "loss": 1.351,
      "theoretical_loss": 3.4330443844371423,
      "tokens_seen": 2191458304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033929789368104314,
      "loss": 1.2236,
      "theoretical_loss": 3.4330360473211923,
      "tokens_seen": 2191523840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003392778335005015,
      "loss": 1.3972,
      "theoretical_loss": 3.4330277105243603,
      "tokens_seen": 2191589376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033925777331995986,
      "loss": 1.2715,
      "theoretical_loss": 3.4330193740466246,
      "tokens_seen": 2191654912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003392377131394183,
      "loss": 1.3,
      "theoretical_loss": 3.4330110378879635,
      "tokens_seen": 2191720448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033921765295887664,
      "loss": 1.4087,
      "theoretical_loss": 3.4330027020483547,
      "tokens_seen": 2191785984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339197592778335,
      "loss": 1.3882,
      "theoretical_loss": 3.432994366527777,
      "tokens_seen": 2191851520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033917753259779336,
      "loss": 1.3216,
      "theoretical_loss": 3.432986031326209,
      "tokens_seen": 2191917056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391574724172518,
      "loss": 1.3639,
      "theoretical_loss": 3.4329776964436274,
      "tokens_seen": 2191982592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033913741223671014,
      "loss": 1.2528,
      "theoretical_loss": 3.432969361880012,
      "tokens_seen": 2192048128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391173520561685,
      "loss": 1.3301,
      "theoretical_loss": 3.432961027635341,
      "tokens_seen": 2192113664
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033909729187562687,
      "loss": 1.3384,
      "theoretical_loss": 3.4329526937095913,
      "tokens_seen": 2192179200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033907723169508523,
      "loss": 1.4283,
      "theoretical_loss": 3.4329443601027427,
      "tokens_seen": 2192244736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033905717151454365,
      "loss": 1.2136,
      "theoretical_loss": 3.432936026814773,
      "tokens_seen": 2192310272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339037111334002,
      "loss": 1.2605,
      "theoretical_loss": 3.43292769384566,
      "tokens_seen": 2192375808
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033901705115346037,
      "loss": 1.3571,
      "theoretical_loss": 3.432919361195382,
      "tokens_seen": 2192441344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033899699097291873,
      "loss": 1.2635,
      "theoretical_loss": 3.4329110288639177,
      "tokens_seen": 2192506880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033897693079237715,
      "loss": 1.3853,
      "theoretical_loss": 3.4329026968512455,
      "tokens_seen": 2192572416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003389568706118355,
      "loss": 1.2294,
      "theoretical_loss": 3.432894365157343,
      "tokens_seen": 2192637952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003389368104312939,
      "loss": 1.2674,
      "theoretical_loss": 3.4328860337821894,
      "tokens_seen": 2192703488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033891675025075224,
      "loss": 1.357,
      "theoretical_loss": 3.432877702725762,
      "tokens_seen": 2192769024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388966900702106,
      "loss": 1.3789,
      "theoretical_loss": 3.4328693719880397,
      "tokens_seen": 2192834560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000338876629889669,
      "loss": 1.3345,
      "theoretical_loss": 3.4328610415690006,
      "tokens_seen": 2192900096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388565697091274,
      "loss": 1.2225,
      "theoretical_loss": 3.432852711468623,
      "tokens_seen": 2192965632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033883650952858574,
      "loss": 1.319,
      "theoretical_loss": 3.432844381686885,
      "tokens_seen": 2193031168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388164493480441,
      "loss": 1.2773,
      "theoretical_loss": 3.4328360522237653,
      "tokens_seen": 2193096704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387963891675025,
      "loss": 1.3942,
      "theoretical_loss": 3.4328277230792423,
      "tokens_seen": 2193162240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387763289869609,
      "loss": 1.2898,
      "theoretical_loss": 3.4328193942532934,
      "tokens_seen": 2193227776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033875626880641924,
      "loss": 1.3954,
      "theoretical_loss": 3.4328110657458977,
      "tokens_seen": 2193293312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387362086258776,
      "loss": 1.2491,
      "theoretical_loss": 3.432802737557033,
      "tokens_seen": 2193358848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033871614844533597,
      "loss": 1.3824,
      "theoretical_loss": 3.4327944096866783,
      "tokens_seen": 2193424384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386960882647944,
      "loss": 1.336,
      "theoretical_loss": 3.432786082134811,
      "tokens_seen": 2193489920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033867602808425275,
      "loss": 1.3324,
      "theoretical_loss": 3.43277775490141,
      "tokens_seen": 2193555456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386559679037111,
      "loss": 1.3167,
      "theoretical_loss": 3.432769427986454,
      "tokens_seen": 2193620992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033863590772316947,
      "loss": 1.3703,
      "theoretical_loss": 3.43276110138992,
      "tokens_seen": 2193686528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386158475426279,
      "loss": 1.4705,
      "theoretical_loss": 3.4327527751117874,
      "tokens_seen": 2193752064
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033859578736208625,
      "loss": 1.2581,
      "theoretical_loss": 3.4327444491520342,
      "tokens_seen": 2193817600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033857572718154467,
      "loss": 1.3588,
      "theoretical_loss": 3.432736123510639,
      "tokens_seen": 2193883136
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 798764,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3836020231246948,
      "objective/train/theoretical_loss": 3.4327277981875794,
      "objective/train/tokens_used": 564407776,
      "theoretical_loss": 3.4327277981875794,
      "tokens_seen": 2193948672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033855566700100303,
      "loss": 1.3783,
      "theoretical_loss": 3.4327277981875794,
      "tokens_seen": 2193948672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033853560682046144,
      "loss": 1.304,
      "theoretical_loss": 3.4327194731828348,
      "tokens_seen": 2194014208
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003385155466399198,
      "loss": 1.4275,
      "theoretical_loss": 3.432711148496382,
      "tokens_seen": 2194079744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033849548645937817,
      "loss": 1.328,
      "theoretical_loss": 3.4327028241282007,
      "tokens_seen": 2194145280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033847542627883653,
      "loss": 1.3396,
      "theoretical_loss": 3.4326945000782687,
      "tokens_seen": 2194210816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384553660982949,
      "loss": 1.2473,
      "theoretical_loss": 3.4326861763465644,
      "tokens_seen": 2194276352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384353059177533,
      "loss": 1.4009,
      "theoretical_loss": 3.432677852933066,
      "tokens_seen": 2194341888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033841524573721167,
      "loss": 1.2956,
      "theoretical_loss": 3.432669529837752,
      "tokens_seen": 2194407424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033839518555667003,
      "loss": 1.377,
      "theoretical_loss": 3.4326612070606006,
      "tokens_seen": 2194472960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383751253761284,
      "loss": 1.2978,
      "theoretical_loss": 3.4326528846015902,
      "tokens_seen": 2194538496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383550651955868,
      "loss": 1.3589,
      "theoretical_loss": 3.432644562460699,
      "tokens_seen": 2194604032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383350050150452,
      "loss": 1.2523,
      "theoretical_loss": 3.4326362406379056,
      "tokens_seen": 2194669568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033831494483450354,
      "loss": 1.3193,
      "theoretical_loss": 3.4326279191331883,
      "tokens_seen": 2194735104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382948846539619,
      "loss": 1.3393,
      "theoretical_loss": 3.4326195979465255,
      "tokens_seen": 2194800640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033827482447342026,
      "loss": 1.447,
      "theoretical_loss": 3.432611277077895,
      "tokens_seen": 2194866176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382547642928787,
      "loss": 1.3865,
      "theoretical_loss": 3.4326029565272758,
      "tokens_seen": 2194931712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033823470411233704,
      "loss": 1.347,
      "theoretical_loss": 3.432594636294646,
      "tokens_seen": 2194997248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382146439317954,
      "loss": 1.3352,
      "theoretical_loss": 3.4325863163799837,
      "tokens_seen": 2195062784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033819458375125376,
      "loss": 1.2813,
      "theoretical_loss": 3.4325779967832677,
      "tokens_seen": 2195128320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381745235707122,
      "loss": 1.3839,
      "theoretical_loss": 3.4325696775044765,
      "tokens_seen": 2195193856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033815446339017054,
      "loss": 1.2652,
      "theoretical_loss": 3.4325613585435875,
      "tokens_seen": 2195259392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381344032096289,
      "loss": 1.2375,
      "theoretical_loss": 3.43255303990058,
      "tokens_seen": 2195324928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033811434302908727,
      "loss": 1.3222,
      "theoretical_loss": 3.4325447215754323,
      "tokens_seen": 2195390464
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033809428284854563,
      "loss": 1.3751,
      "theoretical_loss": 3.4325364035681223,
      "tokens_seen": 2195456000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033807422266800405,
      "loss": 1.2995,
      "theoretical_loss": 3.4325280858786287,
      "tokens_seen": 2195521536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003380541624874624,
      "loss": 1.2695,
      "theoretical_loss": 3.4325197685069297,
      "tokens_seen": 2195587072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033803410230692077,
      "loss": 1.2899,
      "theoretical_loss": 3.4325114514530037,
      "tokens_seen": 2195652608
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033801404212637913,
      "loss": 1.3608,
      "theoretical_loss": 3.4325031347168293,
      "tokens_seen": 2195718144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033799398194583755,
      "loss": 1.4839,
      "theoretical_loss": 3.4324948182983843,
      "tokens_seen": 2195783680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003379739217652959,
      "loss": 1.2855,
      "theoretical_loss": 3.432486502197648,
      "tokens_seen": 2195849216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003379538615847543,
      "loss": 1.3818,
      "theoretical_loss": 3.4324781864145977,
      "tokens_seen": 2195914752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033793380140421264,
      "loss": 1.2523,
      "theoretical_loss": 3.4324698709492125,
      "tokens_seen": 2195980288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337913741223671,
      "loss": 1.3566,
      "theoretical_loss": 3.432461555801471,
      "tokens_seen": 2196045824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378936810431294,
      "loss": 1.3287,
      "theoretical_loss": 3.432453240971351,
      "tokens_seen": 2196111360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378736208625878,
      "loss": 1.2999,
      "theoretical_loss": 3.432444926458831,
      "tokens_seen": 2196176896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033785356068204614,
      "loss": 1.2603,
      "theoretical_loss": 3.4324366122638894,
      "tokens_seen": 2196242432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378335005015045,
      "loss": 1.3015,
      "theoretical_loss": 3.4324282983865047,
      "tokens_seen": 2196307968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378134403209629,
      "loss": 1.3631,
      "theoretical_loss": 3.4324199848266557,
      "tokens_seen": 2196373504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377933801404213,
      "loss": 1.2407,
      "theoretical_loss": 3.43241167158432,
      "tokens_seen": 2196439040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033777331995987964,
      "loss": 1.305,
      "theoretical_loss": 3.432403358659476,
      "tokens_seen": 2196504576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337753259779338,
      "loss": 1.4281,
      "theoretical_loss": 3.432395046052103,
      "tokens_seen": 2196570112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033773319959879637,
      "loss": 1.3704,
      "theoretical_loss": 3.432386733762179,
      "tokens_seen": 2196635648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377131394182548,
      "loss": 1.3947,
      "theoretical_loss": 3.432378421789682,
      "tokens_seen": 2196701184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033769307923771315,
      "loss": 1.3973,
      "theoretical_loss": 3.432370110134591,
      "tokens_seen": 2196766720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376730190571715,
      "loss": 1.2445,
      "theoretical_loss": 3.432361798796884,
      "tokens_seen": 2196832256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033765295887662987,
      "loss": 1.2544,
      "theoretical_loss": 3.4323534877765396,
      "tokens_seen": 2196897792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376328986960883,
      "loss": 1.3833,
      "theoretical_loss": 3.4323451770735356,
      "tokens_seen": 2196963328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033761283851554665,
      "loss": 1.3316,
      "theoretical_loss": 3.4323368666878515,
      "tokens_seen": 2197028864
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337592778335005,
      "loss": 1.3844,
      "theoretical_loss": 3.4323285566194652,
      "tokens_seen": 2197094400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375727181544634,
      "loss": 1.345,
      "theoretical_loss": 3.432320246868355,
      "tokens_seen": 2197159936
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 799965,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.29378080368042,
      "objective/train/theoretical_loss": 3.4323119374344992,
      "objective/train/tokens_used": 567684576,
      "theoretical_loss": 3.4323119374344992,
      "tokens_seen": 2197225472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033755265797392174,
      "loss": 1.3068,
      "theoretical_loss": 3.4323119374344992,
      "tokens_seen": 2197225472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033753259779338015,
      "loss": 1.2938,
      "theoretical_loss": 3.4323036283178765,
      "tokens_seen": 2197291008
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375125376128385,
      "loss": 1.3309,
      "theoretical_loss": 3.4322953195184653,
      "tokens_seen": 2197356544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374924774322969,
      "loss": 1.3439,
      "theoretical_loss": 3.432287011036244,
      "tokens_seen": 2197422080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033747241725175524,
      "loss": 1.3396,
      "theoretical_loss": 3.432278702871191,
      "tokens_seen": 2197487616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033745235707121366,
      "loss": 1.3109,
      "theoretical_loss": 3.432270395023285,
      "tokens_seen": 2197553152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337432296890672,
      "loss": 1.3985,
      "theoretical_loss": 3.432262087492504,
      "tokens_seen": 2197618688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374122367101304,
      "loss": 1.277,
      "theoretical_loss": 3.4322537802788267,
      "tokens_seen": 2197684224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033739217652958874,
      "loss": 1.3417,
      "theoretical_loss": 3.4322454733822316,
      "tokens_seen": 2197749760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373721163490471,
      "loss": 1.294,
      "theoretical_loss": 3.4322371668026967,
      "tokens_seen": 2197815296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373520561685055,
      "loss": 1.461,
      "theoretical_loss": 3.432228860540201,
      "tokens_seen": 2197880832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373319959879639,
      "loss": 1.3607,
      "theoretical_loss": 3.4322205545947226,
      "tokens_seen": 2197946368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033731193580742225,
      "loss": 1.3439,
      "theoretical_loss": 3.4322122489662403,
      "tokens_seen": 2198011904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372918756268806,
      "loss": 1.3775,
      "theoretical_loss": 3.432203943654732,
      "tokens_seen": 2198077440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337271815446339,
      "loss": 1.3674,
      "theoretical_loss": 3.432195638660177,
      "tokens_seen": 2198142976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372517552657974,
      "loss": 1.2804,
      "theoretical_loss": 3.4321873339825526,
      "tokens_seen": 2198208512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033723169508525575,
      "loss": 1.3526,
      "theoretical_loss": 3.432179029621838,
      "tokens_seen": 2198274048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372116349047141,
      "loss": 1.3247,
      "theoretical_loss": 3.432170725578012,
      "tokens_seen": 2198339584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371915747241725,
      "loss": 1.383,
      "theoretical_loss": 3.4321624218510522,
      "tokens_seen": 2198405120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371715145436309,
      "loss": 1.2943,
      "theoretical_loss": 3.4321541184409377,
      "tokens_seen": 2198470656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033715145436308925,
      "loss": 1.2926,
      "theoretical_loss": 3.4321458153476465,
      "tokens_seen": 2198536192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371313941825476,
      "loss": 1.3219,
      "theoretical_loss": 3.432137512571157,
      "tokens_seen": 2198601728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337111334002006,
      "loss": 1.3274,
      "theoretical_loss": 3.432129210111449,
      "tokens_seen": 2198667264
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370912738214644,
      "loss": 1.3279,
      "theoretical_loss": 3.432120907968499,
      "tokens_seen": 2198732800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370712136409228,
      "loss": 1.3657,
      "theoretical_loss": 3.4321126061422866,
      "tokens_seen": 2198798336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033705115346038117,
      "loss": 1.4111,
      "theoretical_loss": 3.43210430463279,
      "tokens_seen": 2198863872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033703109327983953,
      "loss": 1.1815,
      "theoretical_loss": 3.4320960034399883,
      "tokens_seen": 2198929408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033701103309929795,
      "loss": 1.1981,
      "theoretical_loss": 3.4320877025638588,
      "tokens_seen": 2198994944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369909729187563,
      "loss": 1.2987,
      "theoretical_loss": 3.432079402004381,
      "tokens_seen": 2199060480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369709127382147,
      "loss": 1.3709,
      "theoretical_loss": 3.432071101761533,
      "tokens_seen": 2199126016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033695085255767304,
      "loss": 1.3555,
      "theoretical_loss": 3.4320628018352934,
      "tokens_seen": 2199191552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369307923771314,
      "loss": 1.3236,
      "theoretical_loss": 3.4320545022256406,
      "tokens_seen": 2199257088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369107321965898,
      "loss": 1.2404,
      "theoretical_loss": 3.4320462029325527,
      "tokens_seen": 2199322624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368906720160482,
      "loss": 1.2502,
      "theoretical_loss": 3.432037903956009,
      "tokens_seen": 2199388160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033687061183550654,
      "loss": 1.3871,
      "theoretical_loss": 3.432029605295987,
      "tokens_seen": 2199453696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368505516549649,
      "loss": 1.3112,
      "theoretical_loss": 3.4320213069524663,
      "tokens_seen": 2199519232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368304914744233,
      "loss": 1.3443,
      "theoretical_loss": 3.432013008925425,
      "tokens_seen": 2199584768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368104312938817,
      "loss": 1.3577,
      "theoretical_loss": 3.432004711214841,
      "tokens_seen": 2199650304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033679037111334004,
      "loss": 1.3898,
      "theoretical_loss": 3.4319964138206935,
      "tokens_seen": 2199715840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367703109327984,
      "loss": 1.345,
      "theoretical_loss": 3.4319881167429607,
      "tokens_seen": 2199781376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033675025075225677,
      "loss": 1.3213,
      "theoretical_loss": 3.4319798199816214,
      "tokens_seen": 2199846912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367301905717152,
      "loss": 1.314,
      "theoretical_loss": 3.431971523536654,
      "tokens_seen": 2199912448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033671013039117355,
      "loss": 1.2828,
      "theoretical_loss": 3.431963227408036,
      "tokens_seen": 2199977984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366900702106319,
      "loss": 1.3471,
      "theoretical_loss": 3.431954931595748,
      "tokens_seen": 2200043520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033667001003009027,
      "loss": 1.3046,
      "theoretical_loss": 3.4319466360997666,
      "tokens_seen": 2200109056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366499498495487,
      "loss": 1.3941,
      "theoretical_loss": 3.431938340920071,
      "tokens_seen": 2200174592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033662988966900705,
      "loss": 1.3828,
      "theoretical_loss": 3.43193004605664,
      "tokens_seen": 2200240128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366098294884654,
      "loss": 1.308,
      "theoretical_loss": 3.431921751509452,
      "tokens_seen": 2200305664
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365897693079238,
      "loss": 1.3843,
      "theoretical_loss": 3.4319134572784855,
      "tokens_seen": 2200371200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033656970912738214,
      "loss": 1.3077,
      "theoretical_loss": 3.4319051633637185,
      "tokens_seen": 2200436736
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 801168,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3015751838684082,
      "objective/train/theoretical_loss": 3.4318968697651306,
      "objective/train/tokens_used": 570961376,
      "theoretical_loss": 3.4318968697651306,
      "tokens_seen": 2200502272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033654964894684055,
      "loss": 1.2766,
      "theoretical_loss": 3.4318968697651306,
      "tokens_seen": 2200502272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365295887662989,
      "loss": 1.32,
      "theoretical_loss": 3.431888576482699,
      "tokens_seen": 2200567808
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365095285857573,
      "loss": 1.3601,
      "theoretical_loss": 3.4318802835164037,
      "tokens_seen": 2200633344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033648946840521564,
      "loss": 1.1883,
      "theoretical_loss": 3.431871990866222,
      "tokens_seen": 2200698880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033646940822467406,
      "loss": 1.333,
      "theoretical_loss": 3.431863698532133,
      "tokens_seen": 2200764416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364493480441324,
      "loss": 1.2618,
      "theoretical_loss": 3.4318554065141154,
      "tokens_seen": 2200829952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364292878635908,
      "loss": 1.3489,
      "theoretical_loss": 3.4318471148121468,
      "tokens_seen": 2200895488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033640922768304914,
      "loss": 1.3973,
      "theoretical_loss": 3.431838823426207,
      "tokens_seen": 2200961024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363891675025075,
      "loss": 1.2477,
      "theoretical_loss": 3.431830532356274,
      "tokens_seen": 2201026560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363691073219659,
      "loss": 1.2898,
      "theoretical_loss": 3.431822241602326,
      "tokens_seen": 2201092096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363490471414243,
      "loss": 1.3257,
      "theoretical_loss": 3.431813951164342,
      "tokens_seen": 2201157632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033632898696088265,
      "loss": 1.4136,
      "theoretical_loss": 3.4318056610423007,
      "tokens_seen": 2201223168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336308926780341,
      "loss": 1.2406,
      "theoretical_loss": 3.43179737123618,
      "tokens_seen": 2201288704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362888665997994,
      "loss": 1.3525,
      "theoretical_loss": 3.431789081745959,
      "tokens_seen": 2201354240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362688064192578,
      "loss": 1.2073,
      "theoretical_loss": 3.431780792571616,
      "tokens_seen": 2201419776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033624874623871615,
      "loss": 1.3003,
      "theoretical_loss": 3.43177250371313,
      "tokens_seen": 2201485312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362286860581745,
      "loss": 1.3523,
      "theoretical_loss": 3.431764215170479,
      "tokens_seen": 2201550848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033620862587763293,
      "loss": 1.3401,
      "theoretical_loss": 3.4317559269436417,
      "tokens_seen": 2201616384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361885656970913,
      "loss": 1.2454,
      "theoretical_loss": 3.431747639032597,
      "tokens_seen": 2201681920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033616850551654965,
      "loss": 1.2756,
      "theoretical_loss": 3.431739351437323,
      "tokens_seen": 2201747456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336148445336008,
      "loss": 1.3579,
      "theoretical_loss": 3.4317310641577987,
      "tokens_seen": 2201812992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361283851554664,
      "loss": 1.2667,
      "theoretical_loss": 3.431722777194002,
      "tokens_seen": 2201878528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361083249749248,
      "loss": 1.2843,
      "theoretical_loss": 3.4317144905459127,
      "tokens_seen": 2201944064
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033608826479438316,
      "loss": 1.3839,
      "theoretical_loss": 3.4317062042135076,
      "tokens_seen": 2202009600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003360682046138415,
      "loss": 1.337,
      "theoretical_loss": 3.4316979181967673,
      "tokens_seen": 2202075136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003360481444332999,
      "loss": 1.3916,
      "theoretical_loss": 3.4316896324956687,
      "tokens_seen": 2202140672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003360280842527583,
      "loss": 1.2635,
      "theoretical_loss": 3.4316813471101915,
      "tokens_seen": 2202206208
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033600802407221666,
      "loss": 1.3362,
      "theoretical_loss": 3.4316730620403133,
      "tokens_seen": 2202271744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000335987963891675,
      "loss": 1.3782,
      "theoretical_loss": 3.431664777286014,
      "tokens_seen": 2202337280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003359679037111334,
      "loss": 1.3283,
      "theoretical_loss": 3.431656492847271,
      "tokens_seen": 2202402816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033594784353059174,
      "loss": 1.3423,
      "theoretical_loss": 3.431648208724063,
      "tokens_seen": 2202468352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033592778335005016,
      "loss": 1.296,
      "theoretical_loss": 3.4316399249163694,
      "tokens_seen": 2202533888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003359077231695085,
      "loss": 1.2728,
      "theoretical_loss": 3.431631641424168,
      "tokens_seen": 2202599424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003358876629889669,
      "loss": 1.3436,
      "theoretical_loss": 3.4316233582474376,
      "tokens_seen": 2202664960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033586760280842525,
      "loss": 1.2932,
      "theoretical_loss": 3.431615075386157,
      "tokens_seen": 2202730496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033584754262788366,
      "loss": 1.4493,
      "theoretical_loss": 3.431606792840305,
      "tokens_seen": 2202796032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000335827482447342,
      "loss": 1.5376,
      "theoretical_loss": 3.4315985106098594,
      "tokens_seen": 2202861568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358074222668004,
      "loss": 1.35,
      "theoretical_loss": 3.4315902286948,
      "tokens_seen": 2202927104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033578736208625875,
      "loss": 1.3767,
      "theoretical_loss": 3.431581947095104,
      "tokens_seen": 2202992640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357673019057171,
      "loss": 1.3946,
      "theoretical_loss": 3.431573665810751,
      "tokens_seen": 2203058176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033574724172517553,
      "loss": 1.2454,
      "theoretical_loss": 3.43156538484172,
      "tokens_seen": 2203123712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357271815446339,
      "loss": 1.3536,
      "theoretical_loss": 3.4315571041879878,
      "tokens_seen": 2203189248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033570712136409225,
      "loss": 1.4043,
      "theoretical_loss": 3.431548823849535,
      "tokens_seen": 2203254784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356870611835506,
      "loss": 1.3067,
      "theoretical_loss": 3.431540543826339,
      "tokens_seen": 2203320320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033566700100300903,
      "loss": 1.3333,
      "theoretical_loss": 3.431532264118379,
      "tokens_seen": 2203385856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356469408224674,
      "loss": 1.3819,
      "theoretical_loss": 3.4315239847256334,
      "tokens_seen": 2203451392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033562688064192576,
      "loss": 1.2752,
      "theoretical_loss": 3.4315157056480805,
      "tokens_seen": 2203516928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356068204613841,
      "loss": 1.3737,
      "theoretical_loss": 3.4315074268856995,
      "tokens_seen": 2203582464
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355867602808425,
      "loss": 1.2647,
      "theoretical_loss": 3.431499148438469,
      "tokens_seen": 2203648000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033556670010030095,
      "loss": 1.3371,
      "theoretical_loss": 3.4314908703063676,
      "tokens_seen": 2203713536
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 802328,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3683021068572998,
      "objective/train/theoretical_loss": 3.4314825924893735,
      "objective/train/tokens_used": 574238176,
      "theoretical_loss": 3.4314825924893735,
      "tokens_seen": 2203779072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355466399197593,
      "loss": 1.3298,
      "theoretical_loss": 3.4314825924893735,
      "tokens_seen": 2203779072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355265797392177,
      "loss": 1.3622,
      "theoretical_loss": 3.4314743149874656,
      "tokens_seen": 2203844608
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033550651955867604,
      "loss": 1.3813,
      "theoretical_loss": 3.4314660378006225,
      "tokens_seen": 2203910144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033548645937813446,
      "loss": 1.3239,
      "theoretical_loss": 3.431457760928823,
      "tokens_seen": 2203975680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354663991975928,
      "loss": 1.2864,
      "theoretical_loss": 3.4314494843720458,
      "tokens_seen": 2204041216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354463390170512,
      "loss": 1.352,
      "theoretical_loss": 3.431441208130269,
      "tokens_seen": 2204106752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033542627883650954,
      "loss": 1.1718,
      "theoretical_loss": 3.431432932203472,
      "tokens_seen": 2204172288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354062186559679,
      "loss": 1.3385,
      "theoretical_loss": 3.4314246565916324,
      "tokens_seen": 2204237824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353861584754263,
      "loss": 1.294,
      "theoretical_loss": 3.4314163812947305,
      "tokens_seen": 2204303360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353660982948847,
      "loss": 1.4892,
      "theoretical_loss": 3.4314081063127433,
      "tokens_seen": 2204368896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033534603811434305,
      "loss": 1.3164,
      "theoretical_loss": 3.4313998316456504,
      "tokens_seen": 2204434432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353259779338014,
      "loss": 1.3477,
      "theoretical_loss": 3.4313915572934297,
      "tokens_seen": 2204499968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353059177532598,
      "loss": 1.398,
      "theoretical_loss": 3.431383283256061,
      "tokens_seen": 2204565504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352858575727182,
      "loss": 1.2768,
      "theoretical_loss": 3.431375009533522,
      "tokens_seen": 2204631040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033526579739217655,
      "loss": 1.3906,
      "theoretical_loss": 3.4313667361257916,
      "tokens_seen": 2204696576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352457372116349,
      "loss": 1.3745,
      "theoretical_loss": 3.4313584630328484,
      "tokens_seen": 2204762112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033522567703109333,
      "loss": 1.2807,
      "theoretical_loss": 3.4313501902546717,
      "tokens_seen": 2204827648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352056168505517,
      "loss": 1.3377,
      "theoretical_loss": 3.431341917791239,
      "tokens_seen": 2204893184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033518555667001005,
      "loss": 1.3502,
      "theoretical_loss": 3.43133364564253,
      "tokens_seen": 2204958720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351654964894684,
      "loss": 1.238,
      "theoretical_loss": 3.431325373808523,
      "tokens_seen": 2205024256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351454363089268,
      "loss": 1.3744,
      "theoretical_loss": 3.4313171022891966,
      "tokens_seen": 2205089792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351253761283852,
      "loss": 1.3221,
      "theoretical_loss": 3.4313088310845297,
      "tokens_seen": 2205155328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033510531594784356,
      "loss": 1.3182,
      "theoretical_loss": 3.431300560194501,
      "tokens_seen": 2205220864
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350852557673019,
      "loss": 1.3753,
      "theoretical_loss": 3.4312922896190887,
      "tokens_seen": 2205286400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350651955867603,
      "loss": 1.5038,
      "theoretical_loss": 3.4312840193582717,
      "tokens_seen": 2205351936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350451354062187,
      "loss": 1.3746,
      "theoretical_loss": 3.431275749412029,
      "tokens_seen": 2205417472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033502507522567706,
      "loss": 1.3086,
      "theoretical_loss": 3.431267479780339,
      "tokens_seen": 2205483008
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350050150451354,
      "loss": 1.4448,
      "theoretical_loss": 3.4312592104631805,
      "tokens_seen": 2205548544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003349849548645938,
      "loss": 1.2769,
      "theoretical_loss": 3.431250941460532,
      "tokens_seen": 2205614080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033496489468405215,
      "loss": 1.3541,
      "theoretical_loss": 3.431242672772373,
      "tokens_seen": 2205679616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033494483450351056,
      "loss": 1.3862,
      "theoretical_loss": 3.4312344043986807,
      "tokens_seen": 2205745152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003349247743229689,
      "loss": 1.365,
      "theoretical_loss": 3.431226136339435,
      "tokens_seen": 2205810688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003349047141424273,
      "loss": 1.3345,
      "theoretical_loss": 3.431217868594614,
      "tokens_seen": 2205876224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033488465396188565,
      "loss": 1.391,
      "theoretical_loss": 3.431209601164197,
      "tokens_seen": 2205941760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033486459378134406,
      "loss": 1.2849,
      "theoretical_loss": 3.431201334048162,
      "tokens_seen": 2206007296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003348445336008024,
      "loss": 1.352,
      "theoretical_loss": 3.4311930672464888,
      "tokens_seen": 2206072832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003348244734202608,
      "loss": 1.2794,
      "theoretical_loss": 3.4311848007591546,
      "tokens_seen": 2206138368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033480441323971915,
      "loss": 1.2393,
      "theoretical_loss": 3.431176534586139,
      "tokens_seen": 2206203904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347843530591775,
      "loss": 1.4392,
      "theoretical_loss": 3.4311682687274208,
      "tokens_seen": 2206269440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033476429287863593,
      "loss": 1.2982,
      "theoretical_loss": 3.431160003182978,
      "tokens_seen": 2206334976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347442326980943,
      "loss": 1.2828,
      "theoretical_loss": 3.4311517379527903,
      "tokens_seen": 2206400512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033472417251755265,
      "loss": 1.3521,
      "theoretical_loss": 3.431143473036836,
      "tokens_seen": 2206466048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334704112337011,
      "loss": 1.2801,
      "theoretical_loss": 3.4311352084350935,
      "tokens_seen": 2206531584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033468405215646943,
      "loss": 1.2954,
      "theoretical_loss": 3.431126944147542,
      "tokens_seen": 2206597120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346639919759278,
      "loss": 1.3527,
      "theoretical_loss": 3.4311186801741598,
      "tokens_seen": 2206662656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033464393179538616,
      "loss": 1.3429,
      "theoretical_loss": 3.4311104165149255,
      "tokens_seen": 2206728192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346238716148445,
      "loss": 1.3928,
      "theoretical_loss": 3.4311021531698187,
      "tokens_seen": 2206793728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346038114343029,
      "loss": 1.2796,
      "theoretical_loss": 3.4310938901388175,
      "tokens_seen": 2206859264
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345837512537613,
      "loss": 1.3227,
      "theoretical_loss": 3.4310856274219006,
      "tokens_seen": 2206924800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033456369107321966,
      "loss": 1.4448,
      "theoretical_loss": 3.4310773650190467,
      "tokens_seen": 2206990336
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 803600,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3918547630310059,
      "objective/train/theoretical_loss": 3.431069102930235,
      "objective/train/tokens_used": 577514976,
      "theoretical_loss": 3.431069102930235,
      "tokens_seen": 2207055872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334543630892678,
      "loss": 1.3267,
      "theoretical_loss": 3.431069102930235,
      "tokens_seen": 2207055872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345235707121364,
      "loss": 1.308,
      "theoretical_loss": 3.4310608411554435,
      "tokens_seen": 2207121408
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345035105315948,
      "loss": 1.341,
      "theoretical_loss": 3.4310525796946516,
      "tokens_seen": 2207186944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033448345035105316,
      "loss": 1.3016,
      "theoretical_loss": 3.431044318547838,
      "tokens_seen": 2207252480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344633901705115,
      "loss": 1.2061,
      "theoretical_loss": 3.431036057714981,
      "tokens_seen": 2207318016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344433299899699,
      "loss": 1.3517,
      "theoretical_loss": 3.4310277971960597,
      "tokens_seen": 2207383552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033442326980942825,
      "loss": 1.3882,
      "theoretical_loss": 3.4310195369910526,
      "tokens_seen": 2207449088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033440320962888667,
      "loss": 1.3396,
      "theoretical_loss": 3.431011277099939,
      "tokens_seen": 2207514624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033438314944834503,
      "loss": 1.3227,
      "theoretical_loss": 3.431003017522697,
      "tokens_seen": 2207580160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003343630892678034,
      "loss": 1.3462,
      "theoretical_loss": 3.4309947582593057,
      "tokens_seen": 2207645696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033434302908726175,
      "loss": 1.3299,
      "theoretical_loss": 3.4309864993097436,
      "tokens_seen": 2207711232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033432296890672017,
      "loss": 1.3047,
      "theoretical_loss": 3.4309782406739897,
      "tokens_seen": 2207776768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033430290872617853,
      "loss": 1.3421,
      "theoretical_loss": 3.430969982352023,
      "tokens_seen": 2207842304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342828485456369,
      "loss": 1.3673,
      "theoretical_loss": 3.4309617243438217,
      "tokens_seen": 2207907840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033426278836509526,
      "loss": 1.2267,
      "theoretical_loss": 3.4309534666493646,
      "tokens_seen": 2207973376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342427281845536,
      "loss": 1.3883,
      "theoretical_loss": 3.4309452092686312,
      "tokens_seen": 2208038912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033422266800401204,
      "loss": 1.4103,
      "theoretical_loss": 3.4309369522015993,
      "tokens_seen": 2208104448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342026078234704,
      "loss": 1.3459,
      "theoretical_loss": 3.4309286954482485,
      "tokens_seen": 2208169984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033418254764292876,
      "loss": 1.2135,
      "theoretical_loss": 3.430920439008557,
      "tokens_seen": 2208235520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341624874623871,
      "loss": 1.3554,
      "theoretical_loss": 3.430912182882504,
      "tokens_seen": 2208301056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033414242728184554,
      "loss": 1.3319,
      "theoretical_loss": 3.4309039270700676,
      "tokens_seen": 2208366592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341223671013039,
      "loss": 1.2252,
      "theoretical_loss": 3.4308956715712275,
      "tokens_seen": 2208432128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033410230692076226,
      "loss": 1.3462,
      "theoretical_loss": 3.430887416385962,
      "tokens_seen": 2208497664
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340822467402207,
      "loss": 1.3496,
      "theoretical_loss": 3.4308791615142495,
      "tokens_seen": 2208563200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340621865596791,
      "loss": 1.3064,
      "theoretical_loss": 3.43087090695607,
      "tokens_seen": 2208628736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033404212637913746,
      "loss": 1.3807,
      "theoretical_loss": 3.430862652711401,
      "tokens_seen": 2208694272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340220661985958,
      "loss": 1.3175,
      "theoretical_loss": 3.4308543987802214,
      "tokens_seen": 2208759808
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340020060180542,
      "loss": 1.3441,
      "theoretical_loss": 3.4308461451625107,
      "tokens_seen": 2208825344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033398194583751255,
      "loss": 1.3901,
      "theoretical_loss": 3.430837891858247,
      "tokens_seen": 2208890880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033396188565697096,
      "loss": 1.4168,
      "theoretical_loss": 3.43082963886741,
      "tokens_seen": 2208956416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003339418254764293,
      "loss": 1.3062,
      "theoretical_loss": 3.430821386189978,
      "tokens_seen": 2209021952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003339217652958877,
      "loss": 1.3759,
      "theoretical_loss": 3.4308131338259296,
      "tokens_seen": 2209087488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033390170511534605,
      "loss": 1.2549,
      "theoretical_loss": 3.4308048817752437,
      "tokens_seen": 2209153024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033388164493480447,
      "loss": 1.4478,
      "theoretical_loss": 3.430796630037899,
      "tokens_seen": 2209218560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033386158475426283,
      "loss": 1.4023,
      "theoretical_loss": 3.430788378613875,
      "tokens_seen": 2209284096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338415245737212,
      "loss": 1.4027,
      "theoretical_loss": 3.4307801275031498,
      "tokens_seen": 2209349632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033382146439317955,
      "loss": 1.3824,
      "theoretical_loss": 3.4307718767057023,
      "tokens_seen": 2209415168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338014042126379,
      "loss": 1.4902,
      "theoretical_loss": 3.4307636262215113,
      "tokens_seen": 2209480704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033378134403209633,
      "loss": 1.2937,
      "theoretical_loss": 3.4307553760505556,
      "tokens_seen": 2209546240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337612838515547,
      "loss": 1.3211,
      "theoretical_loss": 3.4307471261928146,
      "tokens_seen": 2209611776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033374122367101306,
      "loss": 1.2264,
      "theoretical_loss": 3.4307388766482667,
      "tokens_seen": 2209677312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337211634904714,
      "loss": 1.3344,
      "theoretical_loss": 3.43073062741689,
      "tokens_seen": 2209742848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033370110330992983,
      "loss": 1.3912,
      "theoretical_loss": 3.4307223784986647,
      "tokens_seen": 2209808384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336810431293882,
      "loss": 1.2656,
      "theoretical_loss": 3.430714129893569,
      "tokens_seen": 2209873920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033366098294884656,
      "loss": 1.3834,
      "theoretical_loss": 3.430705881601581,
      "tokens_seen": 2209939456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336409227683049,
      "loss": 1.2637,
      "theoretical_loss": 3.4306976336226804,
      "tokens_seen": 2210004992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336208625877633,
      "loss": 1.3339,
      "theoretical_loss": 3.430689385956846,
      "tokens_seen": 2210070528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336008024072217,
      "loss": 1.2492,
      "theoretical_loss": 3.4306811386040565,
      "tokens_seen": 2210136064
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033358074222668006,
      "loss": 1.3958,
      "theoretical_loss": 3.430672891564291,
      "tokens_seen": 2210201600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335606820461384,
      "loss": 1.3039,
      "theoretical_loss": 3.4306646448375275,
      "tokens_seen": 2210267136
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 804834,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.344569444656372,
      "objective/train/theoretical_loss": 3.4306563984237455,
      "objective/train/tokens_used": 580791776,
      "theoretical_loss": 3.4306563984237455,
      "tokens_seen": 2210332672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335406218655968,
      "loss": 1.2756,
      "theoretical_loss": 3.4306563984237455,
      "tokens_seen": 2210332672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335205616850552,
      "loss": 1.3968,
      "theoretical_loss": 3.4306481523229237,
      "tokens_seen": 2210398208
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033350050150451356,
      "loss": 1.2591,
      "theoretical_loss": 3.430639906535041,
      "tokens_seen": 2210463744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003334804413239719,
      "loss": 1.3303,
      "theoretical_loss": 3.4306316610600764,
      "tokens_seen": 2210529280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003334603811434303,
      "loss": 1.3544,
      "theoretical_loss": 3.4306234158980082,
      "tokens_seen": 2210594816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033344032096288865,
      "loss": 1.3693,
      "theoretical_loss": 3.430615171048816,
      "tokens_seen": 2210660352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033342026078234707,
      "loss": 1.2581,
      "theoretical_loss": 3.4306069265124783,
      "tokens_seen": 2210725888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033340020060180543,
      "loss": 1.2585,
      "theoretical_loss": 3.4305986822889736,
      "tokens_seen": 2210791424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003333801404212638,
      "loss": 1.3018,
      "theoretical_loss": 3.4305904383782813,
      "tokens_seen": 2210856960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033336008024072215,
      "loss": 1.3816,
      "theoretical_loss": 3.4305821947803796,
      "tokens_seen": 2210922496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033334002006018057,
      "loss": 1.2539,
      "theoretical_loss": 3.4305739514952482,
      "tokens_seen": 2210988032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033331995987963893,
      "loss": 1.3827,
      "theoretical_loss": 3.430565708522866,
      "tokens_seen": 2211053568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332998996990973,
      "loss": 1.4157,
      "theoretical_loss": 3.4305574658632105,
      "tokens_seen": 2211119104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033327983951855566,
      "loss": 1.2953,
      "theoretical_loss": 3.4305492235162616,
      "tokens_seen": 2211184640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000333259779338014,
      "loss": 1.4728,
      "theoretical_loss": 3.4305409814819985,
      "tokens_seen": 2211250176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033323971915747244,
      "loss": 1.2951,
      "theoretical_loss": 3.4305327397603995,
      "tokens_seen": 2211315712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332196589769308,
      "loss": 1.2814,
      "theoretical_loss": 3.430524498351444,
      "tokens_seen": 2211381248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033319959879638916,
      "loss": 1.4162,
      "theoretical_loss": 3.4305162572551096,
      "tokens_seen": 2211446784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331795386158475,
      "loss": 1.3947,
      "theoretical_loss": 3.430508016471377,
      "tokens_seen": 2211512320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033315947843530594,
      "loss": 1.2438,
      "theoretical_loss": 3.4304997760002234,
      "tokens_seen": 2211577856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331394182547643,
      "loss": 1.343,
      "theoretical_loss": 3.430491535841629,
      "tokens_seen": 2211643392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033311935807422266,
      "loss": 1.3006,
      "theoretical_loss": 3.4304832959955713,
      "tokens_seen": 2211708928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000333099297893681,
      "loss": 1.4471,
      "theoretical_loss": 3.4304750564620305,
      "tokens_seen": 2211774464
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033307923771313944,
      "loss": 1.4741,
      "theoretical_loss": 3.4304668172409847,
      "tokens_seen": 2211840000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003330591775325978,
      "loss": 1.4021,
      "theoretical_loss": 3.4304585783324133,
      "tokens_seen": 2211905536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033303911735205617,
      "loss": 1.3122,
      "theoretical_loss": 3.430450339736295,
      "tokens_seen": 2211971072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033301905717151453,
      "loss": 1.3565,
      "theoretical_loss": 3.4304421014526083,
      "tokens_seen": 2212036608
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003329989969909729,
      "loss": 1.3251,
      "theoretical_loss": 3.430433863481333,
      "tokens_seen": 2212102144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003329789368104313,
      "loss": 1.2868,
      "theoretical_loss": 3.430425625822447,
      "tokens_seen": 2212167680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033295887662988967,
      "loss": 1.2844,
      "theoretical_loss": 3.43041738847593,
      "tokens_seen": 2212233216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033293881644934803,
      "loss": 1.3552,
      "theoretical_loss": 3.4304091514417605,
      "tokens_seen": 2212298752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003329187562688064,
      "loss": 1.2922,
      "theoretical_loss": 3.4304009147199173,
      "tokens_seen": 2212364288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003328986960882648,
      "loss": 1.3541,
      "theoretical_loss": 3.430392678310379,
      "tokens_seen": 2212429824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003328786359077232,
      "loss": 1.3362,
      "theoretical_loss": 3.4303844422131258,
      "tokens_seen": 2212495360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033285857572718154,
      "loss": 1.3668,
      "theoretical_loss": 3.430376206428135,
      "tokens_seen": 2212560896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003328385155466399,
      "loss": 1.3328,
      "theoretical_loss": 3.4303679709553867,
      "tokens_seen": 2212626432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033281845536609826,
      "loss": 1.4272,
      "theoretical_loss": 3.4303597357948594,
      "tokens_seen": 2212691968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327983951855567,
      "loss": 1.333,
      "theoretical_loss": 3.430351500946532,
      "tokens_seen": 2212757504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033277833500501504,
      "loss": 1.3385,
      "theoretical_loss": 3.4303432664103832,
      "tokens_seen": 2212823040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327582748244734,
      "loss": 1.4265,
      "theoretical_loss": 3.4303350321863926,
      "tokens_seen": 2212888576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033273821464393176,
      "loss": 1.3897,
      "theoretical_loss": 3.4303267982745385,
      "tokens_seen": 2212954112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327181544633902,
      "loss": 1.2506,
      "theoretical_loss": 3.4303185646747996,
      "tokens_seen": 2213019648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033269809428284854,
      "loss": 1.3113,
      "theoretical_loss": 3.4303103313871555,
      "tokens_seen": 2213085184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003326780341023069,
      "loss": 1.3464,
      "theoretical_loss": 3.430302098411585,
      "tokens_seen": 2213150720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033265797392176527,
      "loss": 1.3402,
      "theoretical_loss": 3.430293865748067,
      "tokens_seen": 2213216256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033263791374122363,
      "loss": 1.351,
      "theoretical_loss": 3.43028563339658,
      "tokens_seen": 2213281792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033261785356068205,
      "loss": 1.271,
      "theoretical_loss": 3.4302774013571034,
      "tokens_seen": 2213347328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325977933801404,
      "loss": 1.3101,
      "theoretical_loss": 3.430269169629616,
      "tokens_seen": 2213412864
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325777331995988,
      "loss": 1.2616,
      "theoretical_loss": 3.4302609382140967,
      "tokens_seen": 2213478400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325576730190572,
      "loss": 1.3884,
      "theoretical_loss": 3.430252707110524,
      "tokens_seen": 2213543936
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 805966,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3321585655212402,
      "objective/train/theoretical_loss": 3.430244476318878,
      "objective/train/tokens_used": 584068576,
      "theoretical_loss": 3.430244476318878,
      "tokens_seen": 2213609472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325376128385156,
      "loss": 1.3494,
      "theoretical_loss": 3.430244476318878,
      "tokens_seen": 2213609472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033251755265797396,
      "loss": 1.3688,
      "theoretical_loss": 3.4302362458391364,
      "tokens_seen": 2213675008
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033249749247743233,
      "loss": 1.3037,
      "theoretical_loss": 3.4302280156712794,
      "tokens_seen": 2213740544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003324774322968907,
      "loss": 1.4126,
      "theoretical_loss": 3.430219785815285,
      "tokens_seen": 2213806080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033245737211634905,
      "loss": 1.3166,
      "theoretical_loss": 3.430211556271132,
      "tokens_seen": 2213871616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033243731193580747,
      "loss": 1.3076,
      "theoretical_loss": 3.4302033270388,
      "tokens_seen": 2213937152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033241725175526583,
      "loss": 1.3585,
      "theoretical_loss": 3.4301950981182676,
      "tokens_seen": 2214002688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323971915747242,
      "loss": 1.2564,
      "theoretical_loss": 3.430186869509514,
      "tokens_seen": 2214068224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033237713139418255,
      "loss": 1.3511,
      "theoretical_loss": 3.4301786412125184,
      "tokens_seen": 2214133760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033235707121364097,
      "loss": 1.3087,
      "theoretical_loss": 3.4301704132272586,
      "tokens_seen": 2214199296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033233701103309933,
      "loss": 1.3705,
      "theoretical_loss": 3.4301621855537148,
      "tokens_seen": 2214264832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323169508525577,
      "loss": 1.351,
      "theoretical_loss": 3.4301539581918656,
      "tokens_seen": 2214330368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033229689067201606,
      "loss": 1.288,
      "theoretical_loss": 3.43014573114169,
      "tokens_seen": 2214395904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322768304914744,
      "loss": 1.3084,
      "theoretical_loss": 3.4301375044031666,
      "tokens_seen": 2214461440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033225677031093284,
      "loss": 1.193,
      "theoretical_loss": 3.4301292779762744,
      "tokens_seen": 2214526976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322367101303912,
      "loss": 1.3568,
      "theoretical_loss": 3.430121051860993,
      "tokens_seen": 2214592512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033221664994984956,
      "loss": 1.3756,
      "theoretical_loss": 3.4301128260573006,
      "tokens_seen": 2214658048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321965897693079,
      "loss": 1.4211,
      "theoretical_loss": 3.4301046005651767,
      "tokens_seen": 2214723584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033217652958876634,
      "loss": 1.3621,
      "theoretical_loss": 3.4300963753846005,
      "tokens_seen": 2214789120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321564694082247,
      "loss": 1.2664,
      "theoretical_loss": 3.4300881505155503,
      "tokens_seen": 2214854656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033213640922768306,
      "loss": 1.2217,
      "theoretical_loss": 3.4300799259580055,
      "tokens_seen": 2214920192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321163490471414,
      "loss": 1.3188,
      "theoretical_loss": 3.4300717017119444,
      "tokens_seen": 2214985728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033209628886659984,
      "loss": 1.2905,
      "theoretical_loss": 3.430063477777347,
      "tokens_seen": 2215051264
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003320762286860582,
      "loss": 1.3232,
      "theoretical_loss": 3.430055254154192,
      "tokens_seen": 2215116800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033205616850551657,
      "loss": 1.2838,
      "theoretical_loss": 3.4300470308424584,
      "tokens_seen": 2215182336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033203610832497493,
      "loss": 1.1825,
      "theoretical_loss": 3.430038807842125,
      "tokens_seen": 2215247872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003320160481444333,
      "loss": 1.3243,
      "theoretical_loss": 3.4300305851531703,
      "tokens_seen": 2215313408
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319959879638917,
      "loss": 1.3269,
      "theoretical_loss": 3.4300223627755746,
      "tokens_seen": 2215378944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033197592778335007,
      "loss": 1.2704,
      "theoretical_loss": 3.4300141407093157,
      "tokens_seen": 2215444480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033195586760280843,
      "loss": 1.3595,
      "theoretical_loss": 3.430005918954373,
      "tokens_seen": 2215510016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319358074222668,
      "loss": 1.3242,
      "theoretical_loss": 3.429997697510726,
      "tokens_seen": 2215575552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319157472417252,
      "loss": 1.2508,
      "theoretical_loss": 3.4299894763783527,
      "tokens_seen": 2215641088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318956870611836,
      "loss": 1.3513,
      "theoretical_loss": 3.429981255557233,
      "tokens_seen": 2215706624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033187562688064194,
      "loss": 1.4151,
      "theoretical_loss": 3.429973035047345,
      "tokens_seen": 2215772160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318555667001003,
      "loss": 1.4775,
      "theoretical_loss": 3.4299648148486686,
      "tokens_seen": 2215837696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033183550651955866,
      "loss": 1.3503,
      "theoretical_loss": 3.4299565949611828,
      "tokens_seen": 2215903232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318154463390171,
      "loss": 1.4167,
      "theoretical_loss": 3.429948375384866,
      "tokens_seen": 2215968768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033179538615847544,
      "loss": 1.1877,
      "theoretical_loss": 3.429940156119698,
      "tokens_seen": 2216034304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317753259779338,
      "loss": 1.3089,
      "theoretical_loss": 3.4299319371656565,
      "tokens_seen": 2216099840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033175526579739216,
      "loss": 1.3692,
      "theoretical_loss": 3.429923718522722,
      "tokens_seen": 2216165376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317352056168506,
      "loss": 1.3783,
      "theoretical_loss": 3.429915500190873,
      "tokens_seen": 2216230912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033171514543630894,
      "loss": 1.3537,
      "theoretical_loss": 3.429907282170088,
      "tokens_seen": 2216296448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003316950852557673,
      "loss": 1.3731,
      "theoretical_loss": 3.4298990644603466,
      "tokens_seen": 2216361984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033167502507522567,
      "loss": 1.2365,
      "theoretical_loss": 3.429890847061628,
      "tokens_seen": 2216427520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033165496489468403,
      "loss": 1.3151,
      "theoretical_loss": 3.4298826299739105,
      "tokens_seen": 2216493056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033163490471414245,
      "loss": 1.2895,
      "theoretical_loss": 3.4298744131971737,
      "tokens_seen": 2216558592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003316148445336008,
      "loss": 1.3609,
      "theoretical_loss": 3.429866196731396,
      "tokens_seen": 2216624128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033159478435305917,
      "loss": 1.2262,
      "theoretical_loss": 3.429857980576558,
      "tokens_seen": 2216689664
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033157472417251753,
      "loss": 1.3608,
      "theoretical_loss": 3.429849764732637,
      "tokens_seen": 2216755200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033155466399197595,
      "loss": 1.2636,
      "theoretical_loss": 3.4298415491996126,
      "tokens_seen": 2216820736
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 807026,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.281714677810669,
      "objective/train/theoretical_loss": 3.4298333339774643,
      "objective/train/tokens_used": 587345376,
      "theoretical_loss": 3.4298333339774643,
      "tokens_seen": 2216886272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315346038114343,
      "loss": 1.3427,
      "theoretical_loss": 3.4298333339774643,
      "tokens_seen": 2216886272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315145436308927,
      "loss": 1.2907,
      "theoretical_loss": 3.4298251190661704,
      "tokens_seen": 2216951808
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033149448345035104,
      "loss": 1.2677,
      "theoretical_loss": 3.429816904465711,
      "tokens_seen": 2217017344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314744232698094,
      "loss": 1.4184,
      "theoretical_loss": 3.429808690176064,
      "tokens_seen": 2217082880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314543630892678,
      "loss": 1.2704,
      "theoretical_loss": 3.429800476197209,
      "tokens_seen": 2217148416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314343029087262,
      "loss": 1.2691,
      "theoretical_loss": 3.429792262529125,
      "tokens_seen": 2217213952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033141424272818454,
      "loss": 1.1882,
      "theoretical_loss": 3.4297840491717912,
      "tokens_seen": 2217279488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313941825476429,
      "loss": 1.3455,
      "theoretical_loss": 3.429775836125186,
      "tokens_seen": 2217345024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313741223671013,
      "loss": 1.4167,
      "theoretical_loss": 3.42976762338929,
      "tokens_seen": 2217410560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313540621865597,
      "loss": 1.355,
      "theoretical_loss": 3.4297594109640803,
      "tokens_seen": 2217476096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033133400200601804,
      "loss": 1.3368,
      "theoretical_loss": 3.4297511988495377,
      "tokens_seen": 2217541632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313139418254764,
      "loss": 1.2772,
      "theoretical_loss": 3.42974298704564,
      "tokens_seen": 2217607168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033129388164493477,
      "loss": 1.2859,
      "theoretical_loss": 3.429734775552367,
      "tokens_seen": 2217672704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312738214643932,
      "loss": 1.3015,
      "theoretical_loss": 3.4297265643696973,
      "tokens_seen": 2217738240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033125376128385155,
      "loss": 1.3132,
      "theoretical_loss": 3.42971835349761,
      "tokens_seen": 2217803776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312337011033099,
      "loss": 1.1991,
      "theoretical_loss": 3.4297101429360852,
      "tokens_seen": 2217869312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033121364092276827,
      "loss": 1.3119,
      "theoretical_loss": 3.4297019326851004,
      "tokens_seen": 2217934848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311935807422267,
      "loss": 1.3279,
      "theoretical_loss": 3.4296937227446356,
      "tokens_seen": 2218000384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033117352056168505,
      "loss": 1.409,
      "theoretical_loss": 3.4296855131146695,
      "tokens_seen": 2218065920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311534603811434,
      "loss": 1.3572,
      "theoretical_loss": 3.4296773037951818,
      "tokens_seen": 2218131456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033113340020060177,
      "loss": 1.2476,
      "theoretical_loss": 3.429669094786151,
      "tokens_seen": 2218196992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033111334002006013,
      "loss": 1.3752,
      "theoretical_loss": 3.4296608860875564,
      "tokens_seen": 2218262528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033109327983951855,
      "loss": 1.3066,
      "theoretical_loss": 3.429652677699377,
      "tokens_seen": 2218328064
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033107321965897697,
      "loss": 1.4104,
      "theoretical_loss": 3.429644469621592,
      "tokens_seen": 2218393600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033105315947843533,
      "loss": 1.2898,
      "theoretical_loss": 3.4296362618541805,
      "tokens_seen": 2218459136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003310330992978937,
      "loss": 1.3612,
      "theoretical_loss": 3.429628054397121,
      "tokens_seen": 2218524672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003310130391173521,
      "loss": 1.2916,
      "theoretical_loss": 3.429619847250394,
      "tokens_seen": 2218590208
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033099297893681047,
      "loss": 1.3593,
      "theoretical_loss": 3.4296116404139774,
      "tokens_seen": 2218655744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033097291875626883,
      "loss": 1.2953,
      "theoretical_loss": 3.4296034338878503,
      "tokens_seen": 2218721280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003309528585757272,
      "loss": 1.2146,
      "theoretical_loss": 3.4295952276719923,
      "tokens_seen": 2218786816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003309327983951856,
      "loss": 1.3254,
      "theoretical_loss": 3.4295870217663826,
      "tokens_seen": 2218852352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000330912738214644,
      "loss": 1.3326,
      "theoretical_loss": 3.4295788161709995,
      "tokens_seen": 2218917888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033089267803410234,
      "loss": 1.3057,
      "theoretical_loss": 3.4295706108858233,
      "tokens_seen": 2218983424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003308726178535607,
      "loss": 1.2973,
      "theoretical_loss": 3.4295624059108323,
      "tokens_seen": 2219048960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033085255767301906,
      "loss": 1.3042,
      "theoretical_loss": 3.4295542012460056,
      "tokens_seen": 2219114496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003308324974924775,
      "loss": 1.2186,
      "theoretical_loss": 3.4295459968913224,
      "tokens_seen": 2219180032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033081243731193584,
      "loss": 1.2339,
      "theoretical_loss": 3.4295377928467623,
      "tokens_seen": 2219245568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307923771313942,
      "loss": 1.3124,
      "theoretical_loss": 3.429529589112304,
      "tokens_seen": 2219311104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033077231695085256,
      "loss": 1.4591,
      "theoretical_loss": 3.4295213856879263,
      "tokens_seen": 2219376640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000330752256770311,
      "loss": 1.2751,
      "theoretical_loss": 3.4295131825736087,
      "tokens_seen": 2219442176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033073219658976934,
      "loss": 1.3162,
      "theoretical_loss": 3.4295049797693307,
      "tokens_seen": 2219507712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307121364092277,
      "loss": 1.2756,
      "theoretical_loss": 3.429496777275071,
      "tokens_seen": 2219573248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033069207622868607,
      "loss": 1.249,
      "theoretical_loss": 3.4294885750908084,
      "tokens_seen": 2219638784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033067201604814443,
      "loss": 1.409,
      "theoretical_loss": 3.4294803732165224,
      "tokens_seen": 2219704320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033065195586760285,
      "loss": 1.2564,
      "theoretical_loss": 3.4294721716521925,
      "tokens_seen": 2219769856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003306318956870612,
      "loss": 1.2923,
      "theoretical_loss": 3.4294639703977974,
      "tokens_seen": 2219835392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033061183550651957,
      "loss": 1.2263,
      "theoretical_loss": 3.429455769453316,
      "tokens_seen": 2219900928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033059177532597793,
      "loss": 1.2222,
      "theoretical_loss": 3.4294475688187283,
      "tokens_seen": 2219966464
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033057171514543635,
      "loss": 1.3128,
      "theoretical_loss": 3.4294393684940125,
      "tokens_seen": 2220032000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305516549648947,
      "loss": 1.2149,
      "theoretical_loss": 3.429431168479148,
      "tokens_seen": 2220097536
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 808099,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.131688117980957,
      "objective/train/theoretical_loss": 3.4294229687741145,
      "objective/train/tokens_used": 590622176,
      "theoretical_loss": 3.4294229687741145,
      "tokens_seen": 2220163072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305315947843531,
      "loss": 1.3087,
      "theoretical_loss": 3.4294229687741145,
      "tokens_seen": 2220163072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033051153460381144,
      "loss": 1.3767,
      "theoretical_loss": 3.4294147693788903,
      "tokens_seen": 2220228608
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304914744232698,
      "loss": 1.3219,
      "theoretical_loss": 3.4294065702934553,
      "tokens_seen": 2220294144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304714142427282,
      "loss": 1.2563,
      "theoretical_loss": 3.429398371517788,
      "tokens_seen": 2220359680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304513540621866,
      "loss": 1.3356,
      "theoretical_loss": 3.4293901730518686,
      "tokens_seen": 2220425216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033043129388164494,
      "loss": 1.329,
      "theoretical_loss": 3.4293819748956746,
      "tokens_seen": 2220490752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304112337011033,
      "loss": 1.3432,
      "theoretical_loss": 3.4293737770491868,
      "tokens_seen": 2220556288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303911735205617,
      "loss": 1.2847,
      "theoretical_loss": 3.429365579512383,
      "tokens_seen": 2220621824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303711133400201,
      "loss": 1.3271,
      "theoretical_loss": 3.429357382285244,
      "tokens_seen": 2220687360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033035105315947844,
      "loss": 1.359,
      "theoretical_loss": 3.429349185367747,
      "tokens_seen": 2220752896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303309929789368,
      "loss": 1.4198,
      "theoretical_loss": 3.429340988759873,
      "tokens_seen": 2220818432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033031093279839517,
      "loss": 1.3725,
      "theoretical_loss": 3.4293327924615995,
      "tokens_seen": 2220883968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302908726178536,
      "loss": 1.3385,
      "theoretical_loss": 3.429324596472907,
      "tokens_seen": 2220949504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033027081243731195,
      "loss": 1.347,
      "theoretical_loss": 3.429316400793774,
      "tokens_seen": 2221015040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302507522567703,
      "loss": 1.4334,
      "theoretical_loss": 3.4293082054241797,
      "tokens_seen": 2221080576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033023069207622867,
      "loss": 1.4178,
      "theoretical_loss": 3.429300010364104,
      "tokens_seen": 2221146112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302106318956871,
      "loss": 1.2217,
      "theoretical_loss": 3.429291815613525,
      "tokens_seen": 2221211648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033019057171514545,
      "loss": 1.3204,
      "theoretical_loss": 3.4292836211724222,
      "tokens_seen": 2221277184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003301705115346038,
      "loss": 1.2518,
      "theoretical_loss": 3.4292754270407753,
      "tokens_seen": 2221342720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003301504513540622,
      "loss": 1.3269,
      "theoretical_loss": 3.429267233218563,
      "tokens_seen": 2221408256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033013039117352054,
      "loss": 1.3638,
      "theoretical_loss": 3.4292590397057645,
      "tokens_seen": 2221473792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033011033099297895,
      "loss": 1.256,
      "theoretical_loss": 3.4292508465023595,
      "tokens_seen": 2221539328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300902708124373,
      "loss": 1.3411,
      "theoretical_loss": 3.4292426536083265,
      "tokens_seen": 2221604864
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300702106318957,
      "loss": 1.2805,
      "theoretical_loss": 3.429234461023645,
      "tokens_seen": 2221670400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033005015045135404,
      "loss": 1.3341,
      "theoretical_loss": 3.429226268748294,
      "tokens_seen": 2221735936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033003009027081245,
      "loss": 1.2051,
      "theoretical_loss": 3.4292180767822535,
      "tokens_seen": 2221801472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300100300902708,
      "loss": 1.373,
      "theoretical_loss": 3.429209885125502,
      "tokens_seen": 2221867008
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299899699097292,
      "loss": 1.2722,
      "theoretical_loss": 3.429201693778018,
      "tokens_seen": 2221932544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032996990972918754,
      "loss": 1.2885,
      "theoretical_loss": 3.429193502739782,
      "tokens_seen": 2221998080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299498495486459,
      "loss": 1.3181,
      "theoretical_loss": 3.4291853120107727,
      "tokens_seen": 2222063616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299297893681043,
      "loss": 1.2666,
      "theoretical_loss": 3.4291771215909694,
      "tokens_seen": 2222129152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299097291875627,
      "loss": 1.4222,
      "theoretical_loss": 3.4291689314803513,
      "tokens_seen": 2222194688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032988966900702104,
      "loss": 1.3218,
      "theoretical_loss": 3.4291607416788974,
      "tokens_seen": 2222260224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298696088264794,
      "loss": 1.3142,
      "theoretical_loss": 3.429152552186587,
      "tokens_seen": 2222325760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298495486459378,
      "loss": 1.3737,
      "theoretical_loss": 3.4291443630033993,
      "tokens_seen": 2222391296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298294884653962,
      "loss": 1.3868,
      "theoretical_loss": 3.4291361741293134,
      "tokens_seen": 2222456832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032980942828485455,
      "loss": 1.3815,
      "theoretical_loss": 3.4291279855643086,
      "tokens_seen": 2222522368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297893681043129,
      "loss": 1.2976,
      "theoretical_loss": 3.4291197973083647,
      "tokens_seen": 2222587904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297693079237713,
      "loss": 1.3086,
      "theoretical_loss": 3.42911160936146,
      "tokens_seen": 2222653440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297492477432297,
      "loss": 1.3181,
      "theoretical_loss": 3.4291034217235747,
      "tokens_seen": 2222718976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032972918756268805,
      "loss": 1.3175,
      "theoretical_loss": 3.429095234394687,
      "tokens_seen": 2222784512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297091273821464,
      "loss": 1.3317,
      "theoretical_loss": 3.4290870473747765,
      "tokens_seen": 2222850048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296890672016048,
      "loss": 1.28,
      "theoretical_loss": 3.429078860663823,
      "tokens_seen": 2222915584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296690070210632,
      "loss": 1.3931,
      "theoretical_loss": 3.429070674261805,
      "tokens_seen": 2222981120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032964894684052155,
      "loss": 1.3234,
      "theoretical_loss": 3.429062488168702,
      "tokens_seen": 2223046656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296288866599799,
      "loss": 1.3378,
      "theoretical_loss": 3.4290543023844933,
      "tokens_seen": 2223112192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296088264794383,
      "loss": 1.2408,
      "theoretical_loss": 3.429046116909158,
      "tokens_seen": 2223177728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295887662988967,
      "loss": 1.2164,
      "theoretical_loss": 3.429037931742675,
      "tokens_seen": 2223243264
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295687061183551,
      "loss": 1.3058,
      "theoretical_loss": 3.429029746885025,
      "tokens_seen": 2223308800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295486459378135,
      "loss": 1.4393,
      "theoretical_loss": 3.4290215623361853,
      "tokens_seen": 2223374336
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 809438,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2450859546661377,
      "objective/train/theoretical_loss": 3.4290133780961365,
      "objective/train/tokens_used": 593898976,
      "theoretical_loss": 3.4290133780961365,
      "tokens_seen": 2223439872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032952858575727184,
      "loss": 1.3189,
      "theoretical_loss": 3.4290133780961365,
      "tokens_seen": 2223439872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295085255767302,
      "loss": 1.296,
      "theoretical_loss": 3.429005194164857,
      "tokens_seen": 2223505408
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294884653961886,
      "loss": 1.422,
      "theoretical_loss": 3.428997010542327,
      "tokens_seen": 2223570944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329468405215647,
      "loss": 1.3577,
      "theoretical_loss": 3.4289888272285247,
      "tokens_seen": 2223636480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032944834503510534,
      "loss": 1.316,
      "theoretical_loss": 3.4289806442234303,
      "tokens_seen": 2223702016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294282848545637,
      "loss": 1.4083,
      "theoretical_loss": 3.4289724615270223,
      "tokens_seen": 2223767552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294082246740221,
      "loss": 1.3205,
      "theoretical_loss": 3.4289642791392803,
      "tokens_seen": 2223833088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003293881644934805,
      "loss": 1.4501,
      "theoretical_loss": 3.428956097060184,
      "tokens_seen": 2223898624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032936810431293884,
      "loss": 1.4362,
      "theoretical_loss": 3.4289479152897115,
      "tokens_seen": 2223964160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003293480441323972,
      "loss": 1.2713,
      "theoretical_loss": 3.428939733827843,
      "tokens_seen": 2224029696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032932798395185557,
      "loss": 1.2722,
      "theoretical_loss": 3.4289315526745576,
      "tokens_seen": 2224095232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329307923771314,
      "loss": 1.3474,
      "theoretical_loss": 3.4289233718298346,
      "tokens_seen": 2224160768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032928786359077235,
      "loss": 1.353,
      "theoretical_loss": 3.428915191293653,
      "tokens_seen": 2224226304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292678034102307,
      "loss": 1.4524,
      "theoretical_loss": 3.4289070110659923,
      "tokens_seen": 2224291840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032924774322968907,
      "loss": 1.4401,
      "theoretical_loss": 3.428898831146832,
      "tokens_seen": 2224357376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292276830491475,
      "loss": 1.3023,
      "theoretical_loss": 3.428890651536151,
      "tokens_seen": 2224422912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032920762286860585,
      "loss": 1.1897,
      "theoretical_loss": 3.428882472233928,
      "tokens_seen": 2224488448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291875626880642,
      "loss": 1.3022,
      "theoretical_loss": 3.4288742932401437,
      "tokens_seen": 2224553984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291675025075226,
      "loss": 1.3435,
      "theoretical_loss": 3.4288661145547765,
      "tokens_seen": 2224619520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032914744232698094,
      "loss": 1.2625,
      "theoretical_loss": 3.4288579361778053,
      "tokens_seen": 2224685056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032912738214643935,
      "loss": 1.32,
      "theoretical_loss": 3.4288497581092106,
      "tokens_seen": 2224750592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291073219658977,
      "loss": 1.2711,
      "theoretical_loss": 3.428841580348971,
      "tokens_seen": 2224816128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290872617853561,
      "loss": 1.387,
      "theoretical_loss": 3.428833402897065,
      "tokens_seen": 2224881664
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032906720160481444,
      "loss": 1.4263,
      "theoretical_loss": 3.4288252257534735,
      "tokens_seen": 2224947200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032904714142427286,
      "loss": 1.4015,
      "theoretical_loss": 3.4288170489181744,
      "tokens_seen": 2225012736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290270812437312,
      "loss": 1.3984,
      "theoretical_loss": 3.428808872391148,
      "tokens_seen": 2225078272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290070210631896,
      "loss": 1.3592,
      "theoretical_loss": 3.428800696172373,
      "tokens_seen": 2225143808
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032898696088264794,
      "loss": 1.4372,
      "theoretical_loss": 3.428792520261829,
      "tokens_seen": 2225209344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289669007021063,
      "loss": 1.2257,
      "theoretical_loss": 3.428784344659495,
      "tokens_seen": 2225274880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289468405215647,
      "loss": 1.3395,
      "theoretical_loss": 3.4287761693653502,
      "tokens_seen": 2225340416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289267803410231,
      "loss": 1.3004,
      "theoretical_loss": 3.4287679943793745,
      "tokens_seen": 2225405952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032890672016048145,
      "loss": 1.2406,
      "theoretical_loss": 3.428759819701547,
      "tokens_seen": 2225471488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288866599799398,
      "loss": 1.352,
      "theoretical_loss": 3.4287516453318467,
      "tokens_seen": 2225537024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288665997993982,
      "loss": 1.2338,
      "theoretical_loss": 3.4287434712702534,
      "tokens_seen": 2225602560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288465396188566,
      "loss": 1.3808,
      "theoretical_loss": 3.428735297516746,
      "tokens_seen": 2225668096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032882647943831495,
      "loss": 1.2405,
      "theoretical_loss": 3.4287271240713038,
      "tokens_seen": 2225733632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288064192577733,
      "loss": 1.4105,
      "theoretical_loss": 3.4287189509339067,
      "tokens_seen": 2225799168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287863590772317,
      "loss": 1.2982,
      "theoretical_loss": 3.428710778104533,
      "tokens_seen": 2225864704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287662988966901,
      "loss": 1.3408,
      "theoretical_loss": 3.4287026055831626,
      "tokens_seen": 2225930240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032874623871614845,
      "loss": 1.194,
      "theoretical_loss": 3.4286944333697753,
      "tokens_seen": 2225995776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287261785356068,
      "loss": 1.3261,
      "theoretical_loss": 3.4286862614643496,
      "tokens_seen": 2226061312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287061183550652,
      "loss": 1.2783,
      "theoretical_loss": 3.4286780898668656,
      "tokens_seen": 2226126848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286860581745236,
      "loss": 1.39,
      "theoretical_loss": 3.428669918577302,
      "tokens_seen": 2226192384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032866599799398195,
      "loss": 1.2794,
      "theoretical_loss": 3.4286617475956382,
      "tokens_seen": 2226257920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286459378134403,
      "loss": 1.3567,
      "theoretical_loss": 3.428653576921854,
      "tokens_seen": 2226323456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286258776328987,
      "loss": 1.4562,
      "theoretical_loss": 3.428645406555928,
      "tokens_seen": 2226388992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286058174523571,
      "loss": 1.3584,
      "theoretical_loss": 3.42863723649784,
      "tokens_seen": 2226454528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032858575727181546,
      "loss": 1.3346,
      "theoretical_loss": 3.42862906674757,
      "tokens_seen": 2226520064
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285656970912738,
      "loss": 1.1057,
      "theoretical_loss": 3.428620897305096,
      "tokens_seen": 2226585600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285456369107322,
      "loss": 1.3695,
      "theoretical_loss": 3.4286127281703984,
      "tokens_seen": 2226651136
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 810670,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.169150471687317,
      "objective/train/theoretical_loss": 3.428604559343456,
      "objective/train/tokens_used": 597175776,
      "theoretical_loss": 3.428604559343456,
      "tokens_seen": 2226716672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032852557673019054,
      "loss": 1.2969,
      "theoretical_loss": 3.428604559343456,
      "tokens_seen": 2226716672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032850551654964896,
      "loss": 1.3127,
      "theoretical_loss": 3.4285963908242483,
      "tokens_seen": 2226782208
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284854563691073,
      "loss": 1.3027,
      "theoretical_loss": 3.4285882226127544,
      "tokens_seen": 2226847744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284653961885657,
      "loss": 1.321,
      "theoretical_loss": 3.428580054708954,
      "tokens_seen": 2226913280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032844533600802405,
      "loss": 1.3251,
      "theoretical_loss": 3.428571887112826,
      "tokens_seen": 2226978816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032842527582748246,
      "loss": 1.4209,
      "theoretical_loss": 3.428563719824351,
      "tokens_seen": 2227044352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284052156469408,
      "loss": 1.3158,
      "theoretical_loss": 3.4285555528435068,
      "tokens_seen": 2227109888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283851554663992,
      "loss": 1.291,
      "theoretical_loss": 3.4285473861702735,
      "tokens_seen": 2227175424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032836509528585755,
      "loss": 1.2967,
      "theoretical_loss": 3.4285392198046303,
      "tokens_seen": 2227240960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283450351053159,
      "loss": 1.3661,
      "theoretical_loss": 3.428531053746557,
      "tokens_seen": 2227306496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032832497492477433,
      "loss": 1.2764,
      "theoretical_loss": 3.4285228879960323,
      "tokens_seen": 2227372032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283049147442327,
      "loss": 1.3949,
      "theoretical_loss": 3.428514722553036,
      "tokens_seen": 2227437568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032828485456369105,
      "loss": 1.2696,
      "theoretical_loss": 3.4285065574175473,
      "tokens_seen": 2227503104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282647943831494,
      "loss": 1.2822,
      "theoretical_loss": 3.428498392589546,
      "tokens_seen": 2227568640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032824473420260783,
      "loss": 1.3233,
      "theoretical_loss": 3.428490228069011,
      "tokens_seen": 2227634176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282246740220662,
      "loss": 1.2422,
      "theoretical_loss": 3.4284820638559212,
      "tokens_seen": 2227699712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032820461384152456,
      "loss": 1.4608,
      "theoretical_loss": 3.428473899950257,
      "tokens_seen": 2227765248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281845536609829,
      "loss": 1.2156,
      "theoretical_loss": 3.4284657363519973,
      "tokens_seen": 2227830784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281644934804413,
      "loss": 1.3026,
      "theoretical_loss": 3.4284575730611215,
      "tokens_seen": 2227896320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281444332998997,
      "loss": 1.4539,
      "theoretical_loss": 3.428449410077609,
      "tokens_seen": 2227961856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032812437311935806,
      "loss": 1.2724,
      "theoretical_loss": 3.428441247401439,
      "tokens_seen": 2228027392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281043129388164,
      "loss": 1.3554,
      "theoretical_loss": 3.4284330850325917,
      "tokens_seen": 2228092928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003280842527582748,
      "loss": 1.2716,
      "theoretical_loss": 3.428424922971045,
      "tokens_seen": 2228158464
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032806419257773326,
      "loss": 1.2671,
      "theoretical_loss": 3.4284167612167797,
      "tokens_seen": 2228224000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003280441323971916,
      "loss": 1.209,
      "theoretical_loss": 3.428408599769775,
      "tokens_seen": 2228289536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032802407221665,
      "loss": 1.3737,
      "theoretical_loss": 3.428400438630009,
      "tokens_seen": 2228355072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032800401203610834,
      "loss": 1.3199,
      "theoretical_loss": 3.428392277797463,
      "tokens_seen": 2228420608
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003279839518555667,
      "loss": 1.418,
      "theoretical_loss": 3.428384117272115,
      "tokens_seen": 2228486144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003279638916750251,
      "loss": 1.2577,
      "theoretical_loss": 3.428375957053945,
      "tokens_seen": 2228551680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003279438314944835,
      "loss": 1.3428,
      "theoretical_loss": 3.428367797142932,
      "tokens_seen": 2228617216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032792377131394185,
      "loss": 1.3765,
      "theoretical_loss": 3.4283596375390557,
      "tokens_seen": 2228682752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003279037111334002,
      "loss": 1.3115,
      "theoretical_loss": 3.4283514782422957,
      "tokens_seen": 2228748288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278836509528586,
      "loss": 1.2102,
      "theoretical_loss": 3.4283433192526314,
      "tokens_seen": 2228813824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327863590772317,
      "loss": 1.3156,
      "theoretical_loss": 3.4283351605700414,
      "tokens_seen": 2228879360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032784353059177535,
      "loss": 1.3174,
      "theoretical_loss": 3.4283270021945063,
      "tokens_seen": 2228944896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278234704112337,
      "loss": 1.307,
      "theoretical_loss": 3.4283188441260046,
      "tokens_seen": 2229010432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032780341023069213,
      "loss": 1.3305,
      "theoretical_loss": 3.428310686364516,
      "tokens_seen": 2229075968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277833500501505,
      "loss": 1.4379,
      "theoretical_loss": 3.4283025289100197,
      "tokens_seen": 2229141504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032776328986960885,
      "loss": 1.3166,
      "theoretical_loss": 3.428294371762496,
      "tokens_seen": 2229207040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277432296890672,
      "loss": 1.35,
      "theoretical_loss": 3.4282862149219233,
      "tokens_seen": 2229272576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277231695085256,
      "loss": 1.3105,
      "theoretical_loss": 3.4282780583882815,
      "tokens_seen": 2229338112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327703109327984,
      "loss": 1.3376,
      "theoretical_loss": 3.42826990216155,
      "tokens_seen": 2229403648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032768304914744235,
      "loss": 1.428,
      "theoretical_loss": 3.4282617462417084,
      "tokens_seen": 2229469184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276629889669007,
      "loss": 1.3062,
      "theoretical_loss": 3.428253590628736,
      "tokens_seen": 2229534720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276429287863591,
      "loss": 1.3647,
      "theoretical_loss": 3.428245435322612,
      "tokens_seen": 2229600256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276228686058175,
      "loss": 1.3005,
      "theoretical_loss": 3.428237280323316,
      "tokens_seen": 2229665792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032760280842527586,
      "loss": 1.2641,
      "theoretical_loss": 3.4282291256308275,
      "tokens_seen": 2229731328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275827482447342,
      "loss": 1.4044,
      "theoretical_loss": 3.4282209712451257,
      "tokens_seen": 2229796864
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275626880641926,
      "loss": 1.2942,
      "theoretical_loss": 3.42821281716619,
      "tokens_seen": 2229862400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032754262788365094,
      "loss": 1.3273,
      "theoretical_loss": 3.4282046633940007,
      "tokens_seen": 2229927936
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 811859,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4204270839691162,
      "objective/train/theoretical_loss": 3.428196509928536,
      "objective/train/tokens_used": 600452576,
      "theoretical_loss": 3.428196509928536,
      "tokens_seen": 2229993472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032752256770310936,
      "loss": 1.3451,
      "theoretical_loss": 3.428196509928536,
      "tokens_seen": 2229993472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275025075225677,
      "loss": 1.2626,
      "theoretical_loss": 3.4281883567697764,
      "tokens_seen": 2230059008
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274824473420261,
      "loss": 1.4101,
      "theoretical_loss": 3.4281802039177007,
      "tokens_seen": 2230124544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032746238716148445,
      "loss": 1.406,
      "theoretical_loss": 3.4281720513722886,
      "tokens_seen": 2230190080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032744232698094286,
      "loss": 1.2918,
      "theoretical_loss": 3.4281638991335197,
      "tokens_seen": 2230255616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274222668004012,
      "loss": 1.2664,
      "theoretical_loss": 3.428155747201373,
      "tokens_seen": 2230321152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274022066198596,
      "loss": 1.3028,
      "theoretical_loss": 3.4281475955758287,
      "tokens_seen": 2230386688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032738214643931795,
      "loss": 1.3383,
      "theoretical_loss": 3.4281394442568653,
      "tokens_seen": 2230452224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003273620862587763,
      "loss": 1.3342,
      "theoretical_loss": 3.428131293244463,
      "tokens_seen": 2230517760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032734202607823473,
      "loss": 1.3421,
      "theoretical_loss": 3.4281231425386007,
      "tokens_seen": 2230583296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003273219658976931,
      "loss": 1.3,
      "theoretical_loss": 3.4281149921392586,
      "tokens_seen": 2230648832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032730190571715145,
      "loss": 1.2941,
      "theoretical_loss": 3.4281068420464154,
      "tokens_seen": 2230714368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272818455366098,
      "loss": 1.3592,
      "theoretical_loss": 3.428098692260051,
      "tokens_seen": 2230779904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032726178535606823,
      "loss": 1.3582,
      "theoretical_loss": 3.428090542780145,
      "tokens_seen": 2230845440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272417251755266,
      "loss": 1.2604,
      "theoretical_loss": 3.4280823936066764,
      "tokens_seen": 2230910976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032722166499498496,
      "loss": 1.4166,
      "theoretical_loss": 3.428074244739625,
      "tokens_seen": 2230976512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272016048144433,
      "loss": 1.2577,
      "theoretical_loss": 3.4280660961789704,
      "tokens_seen": 2231042048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271815446339017,
      "loss": 1.2681,
      "theoretical_loss": 3.4280579479246915,
      "tokens_seen": 2231107584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271614844533601,
      "loss": 1.2704,
      "theoretical_loss": 3.428049799976769,
      "tokens_seen": 2231173120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032714142427281846,
      "loss": 1.2891,
      "theoretical_loss": 3.428041652335181,
      "tokens_seen": 2231238656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271213640922768,
      "loss": 1.2308,
      "theoretical_loss": 3.4280335049999073,
      "tokens_seen": 2231304192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271013039117352,
      "loss": 1.2789,
      "theoretical_loss": 3.428025357970928,
      "tokens_seen": 2231369728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270812437311936,
      "loss": 1.2353,
      "theoretical_loss": 3.428017211248222,
      "tokens_seen": 2231435264
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032706118355065196,
      "loss": 1.3457,
      "theoretical_loss": 3.4280090648317696,
      "tokens_seen": 2231500800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270411233701103,
      "loss": 1.3621,
      "theoretical_loss": 3.428000918721549,
      "tokens_seen": 2231566336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270210631895687,
      "loss": 1.351,
      "theoretical_loss": 3.427992772917541,
      "tokens_seen": 2231631872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032700100300902705,
      "loss": 1.426,
      "theoretical_loss": 3.4279846274197237,
      "tokens_seen": 2231697408
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032698094282848547,
      "loss": 1.3148,
      "theoretical_loss": 3.427976482228078,
      "tokens_seen": 2231762944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032696088264794383,
      "loss": 1.4199,
      "theoretical_loss": 3.427968337342583,
      "tokens_seen": 2231828480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003269408224674022,
      "loss": 1.2463,
      "theoretical_loss": 3.4279601927632175,
      "tokens_seen": 2231894016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032692076228686055,
      "loss": 1.379,
      "theoretical_loss": 3.4279520484899617,
      "tokens_seen": 2231959552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032690070210631897,
      "loss": 1.3451,
      "theoretical_loss": 3.4279439045227953,
      "tokens_seen": 2232025088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032688064192577733,
      "loss": 1.3589,
      "theoretical_loss": 3.427935760861697,
      "tokens_seen": 2232090624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268605817452357,
      "loss": 1.3003,
      "theoretical_loss": 3.4279276175066467,
      "tokens_seen": 2232156160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032684052156469406,
      "loss": 1.4179,
      "theoretical_loss": 3.427919474457624,
      "tokens_seen": 2232221696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268204613841524,
      "loss": 1.2903,
      "theoretical_loss": 3.4279113317146086,
      "tokens_seen": 2232287232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032680040120361084,
      "loss": 1.3979,
      "theoretical_loss": 3.4279031892775795,
      "tokens_seen": 2232352768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267803410230692,
      "loss": 1.3002,
      "theoretical_loss": 3.427895047146517,
      "tokens_seen": 2232418304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032676028084252756,
      "loss": 1.3035,
      "theoretical_loss": 3.4278869053213996,
      "tokens_seen": 2232483840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267402206619859,
      "loss": 1.3255,
      "theoretical_loss": 3.4278787638022075,
      "tokens_seen": 2232549376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032672016048144434,
      "loss": 1.3246,
      "theoretical_loss": 3.4278706225889204,
      "tokens_seen": 2232614912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267001003009027,
      "loss": 1.282,
      "theoretical_loss": 3.427862481681517,
      "tokens_seen": 2232680448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032668004012036106,
      "loss": 1.323,
      "theoretical_loss": 3.4278543410799776,
      "tokens_seen": 2232745984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266599799398194,
      "loss": 1.3203,
      "theoretical_loss": 3.4278462007842814,
      "tokens_seen": 2232811520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032663991975927784,
      "loss": 1.318,
      "theoretical_loss": 3.427838060794408,
      "tokens_seen": 2232877056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266198595787362,
      "loss": 1.3438,
      "theoretical_loss": 3.427829921110337,
      "tokens_seen": 2232942592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032659979939819457,
      "loss": 1.282,
      "theoretical_loss": 3.4278217817320478,
      "tokens_seen": 2233008128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032657973921765293,
      "loss": 1.3253,
      "theoretical_loss": 3.42781364265952,
      "tokens_seen": 2233073664
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032655967903711135,
      "loss": 1.3459,
      "theoretical_loss": 3.4278055038927335,
      "tokens_seen": 2233139200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032653961885656976,
      "loss": 1.4018,
      "theoretical_loss": 3.427797365431667,
      "tokens_seen": 2233204736
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 813220,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3615901470184326,
      "objective/train/theoretical_loss": 3.4277892272763006,
      "objective/train/tokens_used": 603729376,
      "theoretical_loss": 3.4277892272763006,
      "tokens_seen": 2233270272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003265195586760281,
      "loss": 1.2716,
      "theoretical_loss": 3.4277892272763006,
      "tokens_seen": 2233270272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264994984954865,
      "loss": 1.3807,
      "theoretical_loss": 3.427781089426614,
      "tokens_seen": 2233335808
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032647943831494485,
      "loss": 1.3369,
      "theoretical_loss": 3.4277729518825866,
      "tokens_seen": 2233401344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032645937813440326,
      "loss": 1.2356,
      "theoretical_loss": 3.4277648146441977,
      "tokens_seen": 2233466880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264393179538616,
      "loss": 1.3094,
      "theoretical_loss": 3.427756677711427,
      "tokens_seen": 2233532416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032641925777332,
      "loss": 1.3786,
      "theoretical_loss": 3.427748541084254,
      "tokens_seen": 2233597952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032639919759277835,
      "loss": 1.308,
      "theoretical_loss": 3.427740404762659,
      "tokens_seen": 2233663488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263791374122367,
      "loss": 1.2941,
      "theoretical_loss": 3.4277322687466203,
      "tokens_seen": 2233729024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032635907723169513,
      "loss": 1.2268,
      "theoretical_loss": 3.427724133036118,
      "tokens_seen": 2233794560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263390170511535,
      "loss": 1.2973,
      "theoretical_loss": 3.427715997631132,
      "tokens_seen": 2233860096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032631895687061185,
      "loss": 1.2539,
      "theoretical_loss": 3.4277078625316415,
      "tokens_seen": 2233925632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262988966900702,
      "loss": 1.3549,
      "theoretical_loss": 3.427699727737626,
      "tokens_seen": 2233991168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032627883650952863,
      "loss": 1.2435,
      "theoretical_loss": 3.4276915932490657,
      "tokens_seen": 2234056704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326258776328987,
      "loss": 1.2663,
      "theoretical_loss": 3.4276834590659395,
      "tokens_seen": 2234122240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032623871614844536,
      "loss": 1.4961,
      "theoretical_loss": 3.4276753251882273,
      "tokens_seen": 2234187776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262186559679037,
      "loss": 1.3321,
      "theoretical_loss": 3.427667191615908,
      "tokens_seen": 2234253312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261985957873621,
      "loss": 1.4481,
      "theoretical_loss": 3.4276590583489623,
      "tokens_seen": 2234318848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261785356068205,
      "loss": 1.3234,
      "theoretical_loss": 3.427650925387369,
      "tokens_seen": 2234384384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032615847542627886,
      "loss": 1.4162,
      "theoretical_loss": 3.4276427927311075,
      "tokens_seen": 2234449920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261384152457372,
      "loss": 1.2544,
      "theoretical_loss": 3.427634660380158,
      "tokens_seen": 2234515456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261183550651956,
      "loss": 1.3533,
      "theoretical_loss": 3.4276265283345,
      "tokens_seen": 2234580992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326098294884654,
      "loss": 1.3068,
      "theoretical_loss": 3.427618396594113,
      "tokens_seen": 2234646528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032607823470411236,
      "loss": 1.3381,
      "theoretical_loss": 3.4276102651589766,
      "tokens_seen": 2234712064
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003260581745235707,
      "loss": 1.2924,
      "theoretical_loss": 3.42760213402907,
      "tokens_seen": 2234777600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003260381143430291,
      "loss": 1.3524,
      "theoretical_loss": 3.4275940032043732,
      "tokens_seen": 2234843136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032601805416248745,
      "loss": 1.3387,
      "theoretical_loss": 3.427585872684866,
      "tokens_seen": 2234908672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032599799398194587,
      "loss": 1.4948,
      "theoretical_loss": 3.427577742470527,
      "tokens_seen": 2234974208
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032597793380140423,
      "loss": 1.3096,
      "theoretical_loss": 3.4275696125613373,
      "tokens_seen": 2235039744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003259578736208626,
      "loss": 1.2825,
      "theoretical_loss": 3.427561482957275,
      "tokens_seen": 2235105280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032593781344032095,
      "loss": 1.3896,
      "theoretical_loss": 3.427553353658321,
      "tokens_seen": 2235170816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032591775325977937,
      "loss": 1.2537,
      "theoretical_loss": 3.427545224664454,
      "tokens_seen": 2235236352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032589769307923773,
      "loss": 1.3955,
      "theoretical_loss": 3.4275370959756537,
      "tokens_seen": 2235301888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003258776328986961,
      "loss": 1.3798,
      "theoretical_loss": 3.4275289675919,
      "tokens_seen": 2235367424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032585757271815446,
      "loss": 1.242,
      "theoretical_loss": 3.4275208395131727,
      "tokens_seen": 2235432960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003258375125376128,
      "loss": 1.3125,
      "theoretical_loss": 3.427512711739451,
      "tokens_seen": 2235498496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032581745235707124,
      "loss": 1.3612,
      "theoretical_loss": 3.427504584270715,
      "tokens_seen": 2235564032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003257973921765296,
      "loss": 1.3237,
      "theoretical_loss": 3.427496457106943,
      "tokens_seen": 2235629568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032577733199598796,
      "loss": 1.3474,
      "theoretical_loss": 3.4274883302481163,
      "tokens_seen": 2235695104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257572718154463,
      "loss": 1.4031,
      "theoretical_loss": 3.4274802036942136,
      "tokens_seen": 2235760640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032573721163490474,
      "loss": 1.314,
      "theoretical_loss": 3.427472077445215,
      "tokens_seen": 2235826176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257171514543631,
      "loss": 1.3287,
      "theoretical_loss": 3.4274639515010996,
      "tokens_seen": 2235891712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032569709127382146,
      "loss": 1.2919,
      "theoretical_loss": 3.4274558258618475,
      "tokens_seen": 2235957248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256770310932798,
      "loss": 1.3112,
      "theoretical_loss": 3.4274477005274377,
      "tokens_seen": 2236022784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032565697091273824,
      "loss": 1.2952,
      "theoretical_loss": 3.4274395754978504,
      "tokens_seen": 2236088320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256369107321966,
      "loss": 1.2735,
      "theoretical_loss": 3.427431450773065,
      "tokens_seen": 2236153856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032561685055165497,
      "loss": 1.2751,
      "theoretical_loss": 3.4274233263530616,
      "tokens_seen": 2236219392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032559679037111333,
      "loss": 1.3584,
      "theoretical_loss": 3.427415202237819,
      "tokens_seen": 2236284928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255767301905717,
      "loss": 1.3446,
      "theoretical_loss": 3.427407078427317,
      "tokens_seen": 2236350464
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255566700100301,
      "loss": 1.3523,
      "theoretical_loss": 3.4273989549215362,
      "tokens_seen": 2236416000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032553660982948847,
      "loss": 1.2586,
      "theoretical_loss": 3.427390831720455,
      "tokens_seen": 2236481536
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 814455,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4246420860290527,
      "objective/train/theoretical_loss": 3.4273827088240543,
      "objective/train/tokens_used": 607006176,
      "theoretical_loss": 3.4273827088240543,
      "tokens_seen": 2236547072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032551654964894683,
      "loss": 1.3625,
      "theoretical_loss": 3.4273827088240543,
      "tokens_seen": 2236547072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254964894684052,
      "loss": 1.4448,
      "theoretical_loss": 3.4273745862323124,
      "tokens_seen": 2236612608
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254764292878636,
      "loss": 1.381,
      "theoretical_loss": 3.4273664639452095,
      "tokens_seen": 2236678144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000325456369107322,
      "loss": 1.3451,
      "theoretical_loss": 3.4273583419627256,
      "tokens_seen": 2236743680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032543630892678034,
      "loss": 1.3672,
      "theoretical_loss": 3.42735022028484,
      "tokens_seen": 2236809216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254162487462387,
      "loss": 1.2712,
      "theoretical_loss": 3.4273420989115326,
      "tokens_seen": 2236874752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032539618856569706,
      "loss": 1.4407,
      "theoretical_loss": 3.4273339778427827,
      "tokens_seen": 2236940288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253761283851555,
      "loss": 1.339,
      "theoretical_loss": 3.42732585707857,
      "tokens_seen": 2237005824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032535606820461384,
      "loss": 1.3412,
      "theoretical_loss": 3.4273177366188747,
      "tokens_seen": 2237071360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253360080240722,
      "loss": 1.2962,
      "theoretical_loss": 3.427309616463676,
      "tokens_seen": 2237136896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032531594784353056,
      "loss": 1.3918,
      "theoretical_loss": 3.4273014966129534,
      "tokens_seen": 2237202432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000325295887662989,
      "loss": 1.4419,
      "theoretical_loss": 3.4272933770666865,
      "tokens_seen": 2237267968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032527582748244734,
      "loss": 1.3683,
      "theoretical_loss": 3.4272852578248556,
      "tokens_seen": 2237333504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003252557673019057,
      "loss": 1.3307,
      "theoretical_loss": 3.4272771388874403,
      "tokens_seen": 2237399040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032523570712136407,
      "loss": 1.3817,
      "theoretical_loss": 3.4272690202544194,
      "tokens_seen": 2237464576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032521564694082243,
      "loss": 1.2603,
      "theoretical_loss": 3.4272609019257736,
      "tokens_seen": 2237530112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032519558676028084,
      "loss": 1.3102,
      "theoretical_loss": 3.427252783901482,
      "tokens_seen": 2237595648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003251755265797392,
      "loss": 1.3353,
      "theoretical_loss": 3.4272446661815246,
      "tokens_seen": 2237661184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032515546639919757,
      "loss": 1.3055,
      "theoretical_loss": 3.4272365487658805,
      "tokens_seen": 2237726720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032513540621865593,
      "loss": 1.251,
      "theoretical_loss": 3.4272284316545303,
      "tokens_seen": 2237792256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032511534603811435,
      "loss": 1.3998,
      "theoretical_loss": 3.4272203148474527,
      "tokens_seen": 2237857792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003250952858575727,
      "loss": 1.444,
      "theoretical_loss": 3.4272121983446278,
      "tokens_seen": 2237923328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032507522567703107,
      "loss": 1.391,
      "theoretical_loss": 3.427204082146036,
      "tokens_seen": 2237988864
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003250551654964895,
      "loss": 1.3426,
      "theoretical_loss": 3.427195966251656,
      "tokens_seen": 2238054400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032503510531594785,
      "loss": 1.2968,
      "theoretical_loss": 3.4271878506614675,
      "tokens_seen": 2238119936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032501504513540627,
      "loss": 1.2998,
      "theoretical_loss": 3.4271797353754505,
      "tokens_seen": 2238185472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032499498495486463,
      "loss": 1.43,
      "theoretical_loss": 3.427171620393585,
      "tokens_seen": 2238251008
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324974924774323,
      "loss": 1.3812,
      "theoretical_loss": 3.42716350571585,
      "tokens_seen": 2238316544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032495486459378135,
      "loss": 1.4087,
      "theoretical_loss": 3.427155391342226,
      "tokens_seen": 2238382080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032493480441323977,
      "loss": 1.2584,
      "theoretical_loss": 3.427147277272692,
      "tokens_seen": 2238447616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032491474423269813,
      "loss": 1.385,
      "theoretical_loss": 3.427139163507228,
      "tokens_seen": 2238513152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248946840521565,
      "loss": 1.381,
      "theoretical_loss": 3.427131050045814,
      "tokens_seen": 2238578688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032487462387161486,
      "loss": 1.337,
      "theoretical_loss": 3.427122936888429,
      "tokens_seen": 2238644224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248545636910732,
      "loss": 1.2905,
      "theoretical_loss": 3.427114824035053,
      "tokens_seen": 2238709760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032483450351053164,
      "loss": 1.3592,
      "theoretical_loss": 3.427106711485666,
      "tokens_seen": 2238775296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032481444332999,
      "loss": 1.4158,
      "theoretical_loss": 3.4270985992402476,
      "tokens_seen": 2238840832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032479438314944836,
      "loss": 1.3595,
      "theoretical_loss": 3.427090487298778,
      "tokens_seen": 2238906368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247743229689067,
      "loss": 1.3745,
      "theoretical_loss": 3.4270823756612354,
      "tokens_seen": 2238971904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032475426278836514,
      "loss": 1.3984,
      "theoretical_loss": 3.427074264327601,
      "tokens_seen": 2239037440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247342026078235,
      "loss": 1.2992,
      "theoretical_loss": 3.4270661532978535,
      "tokens_seen": 2239102976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032471414242728186,
      "loss": 1.3502,
      "theoretical_loss": 3.4270580425719737,
      "tokens_seen": 2239168512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003246940822467402,
      "loss": 1.2706,
      "theoretical_loss": 3.4270499321499406,
      "tokens_seen": 2239234048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032467402206619864,
      "loss": 1.273,
      "theoretical_loss": 3.427041822031734,
      "tokens_seen": 2239299584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324653961885657,
      "loss": 1.3548,
      "theoretical_loss": 3.4270337122173333,
      "tokens_seen": 2239365120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032463390170511537,
      "loss": 1.3315,
      "theoretical_loss": 3.427025602706719,
      "tokens_seen": 2239430656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032461384152457373,
      "loss": 1.2724,
      "theoretical_loss": 3.4270174934998705,
      "tokens_seen": 2239496192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003245937813440321,
      "loss": 1.3239,
      "theoretical_loss": 3.4270093845967673,
      "tokens_seen": 2239561728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003245737211634905,
      "loss": 1.2367,
      "theoretical_loss": 3.427001275997389,
      "tokens_seen": 2239627264
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032455366098294887,
      "loss": 1.355,
      "theoretical_loss": 3.4269931677017165,
      "tokens_seen": 2239692800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032453360080240723,
      "loss": 1.2892,
      "theoretical_loss": 3.426985059709728,
      "tokens_seen": 2239758336
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 815683,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.283194899559021,
      "objective/train/theoretical_loss": 3.426976952021404,
      "objective/train/tokens_used": 610282976,
      "theoretical_loss": 3.426976952021404,
      "tokens_seen": 2239823872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003245135406218656,
      "loss": 1.3528,
      "theoretical_loss": 3.426976952021404,
      "tokens_seen": 2239823872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324493480441324,
      "loss": 1.2866,
      "theoretical_loss": 3.4269688446367246,
      "tokens_seen": 2239889408
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244734202607824,
      "loss": 1.314,
      "theoretical_loss": 3.4269607375556683,
      "tokens_seen": 2239954944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032445336008024074,
      "loss": 1.4064,
      "theoretical_loss": 3.4269526307782163,
      "tokens_seen": 2240020480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244332998996991,
      "loss": 1.3326,
      "theoretical_loss": 3.4269445243043477,
      "tokens_seen": 2240086016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032441323971915746,
      "loss": 1.2591,
      "theoretical_loss": 3.426936418134042,
      "tokens_seen": 2240151552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243931795386159,
      "loss": 1.372,
      "theoretical_loss": 3.4269283122672793,
      "tokens_seen": 2240217088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032437311935807424,
      "loss": 1.3037,
      "theoretical_loss": 3.426920206704039,
      "tokens_seen": 2240282624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243530591775326,
      "loss": 1.4723,
      "theoretical_loss": 3.4269121014443016,
      "tokens_seen": 2240348160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032433299899699096,
      "loss": 1.437,
      "theoretical_loss": 3.426903996488046,
      "tokens_seen": 2240413696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243129388164494,
      "loss": 1.3004,
      "theoretical_loss": 3.426895891835253,
      "tokens_seen": 2240479232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032429287863590774,
      "loss": 1.311,
      "theoretical_loss": 3.426887787485901,
      "tokens_seen": 2240544768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003242728184553661,
      "loss": 1.3843,
      "theoretical_loss": 3.426879683439971,
      "tokens_seen": 2240610304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032425275827482447,
      "loss": 1.3201,
      "theoretical_loss": 3.4268715796974414,
      "tokens_seen": 2240675840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032423269809428283,
      "loss": 1.3486,
      "theoretical_loss": 3.4268634762582937,
      "tokens_seen": 2240741376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032421263791374125,
      "loss": 1.3286,
      "theoretical_loss": 3.426855373122506,
      "tokens_seen": 2240806912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241925777331996,
      "loss": 1.3047,
      "theoretical_loss": 3.426847270290059,
      "tokens_seen": 2240872448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032417251755265797,
      "loss": 1.3084,
      "theoretical_loss": 3.426839167760933,
      "tokens_seen": 2240937984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032415245737211633,
      "loss": 1.39,
      "theoretical_loss": 3.4268310655351066,
      "tokens_seen": 2241003520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032413239719157475,
      "loss": 1.3173,
      "theoretical_loss": 3.42682296361256,
      "tokens_seen": 2241069056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241123370110331,
      "loss": 1.3257,
      "theoretical_loss": 3.426814861993273,
      "tokens_seen": 2241134592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032409227683049147,
      "loss": 1.2635,
      "theoretical_loss": 3.4268067606772257,
      "tokens_seen": 2241200128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032407221664994984,
      "loss": 1.3251,
      "theoretical_loss": 3.4267986596643976,
      "tokens_seen": 2241265664
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003240521564694082,
      "loss": 1.3612,
      "theoretical_loss": 3.4267905589547683,
      "tokens_seen": 2241331200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003240320962888666,
      "loss": 1.2525,
      "theoretical_loss": 3.4267824585483178,
      "tokens_seen": 2241396736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324012036108325,
      "loss": 1.3227,
      "theoretical_loss": 3.4267743584450256,
      "tokens_seen": 2241462272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032399197592778334,
      "loss": 1.2418,
      "theoretical_loss": 3.4267662586448724,
      "tokens_seen": 2241527808
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003239719157472417,
      "loss": 1.3262,
      "theoretical_loss": 3.4267581591478367,
      "tokens_seen": 2241593344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003239518555667001,
      "loss": 1.3126,
      "theoretical_loss": 3.4267500599538994,
      "tokens_seen": 2241658880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003239317953861585,
      "loss": 1.3415,
      "theoretical_loss": 3.4267419610630396,
      "tokens_seen": 2241724416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032391173520561684,
      "loss": 1.3013,
      "theoretical_loss": 3.4267338624752375,
      "tokens_seen": 2241789952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003238916750250752,
      "loss": 1.3569,
      "theoretical_loss": 3.426725764190473,
      "tokens_seen": 2241855488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032387161484453357,
      "loss": 1.3318,
      "theoretical_loss": 3.426717666208725,
      "tokens_seen": 2241921024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000323851554663992,
      "loss": 1.3828,
      "theoretical_loss": 3.426709568529974,
      "tokens_seen": 2241986560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032383149448345034,
      "loss": 1.3944,
      "theoretical_loss": 3.4267014711542005,
      "tokens_seen": 2242052096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003238114343029087,
      "loss": 1.2655,
      "theoretical_loss": 3.426693374081383,
      "tokens_seen": 2242117632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032379137412236707,
      "loss": 1.2596,
      "theoretical_loss": 3.4266852773115017,
      "tokens_seen": 2242183168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003237713139418255,
      "loss": 1.3989,
      "theoretical_loss": 3.4266771808445373,
      "tokens_seen": 2242248704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032375125376128385,
      "loss": 1.2744,
      "theoretical_loss": 3.4266690846804684,
      "tokens_seen": 2242314240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003237311935807422,
      "loss": 1.2889,
      "theoretical_loss": 3.4266609888192754,
      "tokens_seen": 2242379776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032371113340020057,
      "loss": 1.3091,
      "theoretical_loss": 3.426652893260938,
      "tokens_seen": 2242445312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032369107321965893,
      "loss": 1.3262,
      "theoretical_loss": 3.426644798005436,
      "tokens_seen": 2242510848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032367101303911735,
      "loss": 1.3671,
      "theoretical_loss": 3.4266367030527496,
      "tokens_seen": 2242576384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236509528585757,
      "loss": 1.2783,
      "theoretical_loss": 3.426628608402858,
      "tokens_seen": 2242641920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236308926780341,
      "loss": 1.3074,
      "theoretical_loss": 3.4266205140557413,
      "tokens_seen": 2242707456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032361083249749244,
      "loss": 1.4128,
      "theoretical_loss": 3.4266124200113794,
      "tokens_seen": 2242772992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032359077231695085,
      "loss": 1.3533,
      "theoretical_loss": 3.4266043262697523,
      "tokens_seen": 2242838528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032357071213640927,
      "loss": 1.2707,
      "theoretical_loss": 3.4265962328308395,
      "tokens_seen": 2242904064
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032355065195586763,
      "loss": 1.3119,
      "theoretical_loss": 3.426588139694621,
      "tokens_seen": 2242969600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000323530591775326,
      "loss": 1.2203,
      "theoretical_loss": 3.4265800468610763,
      "tokens_seen": 2243035136
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 816816,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2776150703430176,
      "objective/train/theoretical_loss": 3.4265719543301856,
      "objective/train/tokens_used": 613559776,
      "theoretical_loss": 3.4265719543301856,
      "tokens_seen": 2243100672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235105315947844,
      "loss": 1.3051,
      "theoretical_loss": 3.4265719543301856,
      "tokens_seen": 2243100672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234904714142428,
      "loss": 1.1838,
      "theoretical_loss": 3.4265638621019288,
      "tokens_seen": 2243166208
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032347041123370114,
      "loss": 1.2919,
      "theoretical_loss": 3.426555770176286,
      "tokens_seen": 2243231744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234503510531595,
      "loss": 1.3196,
      "theoretical_loss": 3.426547678553236,
      "tokens_seen": 2243297280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032343029087261786,
      "loss": 1.3738,
      "theoretical_loss": 3.4265395872327593,
      "tokens_seen": 2243362816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234102306920763,
      "loss": 1.3357,
      "theoretical_loss": 3.426531496214836,
      "tokens_seen": 2243428352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032339017051153464,
      "loss": 1.3721,
      "theoretical_loss": 3.426523405499446,
      "tokens_seen": 2243493888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000323370110330993,
      "loss": 1.3602,
      "theoretical_loss": 3.4265153150865686,
      "tokens_seen": 2243559424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032335005015045136,
      "loss": 1.4463,
      "theoretical_loss": 3.4265072249761834,
      "tokens_seen": 2243624960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003233299899699098,
      "loss": 1.3772,
      "theoretical_loss": 3.4264991351682714,
      "tokens_seen": 2243690496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032330992978936814,
      "loss": 1.4138,
      "theoretical_loss": 3.4264910456628117,
      "tokens_seen": 2243756032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003232898696088265,
      "loss": 1.2472,
      "theoretical_loss": 3.4264829564597843,
      "tokens_seen": 2243821568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032326980942828487,
      "loss": 1.3262,
      "theoretical_loss": 3.4264748675591687,
      "tokens_seen": 2243887104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032324974924774323,
      "loss": 1.3571,
      "theoretical_loss": 3.426466778960945,
      "tokens_seen": 2243952640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032322968906720165,
      "loss": 1.4162,
      "theoretical_loss": 3.4264586906650933,
      "tokens_seen": 2244018176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032320962888666,
      "loss": 1.345,
      "theoretical_loss": 3.4264506026715935,
      "tokens_seen": 2244083712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032318956870611837,
      "loss": 1.3167,
      "theoretical_loss": 3.426442514980425,
      "tokens_seen": 2244149248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032316950852557673,
      "loss": 1.4397,
      "theoretical_loss": 3.4264344275915684,
      "tokens_seen": 2244214784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032314944834503515,
      "loss": 1.4028,
      "theoretical_loss": 3.426426340505003,
      "tokens_seen": 2244280320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231293881644935,
      "loss": 1.3731,
      "theoretical_loss": 3.4264182537207084,
      "tokens_seen": 2244345856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231093279839519,
      "loss": 1.3621,
      "theoretical_loss": 3.4264101672386653,
      "tokens_seen": 2244411392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032308926780341024,
      "loss": 1.2781,
      "theoretical_loss": 3.4264020810588534,
      "tokens_seen": 2244476928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230692076228686,
      "loss": 1.2786,
      "theoretical_loss": 3.4263939951812517,
      "tokens_seen": 2244542464
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000323049147442327,
      "loss": 1.3825,
      "theoretical_loss": 3.426385909605841,
      "tokens_seen": 2244608000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230290872617854,
      "loss": 1.2917,
      "theoretical_loss": 3.426377824332601,
      "tokens_seen": 2244673536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032300902708124374,
      "loss": 1.4216,
      "theoretical_loss": 3.4263697393615113,
      "tokens_seen": 2244739072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229889669007021,
      "loss": 1.2464,
      "theoretical_loss": 3.426361654692552,
      "tokens_seen": 2244804608
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229689067201605,
      "loss": 1.2479,
      "theoretical_loss": 3.426353570325703,
      "tokens_seen": 2244870144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229488465396189,
      "loss": 1.4348,
      "theoretical_loss": 3.426345486260944,
      "tokens_seen": 2244935680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032292878635907724,
      "loss": 1.3056,
      "theoretical_loss": 3.4263374024982554,
      "tokens_seen": 2245001216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229087261785356,
      "loss": 1.4138,
      "theoretical_loss": 3.4263293190376167,
      "tokens_seen": 2245066752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032288866599799397,
      "loss": 1.335,
      "theoretical_loss": 3.4263212358790076,
      "tokens_seen": 2245132288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228686058174524,
      "loss": 1.4341,
      "theoretical_loss": 3.4263131530224085,
      "tokens_seen": 2245197824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032284854563691074,
      "loss": 1.3424,
      "theoretical_loss": 3.426305070467799,
      "tokens_seen": 2245263360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228284854563691,
      "loss": 1.3051,
      "theoretical_loss": 3.426296988215159,
      "tokens_seen": 2245328896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032280842527582747,
      "loss": 1.293,
      "theoretical_loss": 3.4262889062644684,
      "tokens_seen": 2245394432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003227883650952859,
      "loss": 1.3867,
      "theoretical_loss": 3.4262808246157066,
      "tokens_seen": 2245459968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032276830491474425,
      "loss": 1.3451,
      "theoretical_loss": 3.426272743268855,
      "tokens_seen": 2245525504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003227482447342026,
      "loss": 1.323,
      "theoretical_loss": 3.426264662223892,
      "tokens_seen": 2245591040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032272818455366097,
      "loss": 1.3401,
      "theoretical_loss": 3.426256581480798,
      "tokens_seen": 2245656576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032270812437311933,
      "loss": 1.3086,
      "theoretical_loss": 3.4262485010395536,
      "tokens_seen": 2245722112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032268806419257775,
      "loss": 1.3879,
      "theoretical_loss": 3.426240420900138,
      "tokens_seen": 2245787648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226680040120361,
      "loss": 1.4148,
      "theoretical_loss": 3.426232341062531,
      "tokens_seen": 2245853184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226479438314945,
      "loss": 1.3904,
      "theoretical_loss": 3.4262242615267127,
      "tokens_seen": 2245918720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032262788365095284,
      "loss": 1.364,
      "theoretical_loss": 3.4262161822926633,
      "tokens_seen": 2245984256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032260782347041125,
      "loss": 1.2714,
      "theoretical_loss": 3.4262081033603624,
      "tokens_seen": 2246049792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225877632898696,
      "loss": 1.2822,
      "theoretical_loss": 3.42620002472979,
      "tokens_seen": 2246115328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322567703109328,
      "loss": 1.3282,
      "theoretical_loss": 3.426191946400926,
      "tokens_seen": 2246180864
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032254764292878634,
      "loss": 1.3302,
      "theoretical_loss": 3.4261838683737507,
      "tokens_seen": 2246246400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225275827482447,
      "loss": 1.3235,
      "theoretical_loss": 3.4261757906482435,
      "tokens_seen": 2246311936
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 817879,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2966686487197876,
      "objective/train/theoretical_loss": 3.4261677132243844,
      "objective/train/tokens_used": 616836576,
      "theoretical_loss": 3.4261677132243844,
      "tokens_seen": 2246377472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225075225677031,
      "loss": 1.3322,
      "theoretical_loss": 3.4261677132243844,
      "tokens_seen": 2246377472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224874623871615,
      "loss": 1.295,
      "theoretical_loss": 3.4261596361021533,
      "tokens_seen": 2246443008
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032246740220661984,
      "loss": 1.4355,
      "theoretical_loss": 3.426151559281531,
      "tokens_seen": 2246508544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224473420260782,
      "loss": 1.2629,
      "theoretical_loss": 3.4261434827624964,
      "tokens_seen": 2246574080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224272818455366,
      "loss": 1.2509,
      "theoretical_loss": 3.42613540654503,
      "tokens_seen": 2246639616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322407221664995,
      "loss": 1.4527,
      "theoretical_loss": 3.4261273306291113,
      "tokens_seen": 2246705152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032238716148445335,
      "loss": 1.4129,
      "theoretical_loss": 3.4261192550147204,
      "tokens_seen": 2246770688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223671013039117,
      "loss": 1.4208,
      "theoretical_loss": 3.4261111797018375,
      "tokens_seen": 2246836224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223470411233701,
      "loss": 1.3964,
      "theoretical_loss": 3.4261031046904424,
      "tokens_seen": 2246901760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223269809428285,
      "loss": 1.2715,
      "theoretical_loss": 3.426095029980515,
      "tokens_seen": 2246967296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032230692076228685,
      "loss": 1.3555,
      "theoretical_loss": 3.4260869555720355,
      "tokens_seen": 2247032832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222868605817452,
      "loss": 1.3502,
      "theoretical_loss": 3.4260788814649836,
      "tokens_seen": 2247098368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222668004012036,
      "loss": 1.2704,
      "theoretical_loss": 3.4260708076593396,
      "tokens_seen": 2247163904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322246740220662,
      "loss": 1.2961,
      "theoretical_loss": 3.4260627341550824,
      "tokens_seen": 2247229440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032222668004012035,
      "loss": 1.3102,
      "theoretical_loss": 3.426054660952193,
      "tokens_seen": 2247294976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222066198595787,
      "loss": 1.3569,
      "theoretical_loss": 3.4260465880506517,
      "tokens_seen": 2247360512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221865596790371,
      "loss": 1.3768,
      "theoretical_loss": 3.426038515450437,
      "tokens_seen": 2247426048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221664994984955,
      "loss": 1.3061,
      "theoretical_loss": 3.4260304431515305,
      "tokens_seen": 2247491584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032214643931795386,
      "loss": 1.3181,
      "theoretical_loss": 3.426022371153911,
      "tokens_seen": 2247557120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221263791374122,
      "loss": 1.3183,
      "theoretical_loss": 3.4260142994575586,
      "tokens_seen": 2247622656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221063189568706,
      "loss": 1.1854,
      "theoretical_loss": 3.426006228062454,
      "tokens_seen": 2247688192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032208625877632894,
      "loss": 1.2646,
      "theoretical_loss": 3.4259981569685767,
      "tokens_seen": 2247753728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220661985957874,
      "loss": 1.3579,
      "theoretical_loss": 3.425990086175906,
      "tokens_seen": 2247819264
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220461384152458,
      "loss": 1.3166,
      "theoretical_loss": 3.4259820156844234,
      "tokens_seen": 2247884800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032202607823470414,
      "loss": 1.4076,
      "theoretical_loss": 3.425973945494108,
      "tokens_seen": 2247950336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220060180541625,
      "loss": 1.261,
      "theoretical_loss": 3.425965875604939,
      "tokens_seen": 2248015872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003219859578736209,
      "loss": 1.4944,
      "theoretical_loss": 3.425957806016898,
      "tokens_seen": 2248081408
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003219658976930793,
      "loss": 1.2509,
      "theoretical_loss": 3.425949736729964,
      "tokens_seen": 2248146944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032194583751253764,
      "loss": 1.4037,
      "theoretical_loss": 3.425941667744117,
      "tokens_seen": 2248212480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321925777331996,
      "loss": 1.3695,
      "theoretical_loss": 3.425933599059337,
      "tokens_seen": 2248278016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032190571715145437,
      "loss": 1.3075,
      "theoretical_loss": 3.4259255306756042,
      "tokens_seen": 2248343552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218856569709128,
      "loss": 1.4285,
      "theoretical_loss": 3.425917462592899,
      "tokens_seen": 2248409088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032186559679037115,
      "loss": 1.2572,
      "theoretical_loss": 3.4259093948112005,
      "tokens_seen": 2248474624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218455366098295,
      "loss": 1.3999,
      "theoretical_loss": 3.4259013273304895,
      "tokens_seen": 2248540160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032182547642928787,
      "loss": 1.3358,
      "theoretical_loss": 3.4258932601507457,
      "tokens_seen": 2248605696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218054162487463,
      "loss": 1.5248,
      "theoretical_loss": 3.4258851932719487,
      "tokens_seen": 2248671232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032178535606820465,
      "loss": 1.2936,
      "theoretical_loss": 3.425877126694079,
      "tokens_seen": 2248736768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321765295887663,
      "loss": 1.2266,
      "theoretical_loss": 3.4258690604171163,
      "tokens_seen": 2248802304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032174523570712137,
      "loss": 1.309,
      "theoretical_loss": 3.425860994441041,
      "tokens_seen": 2248867840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032172517552657974,
      "loss": 1.3537,
      "theoretical_loss": 3.425852928765833,
      "tokens_seen": 2248933376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032170511534603815,
      "loss": 1.4119,
      "theoretical_loss": 3.4258448633914718,
      "tokens_seen": 2248998912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216850551654965,
      "loss": 1.3467,
      "theoretical_loss": 3.425836798317938,
      "tokens_seen": 2249064448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216649949849549,
      "loss": 1.204,
      "theoretical_loss": 3.4258287335452113,
      "tokens_seen": 2249129984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032164493480441324,
      "loss": 1.3691,
      "theoretical_loss": 3.4258206690732718,
      "tokens_seen": 2249195520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032162487462387165,
      "loss": 1.2847,
      "theoretical_loss": 3.4258126049020996,
      "tokens_seen": 2249261056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032160481444333,
      "loss": 1.2781,
      "theoretical_loss": 3.425804541031675,
      "tokens_seen": 2249326592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215847542627884,
      "loss": 1.2514,
      "theoretical_loss": 3.425796477461977,
      "tokens_seen": 2249392128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032156469408224674,
      "loss": 1.3658,
      "theoretical_loss": 3.425788414192987,
      "tokens_seen": 2249457664
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215446339017051,
      "loss": 1.2879,
      "theoretical_loss": 3.4257803512246836,
      "tokens_seen": 2249523200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215245737211635,
      "loss": 1.2631,
      "theoretical_loss": 3.425772288557048,
      "tokens_seen": 2249588736
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 819025,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2979280948638916,
      "objective/train/theoretical_loss": 3.4257642261900596,
      "objective/train/tokens_used": 620113376,
      "theoretical_loss": 3.4257642261900596,
      "tokens_seen": 2249654272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215045135406219,
      "loss": 1.3743,
      "theoretical_loss": 3.4257642261900596,
      "tokens_seen": 2249654272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032148445336008024,
      "loss": 1.2688,
      "theoretical_loss": 3.425756164123699,
      "tokens_seen": 2249719808
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214643931795386,
      "loss": 1.2962,
      "theoretical_loss": 3.4257481023579457,
      "tokens_seen": 2249785344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321444332998997,
      "loss": 1.3528,
      "theoretical_loss": 3.425740040892779,
      "tokens_seen": 2249850880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214242728184554,
      "loss": 1.3516,
      "theoretical_loss": 3.425731979728181,
      "tokens_seen": 2249916416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032140421263791375,
      "loss": 1.3329,
      "theoretical_loss": 3.42572391886413,
      "tokens_seen": 2249981952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213841524573721,
      "loss": 1.3181,
      "theoretical_loss": 3.425715858300607,
      "tokens_seen": 2250047488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213640922768305,
      "loss": 1.377,
      "theoretical_loss": 3.425707798037591,
      "tokens_seen": 2250113024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213440320962889,
      "loss": 1.3133,
      "theoretical_loss": 3.425699738075063,
      "tokens_seen": 2250178560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032132397191574725,
      "loss": 1.3571,
      "theoretical_loss": 3.425691678413003,
      "tokens_seen": 2250244096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213039117352056,
      "loss": 1.4143,
      "theoretical_loss": 3.425683619051391,
      "tokens_seen": 2250309632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321283851554664,
      "loss": 1.2556,
      "theoretical_loss": 3.4256755599902062,
      "tokens_seen": 2250375168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212637913741224,
      "loss": 1.2576,
      "theoretical_loss": 3.4256675012294293,
      "tokens_seen": 2250440704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032124373119358075,
      "loss": 1.2748,
      "theoretical_loss": 3.425659442769041,
      "tokens_seen": 2250506240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212236710130391,
      "loss": 1.2606,
      "theoretical_loss": 3.42565138460902,
      "tokens_seen": 2250571776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212036108324975,
      "loss": 1.4338,
      "theoretical_loss": 3.4256433267493476,
      "tokens_seen": 2250637312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211835506519559,
      "loss": 1.2458,
      "theoretical_loss": 3.425635269190003,
      "tokens_seen": 2250702848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032116349047141426,
      "loss": 1.2956,
      "theoretical_loss": 3.4256272119309665,
      "tokens_seen": 2250768384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211434302908726,
      "loss": 1.2375,
      "theoretical_loss": 3.4256191549722184,
      "tokens_seen": 2250833920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321123370110331,
      "loss": 1.3663,
      "theoretical_loss": 3.4256110983137384,
      "tokens_seen": 2250899456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032110330992978934,
      "loss": 1.3749,
      "theoretical_loss": 3.4256030419555072,
      "tokens_seen": 2250964992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032108324974924776,
      "loss": 1.4324,
      "theoretical_loss": 3.4255949858975043,
      "tokens_seen": 2251030528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210631895687061,
      "loss": 1.3478,
      "theoretical_loss": 3.4255869301397097,
      "tokens_seen": 2251096064
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210431293881645,
      "loss": 1.3242,
      "theoretical_loss": 3.4255788746821034,
      "tokens_seen": 2251161600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032102306920762285,
      "loss": 1.3252,
      "theoretical_loss": 3.4255708195246664,
      "tokens_seen": 2251227136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032100300902708126,
      "loss": 1.3484,
      "theoretical_loss": 3.4255627646673776,
      "tokens_seen": 2251292672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003209829488465396,
      "loss": 1.3683,
      "theoretical_loss": 3.425554710110218,
      "tokens_seen": 2251358208
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000320962888665998,
      "loss": 1.3374,
      "theoretical_loss": 3.4255466558531675,
      "tokens_seen": 2251423744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032094282848545635,
      "loss": 1.336,
      "theoretical_loss": 3.4255386018962057,
      "tokens_seen": 2251489280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003209227683049147,
      "loss": 1.4272,
      "theoretical_loss": 3.425530548239313,
      "tokens_seen": 2251554816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032090270812437313,
      "loss": 1.2396,
      "theoretical_loss": 3.425522494882469,
      "tokens_seen": 2251620352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208826479438315,
      "loss": 1.3155,
      "theoretical_loss": 3.425514441825655,
      "tokens_seen": 2251685888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032086258776328985,
      "loss": 1.3442,
      "theoretical_loss": 3.42550638906885,
      "tokens_seen": 2251751424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208425275827482,
      "loss": 1.2906,
      "theoretical_loss": 3.4254983366120344,
      "tokens_seen": 2251816960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032082246740220663,
      "loss": 1.3499,
      "theoretical_loss": 3.425490284455188,
      "tokens_seen": 2251882496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000320802407221665,
      "loss": 1.2708,
      "theoretical_loss": 3.425482232598292,
      "tokens_seen": 2251948032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032078234704112336,
      "loss": 1.3997,
      "theoretical_loss": 3.4254741810413254,
      "tokens_seen": 2252013568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003207622868605817,
      "loss": 1.3667,
      "theoretical_loss": 3.4254661297842683,
      "tokens_seen": 2252079104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003207422266800401,
      "loss": 1.2637,
      "theoretical_loss": 3.4254580788271016,
      "tokens_seen": 2252144640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003207221664994985,
      "loss": 1.3521,
      "theoretical_loss": 3.425450028169805,
      "tokens_seen": 2252210176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032070210631895686,
      "loss": 1.3381,
      "theoretical_loss": 3.4254419778123575,
      "tokens_seen": 2252275712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206820461384152,
      "loss": 1.2596,
      "theoretical_loss": 3.425433927754741,
      "tokens_seen": 2252341248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206619859578736,
      "loss": 1.3667,
      "theoretical_loss": 3.425425877996935,
      "tokens_seen": 2252406784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000320641925777332,
      "loss": 1.3873,
      "theoretical_loss": 3.4254178285389187,
      "tokens_seen": 2252472320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032062186559679036,
      "loss": 1.2714,
      "theoretical_loss": 3.425409779380674,
      "tokens_seen": 2252537856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206018054162487,
      "loss": 1.3055,
      "theoretical_loss": 3.425401730522179,
      "tokens_seen": 2252603392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205817452357071,
      "loss": 1.2931,
      "theoretical_loss": 3.4253936819634156,
      "tokens_seen": 2252668928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205616850551655,
      "loss": 1.3049,
      "theoretical_loss": 3.4253856337043627,
      "tokens_seen": 2252734464
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205416248746239,
      "loss": 1.3062,
      "theoretical_loss": 3.425377585745001,
      "tokens_seen": 2252800000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205215646940823,
      "loss": 1.3105,
      "theoretical_loss": 3.4253695380853104,
      "tokens_seen": 2252865536
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 820362,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4370191097259521,
      "objective/train/theoretical_loss": 3.4253614907252707,
      "objective/train/tokens_used": 623390176,
      "theoretical_loss": 3.4253614907252707,
      "tokens_seen": 2252931072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032050150451354064,
      "loss": 1.394,
      "theoretical_loss": 3.4253614907252707,
      "tokens_seen": 2252931072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000320481444332999,
      "loss": 1.3374,
      "theoretical_loss": 3.425353443664863,
      "tokens_seen": 2252996608
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204613841524574,
      "loss": 1.2829,
      "theoretical_loss": 3.4253453969040666,
      "tokens_seen": 2253062144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204413239719158,
      "loss": 1.2971,
      "theoretical_loss": 3.425337350442862,
      "tokens_seen": 2253127680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032042126379137415,
      "loss": 1.3441,
      "theoretical_loss": 3.425329304281229,
      "tokens_seen": 2253193216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204012036108325,
      "loss": 1.4001,
      "theoretical_loss": 3.4253212584191477,
      "tokens_seen": 2253258752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203811434302909,
      "loss": 1.3022,
      "theoretical_loss": 3.425313212856599,
      "tokens_seen": 2253324288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203610832497493,
      "loss": 1.3169,
      "theoretical_loss": 3.4253051675935624,
      "tokens_seen": 2253389824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032034102306920765,
      "loss": 1.2853,
      "theoretical_loss": 3.425297122630018,
      "tokens_seen": 2253455360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000320320962888666,
      "loss": 1.3437,
      "theoretical_loss": 3.4252890779659464,
      "tokens_seen": 2253520896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203009027081244,
      "loss": 1.2632,
      "theoretical_loss": 3.425281033601327,
      "tokens_seen": 2253586432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202808425275828,
      "loss": 1.3172,
      "theoretical_loss": 3.4252729895361407,
      "tokens_seen": 2253651968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032026078234704115,
      "loss": 1.4267,
      "theoretical_loss": 3.425264945770367,
      "tokens_seen": 2253717504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202407221664995,
      "loss": 1.3465,
      "theoretical_loss": 3.4252569023039867,
      "tokens_seen": 2253783040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202206619859579,
      "loss": 1.3125,
      "theoretical_loss": 3.4252488591369796,
      "tokens_seen": 2253848576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202006018054163,
      "loss": 1.3946,
      "theoretical_loss": 3.4252408162693255,
      "tokens_seen": 2253914112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032018054162487466,
      "loss": 1.421,
      "theoretical_loss": 3.4252327737010053,
      "tokens_seen": 2253979648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000320160481444333,
      "loss": 1.3479,
      "theoretical_loss": 3.425224731431999,
      "tokens_seen": 2254045184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201404212637914,
      "loss": 1.3725,
      "theoretical_loss": 3.425216689462286,
      "tokens_seen": 2254110720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032012036108324974,
      "loss": 1.3045,
      "theoretical_loss": 3.4252086477918473,
      "tokens_seen": 2254176256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032010030090270816,
      "loss": 1.3119,
      "theoretical_loss": 3.425200606420663,
      "tokens_seen": 2254241792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003200802407221665,
      "loss": 1.2955,
      "theoretical_loss": 3.4251925653487127,
      "tokens_seen": 2254307328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003200601805416249,
      "loss": 1.3212,
      "theoretical_loss": 3.425184524575977,
      "tokens_seen": 2254372864
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032004012036108325,
      "loss": 1.4203,
      "theoretical_loss": 3.425176484102436,
      "tokens_seen": 2254438400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032002006018054166,
      "loss": 1.3079,
      "theoretical_loss": 3.4251684439280696,
      "tokens_seen": 2254503936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032,
      "loss": 1.3803,
      "theoretical_loss": 3.4251604040528587,
      "tokens_seen": 2254569472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003199799398194584,
      "loss": 1.2478,
      "theoretical_loss": 3.4251523644767827,
      "tokens_seen": 2254635008
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031995987963891675,
      "loss": 1.3882,
      "theoretical_loss": 3.425144325199822,
      "tokens_seen": 2254700544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003199398194583751,
      "loss": 1.1777,
      "theoretical_loss": 3.425136286221957,
      "tokens_seen": 2254766080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031991975927783353,
      "loss": 1.3226,
      "theoretical_loss": 3.4251282475431677,
      "tokens_seen": 2254831616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198996990972919,
      "loss": 1.3984,
      "theoretical_loss": 3.4251202091634343,
      "tokens_seen": 2254897152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031987963891675025,
      "loss": 1.3665,
      "theoretical_loss": 3.4251121710827372,
      "tokens_seen": 2254962688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198595787362086,
      "loss": 1.3718,
      "theoretical_loss": 3.425104133301056,
      "tokens_seen": 2255028224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031983951855566703,
      "loss": 1.3711,
      "theoretical_loss": 3.4250960958183714,
      "tokens_seen": 2255093760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198194583751254,
      "loss": 1.3844,
      "theoretical_loss": 3.4250880586346635,
      "tokens_seen": 2255159296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031979939819458376,
      "loss": 1.3275,
      "theoretical_loss": 3.4250800217499124,
      "tokens_seen": 2255224832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197793380140421,
      "loss": 1.2946,
      "theoretical_loss": 3.4250719851640983,
      "tokens_seen": 2255290368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197592778335005,
      "loss": 1.3391,
      "theoretical_loss": 3.4250639488772014,
      "tokens_seen": 2255355904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197392176529589,
      "loss": 1.285,
      "theoretical_loss": 3.425055912889202,
      "tokens_seen": 2255421440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031971915747241726,
      "loss": 1.3366,
      "theoretical_loss": 3.4250478772000803,
      "tokens_seen": 2255486976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196990972918756,
      "loss": 1.3796,
      "theoretical_loss": 3.4250398418098165,
      "tokens_seen": 2255552512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319679037111334,
      "loss": 1.3664,
      "theoretical_loss": 3.4250318067183905,
      "tokens_seen": 2255618048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196589769307924,
      "loss": 1.2685,
      "theoretical_loss": 3.4250237719257823,
      "tokens_seen": 2255683584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031963891675025076,
      "loss": 1.2698,
      "theoretical_loss": 3.4250157374319734,
      "tokens_seen": 2255749120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031963891675025076,
      "loss": 1.4421,
      "theoretical_loss": 3.4250077032369424,
      "tokens_seen": 2255814656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196188565697091,
      "loss": 1.337,
      "theoretical_loss": 3.4249996693406706,
      "tokens_seen": 2255880192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003195987963891675,
      "loss": 1.234,
      "theoretical_loss": 3.424991635743138,
      "tokens_seen": 2255945728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031957873620862585,
      "loss": 1.3528,
      "theoretical_loss": 3.4249836024443248,
      "tokens_seen": 2256011264
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031955867602808427,
      "loss": 1.35,
      "theoretical_loss": 3.4249755694442108,
      "tokens_seen": 2256076800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031953861584754263,
      "loss": 1.2912,
      "theoretical_loss": 3.4249675367427765,
      "tokens_seen": 2256142336
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 821597,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2336771488189697,
      "objective/train/theoretical_loss": 3.424959504340002,
      "objective/train/tokens_used": 626666976,
      "theoretical_loss": 3.424959504340002,
      "tokens_seen": 2256207872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319518555667001,
      "loss": 1.2972,
      "theoretical_loss": 3.424959504340002,
      "tokens_seen": 2256207872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031949849548645935,
      "loss": 1.2471,
      "theoretical_loss": 3.424951472235868,
      "tokens_seen": 2256273408
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031947843530591777,
      "loss": 1.3624,
      "theoretical_loss": 3.4249434404303543,
      "tokens_seen": 2256338944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031945837512537613,
      "loss": 1.3168,
      "theoretical_loss": 3.424935408923441,
      "tokens_seen": 2256404480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003194383149448345,
      "loss": 1.3819,
      "theoretical_loss": 3.4249273777151084,
      "tokens_seen": 2256470016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031941825476429286,
      "loss": 1.3789,
      "theoretical_loss": 3.4249193468053374,
      "tokens_seen": 2256535552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193981945837512,
      "loss": 1.3234,
      "theoretical_loss": 3.424911316194107,
      "tokens_seen": 2256601088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031937813440320964,
      "loss": 1.2714,
      "theoretical_loss": 3.4249032858813986,
      "tokens_seen": 2256666624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319358074222668,
      "loss": 1.3801,
      "theoretical_loss": 3.4248952558671917,
      "tokens_seen": 2256732160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031933801404212636,
      "loss": 1.3342,
      "theoretical_loss": 3.424887226151467,
      "tokens_seen": 2256797696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193179538615847,
      "loss": 1.4121,
      "theoretical_loss": 3.4248791967342043,
      "tokens_seen": 2256863232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031929789368104314,
      "loss": 1.3731,
      "theoretical_loss": 3.4248711676153842,
      "tokens_seen": 2256928768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192778335005015,
      "loss": 1.3239,
      "theoretical_loss": 3.424863138794987,
      "tokens_seen": 2256994304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031925777331995986,
      "loss": 1.286,
      "theoretical_loss": 3.4248551102729925,
      "tokens_seen": 2257059840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192377131394182,
      "loss": 1.3224,
      "theoretical_loss": 3.4248470820493813,
      "tokens_seen": 2257125376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192176529588766,
      "loss": 1.2979,
      "theoretical_loss": 3.424839054124133,
      "tokens_seen": 2257190912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319197592778335,
      "loss": 1.4244,
      "theoretical_loss": 3.424831026497229,
      "tokens_seen": 2257256448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031917753259779337,
      "loss": 1.3088,
      "theoretical_loss": 3.424822999168649,
      "tokens_seen": 2257321984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031915747241725173,
      "loss": 1.3579,
      "theoretical_loss": 3.424814972138373,
      "tokens_seen": 2257387520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191374122367101,
      "loss": 1.3548,
      "theoretical_loss": 3.4248069454063814,
      "tokens_seen": 2257453056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191173520561685,
      "loss": 1.2612,
      "theoretical_loss": 3.4247989189726544,
      "tokens_seen": 2257518592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031909729187562687,
      "loss": 1.3465,
      "theoretical_loss": 3.4247908928371724,
      "tokens_seen": 2257584128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031907723169508523,
      "loss": 1.2989,
      "theoretical_loss": 3.424782866999916,
      "tokens_seen": 2257649664
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031905717151454365,
      "loss": 1.3458,
      "theoretical_loss": 3.4247748414608647,
      "tokens_seen": 2257715200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031903711133400206,
      "loss": 1.3281,
      "theoretical_loss": 3.424766816219999,
      "tokens_seen": 2257780736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003190170511534604,
      "loss": 1.2821,
      "theoretical_loss": 3.4247587912773,
      "tokens_seen": 2257846272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003189969909729188,
      "loss": 1.309,
      "theoretical_loss": 3.424750766632747,
      "tokens_seen": 2257911808
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031897693079237715,
      "loss": 1.2786,
      "theoretical_loss": 3.42474274228632,
      "tokens_seen": 2257977344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003189568706118355,
      "loss": 1.3446,
      "theoretical_loss": 3.4247347182380006,
      "tokens_seen": 2258042880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031893681043129393,
      "loss": 1.4057,
      "theoretical_loss": 3.424726694487768,
      "tokens_seen": 2258108416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003189167502507523,
      "loss": 1.3338,
      "theoretical_loss": 3.4247186710356026,
      "tokens_seen": 2258173952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031889669007021065,
      "loss": 1.4403,
      "theoretical_loss": 3.424710647881485,
      "tokens_seen": 2258239488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318876629889669,
      "loss": 1.3422,
      "theoretical_loss": 3.4247026250253954,
      "tokens_seen": 2258305024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031885656970912743,
      "loss": 1.3196,
      "theoretical_loss": 3.424694602467314,
      "tokens_seen": 2258370560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003188365095285858,
      "loss": 1.3789,
      "theoretical_loss": 3.424686580207221,
      "tokens_seen": 2258436096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031881644934804416,
      "loss": 1.3325,
      "theoretical_loss": 3.4246785582450965,
      "tokens_seen": 2258501632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187963891675025,
      "loss": 1.381,
      "theoretical_loss": 3.424670536580922,
      "tokens_seen": 2258567168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187763289869609,
      "loss": 1.2465,
      "theoretical_loss": 3.424662515214676,
      "tokens_seen": 2258632704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187562688064193,
      "loss": 1.3982,
      "theoretical_loss": 3.42465449414634,
      "tokens_seen": 2258698240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031873620862587766,
      "loss": 1.3873,
      "theoretical_loss": 3.4246464733758937,
      "tokens_seen": 2258763776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318716148445336,
      "loss": 1.2336,
      "theoretical_loss": 3.424638452903318,
      "tokens_seen": 2258829312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186960882647944,
      "loss": 1.2054,
      "theoretical_loss": 3.4246304327285926,
      "tokens_seen": 2258894848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186760280842528,
      "loss": 1.3343,
      "theoretical_loss": 3.424622412851698,
      "tokens_seen": 2258960384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031865596790371116,
      "loss": 1.3511,
      "theoretical_loss": 3.4246143932726145,
      "tokens_seen": 2259025920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186359077231695,
      "loss": 1.3493,
      "theoretical_loss": 3.424606373991322,
      "tokens_seen": 2259091456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186158475426279,
      "loss": 1.3013,
      "theoretical_loss": 3.4245983550078023,
      "tokens_seen": 2259156992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031859578736208625,
      "loss": 1.2058,
      "theoretical_loss": 3.424590336322034,
      "tokens_seen": 2259222528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031857572718154467,
      "loss": 1.3551,
      "theoretical_loss": 3.4245823179339983,
      "tokens_seen": 2259288064
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031855566700100303,
      "loss": 1.3394,
      "theoretical_loss": 3.424574299843675,
      "tokens_seen": 2259353600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003185356068204614,
      "loss": 1.3188,
      "theoretical_loss": 3.4245662820510443,
      "tokens_seen": 2259419136
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 822867,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.5087814331054688,
      "objective/train/theoretical_loss": 3.4245582645560875,
      "objective/train/tokens_used": 629943776,
      "theoretical_loss": 3.4245582645560875,
      "tokens_seen": 2259484672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031851554663991975,
      "loss": 1.3787,
      "theoretical_loss": 3.4245582645560875,
      "tokens_seen": 2259484672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031849548645937817,
      "loss": 1.2638,
      "theoretical_loss": 3.424550247358784,
      "tokens_seen": 2259550208
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031847542627883653,
      "loss": 1.2854,
      "theoretical_loss": 3.4245422304591147,
      "tokens_seen": 2259615744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184553660982949,
      "loss": 1.2466,
      "theoretical_loss": 3.4245342138570596,
      "tokens_seen": 2259681280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031843530591775326,
      "loss": 1.3687,
      "theoretical_loss": 3.424526197552599,
      "tokens_seen": 2259746816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184152457372116,
      "loss": 1.3883,
      "theoretical_loss": 3.424518181545713,
      "tokens_seen": 2259812352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031839518555667004,
      "loss": 1.2954,
      "theoretical_loss": 3.4245101658363826,
      "tokens_seen": 2259877888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183751253761284,
      "loss": 1.2757,
      "theoretical_loss": 3.4245021504245874,
      "tokens_seen": 2259943424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031835506519558676,
      "loss": 1.2809,
      "theoretical_loss": 3.424494135310308,
      "tokens_seen": 2260008960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183350050150451,
      "loss": 1.2842,
      "theoretical_loss": 3.4244861204935253,
      "tokens_seen": 2260074496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031831494483450354,
      "loss": 1.3217,
      "theoretical_loss": 3.4244781059742184,
      "tokens_seen": 2260140032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182948846539619,
      "loss": 1.3503,
      "theoretical_loss": 3.424470091752369,
      "tokens_seen": 2260205568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031827482447342026,
      "loss": 1.1934,
      "theoretical_loss": 3.4244620778279566,
      "tokens_seen": 2260271104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182547642928786,
      "loss": 1.2986,
      "theoretical_loss": 3.4244540642009613,
      "tokens_seen": 2260336640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318234704112337,
      "loss": 1.3644,
      "theoretical_loss": 3.4244460508713646,
      "tokens_seen": 2260402176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182146439317954,
      "loss": 1.3134,
      "theoretical_loss": 3.4244380378391455,
      "tokens_seen": 2260467712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031819458375125377,
      "loss": 1.266,
      "theoretical_loss": 3.424430025104285,
      "tokens_seen": 2260533248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031817452357071213,
      "loss": 1.3428,
      "theoretical_loss": 3.4244220126667635,
      "tokens_seen": 2260598784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003181544633901705,
      "loss": 1.299,
      "theoretical_loss": 3.4244140005265615,
      "tokens_seen": 2260664320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003181344032096289,
      "loss": 1.379,
      "theoretical_loss": 3.4244059886836586,
      "tokens_seen": 2260729856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031811434302908727,
      "loss": 1.4492,
      "theoretical_loss": 3.424397977138036,
      "tokens_seen": 2260795392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031809428284854563,
      "loss": 1.446,
      "theoretical_loss": 3.4243899658896737,
      "tokens_seen": 2260860928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318074222668004,
      "loss": 1.3207,
      "theoretical_loss": 3.4243819549385517,
      "tokens_seen": 2260926464
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003180541624874624,
      "loss": 1.3248,
      "theoretical_loss": 3.424373944284651,
      "tokens_seen": 2260992000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031803410230692077,
      "loss": 1.2993,
      "theoretical_loss": 3.4243659339279517,
      "tokens_seen": 2261057536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031801404212637913,
      "loss": 1.3959,
      "theoretical_loss": 3.424357923868434,
      "tokens_seen": 2261123072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003179939819458375,
      "loss": 1.332,
      "theoretical_loss": 3.4243499141060783,
      "tokens_seen": 2261188608
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031797392176529586,
      "loss": 1.2128,
      "theoretical_loss": 3.4243419046408654,
      "tokens_seen": 2261254144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003179538615847543,
      "loss": 1.3677,
      "theoretical_loss": 3.424333895472775,
      "tokens_seen": 2261319680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031793380140421264,
      "loss": 1.3694,
      "theoretical_loss": 3.4243258866017876,
      "tokens_seen": 2261385216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317913741223671,
      "loss": 1.2748,
      "theoretical_loss": 3.4243178780278836,
      "tokens_seen": 2261450752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031789368104312936,
      "loss": 1.3419,
      "theoretical_loss": 3.424309869751044,
      "tokens_seen": 2261516288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178736208625878,
      "loss": 1.3376,
      "theoretical_loss": 3.424301861771249,
      "tokens_seen": 2261581824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031785356068204614,
      "loss": 1.3809,
      "theoretical_loss": 3.4242938540884778,
      "tokens_seen": 2261647360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178335005015045,
      "loss": 1.3321,
      "theoretical_loss": 3.424285846702712,
      "tokens_seen": 2261712896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031781344032096287,
      "loss": 1.4432,
      "theoretical_loss": 3.4242778396139317,
      "tokens_seen": 2261778432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031779338014042123,
      "loss": 1.3648,
      "theoretical_loss": 3.4242698328221173,
      "tokens_seen": 2261843968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031777331995987964,
      "loss": 1.404,
      "theoretical_loss": 3.424261826327249,
      "tokens_seen": 2261909504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317753259779338,
      "loss": 1.2429,
      "theoretical_loss": 3.424253820129307,
      "tokens_seen": 2261975040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031773319959879637,
      "loss": 1.3101,
      "theoretical_loss": 3.4242458142282723,
      "tokens_seen": 2262040576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031771313941825473,
      "loss": 1.1376,
      "theoretical_loss": 3.4242378086241247,
      "tokens_seen": 2262106112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031769307923771315,
      "loss": 1.4223,
      "theoretical_loss": 3.424229803316845,
      "tokens_seen": 2262171648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176730190571715,
      "loss": 1.2826,
      "theoretical_loss": 3.424221798306413,
      "tokens_seen": 2262237184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031765295887662987,
      "loss": 1.3168,
      "theoretical_loss": 3.42421379359281,
      "tokens_seen": 2262302720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031763289869608823,
      "loss": 1.3673,
      "theoretical_loss": 3.424205789176016,
      "tokens_seen": 2262368256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176128385155466,
      "loss": 1.4524,
      "theoretical_loss": 3.424197785056011,
      "tokens_seen": 2262433792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317592778335005,
      "loss": 1.3922,
      "theoretical_loss": 3.4241897812327755,
      "tokens_seen": 2262499328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175727181544634,
      "loss": 1.3567,
      "theoretical_loss": 3.4241817777062904,
      "tokens_seen": 2262564864
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175526579739218,
      "loss": 1.3194,
      "theoretical_loss": 3.424173774476536,
      "tokens_seen": 2262630400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031753259779338015,
      "loss": 1.3001,
      "theoretical_loss": 3.4241657715434917,
      "tokens_seen": 2262695936
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 824193,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4341137409210205,
      "objective/train/theoretical_loss": 3.4241577689071394,
      "objective/train/tokens_used": 633220576,
      "theoretical_loss": 3.4241577689071394,
      "tokens_seen": 2262761472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031751253761283857,
      "loss": 1.3582,
      "theoretical_loss": 3.4241577689071394,
      "tokens_seen": 2262761472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031749247743229693,
      "loss": 1.3299,
      "theoretical_loss": 3.4241497665674587,
      "tokens_seen": 2262827008
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003174724172517553,
      "loss": 1.351,
      "theoretical_loss": 3.42414176452443,
      "tokens_seen": 2262892544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031745235707121366,
      "loss": 1.2903,
      "theoretical_loss": 3.424133762778034,
      "tokens_seen": 2262958080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317432296890672,
      "loss": 1.3457,
      "theoretical_loss": 3.424125761328251,
      "tokens_seen": 2263023616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031741223671013044,
      "loss": 1.2391,
      "theoretical_loss": 3.4241177601750614,
      "tokens_seen": 2263089152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173921765295888,
      "loss": 1.2918,
      "theoretical_loss": 3.4241097593184455,
      "tokens_seen": 2263154688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031737211634904716,
      "loss": 1.3542,
      "theoretical_loss": 3.424101758758384,
      "tokens_seen": 2263220224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173520561685055,
      "loss": 1.3067,
      "theoretical_loss": 3.424093758494857,
      "tokens_seen": 2263285760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031733199598796394,
      "loss": 1.345,
      "theoretical_loss": 3.424085758527845,
      "tokens_seen": 2263351296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173119358074223,
      "loss": 1.3262,
      "theoretical_loss": 3.4240777588573286,
      "tokens_seen": 2263416832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031729187562688066,
      "loss": 1.3295,
      "theoretical_loss": 3.424069759483288,
      "tokens_seen": 2263482368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317271815446339,
      "loss": 1.3212,
      "theoretical_loss": 3.4240617604057038,
      "tokens_seen": 2263547904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172517552657974,
      "loss": 1.4052,
      "theoretical_loss": 3.424053761624556,
      "tokens_seen": 2263613440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172316950852558,
      "loss": 1.3852,
      "theoretical_loss": 3.424045763139826,
      "tokens_seen": 2263678976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031721163490471417,
      "loss": 1.3656,
      "theoretical_loss": 3.4240377649514935,
      "tokens_seen": 2263744512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031719157472417253,
      "loss": 1.3539,
      "theoretical_loss": 3.4240297670595385,
      "tokens_seen": 2263810048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171715145436309,
      "loss": 1.3526,
      "theoretical_loss": 3.4240217694639425,
      "tokens_seen": 2263875584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171514543630893,
      "loss": 1.3976,
      "theoretical_loss": 3.4240137721646855,
      "tokens_seen": 2263941120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031713139418254767,
      "loss": 1.2782,
      "theoretical_loss": 3.424005775161748,
      "tokens_seen": 2264006656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031711133400200603,
      "loss": 1.4166,
      "theoretical_loss": 3.42399777845511,
      "tokens_seen": 2264072192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170912738214644,
      "loss": 1.384,
      "theoretical_loss": 3.4239897820447522,
      "tokens_seen": 2264137728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170712136409228,
      "loss": 1.2788,
      "theoretical_loss": 3.4239817859306556,
      "tokens_seen": 2264203264
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170511534603812,
      "loss": 1.3884,
      "theoretical_loss": 3.4239737901128,
      "tokens_seen": 2264268800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031703109327983954,
      "loss": 1.3159,
      "theoretical_loss": 3.423965794591166,
      "tokens_seen": 2264334336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170110330992979,
      "loss": 1.2609,
      "theoretical_loss": 3.423957799365734,
      "tokens_seen": 2264399872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031699097291875626,
      "loss": 1.2845,
      "theoretical_loss": 3.4239498044364844,
      "tokens_seen": 2264465408
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003169709127382147,
      "loss": 1.3472,
      "theoretical_loss": 3.423941809803398,
      "tokens_seen": 2264530944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031695085255767304,
      "loss": 1.2235,
      "theoretical_loss": 3.423933815466455,
      "tokens_seen": 2264596480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003169307923771314,
      "loss": 1.3438,
      "theoretical_loss": 3.423925821425636,
      "tokens_seen": 2264662016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031691073219658976,
      "loss": 1.3512,
      "theoretical_loss": 3.4239178276809215,
      "tokens_seen": 2264727552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003168906720160482,
      "loss": 1.3445,
      "theoretical_loss": 3.4239098342322913,
      "tokens_seen": 2264793088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031687061183550654,
      "loss": 1.3948,
      "theoretical_loss": 3.4239018410797266,
      "tokens_seen": 2264858624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003168505516549649,
      "loss": 1.3745,
      "theoretical_loss": 3.4238938482232077,
      "tokens_seen": 2264924160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031683049147442327,
      "loss": 1.367,
      "theoretical_loss": 3.423885855662715,
      "tokens_seen": 2264989696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031681043129388163,
      "loss": 1.3853,
      "theoretical_loss": 3.4238778633982294,
      "tokens_seen": 2265055232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031679037111334004,
      "loss": 1.3939,
      "theoretical_loss": 3.423869871429731,
      "tokens_seen": 2265120768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003167703109327984,
      "loss": 1.3539,
      "theoretical_loss": 3.4238618797572,
      "tokens_seen": 2265186304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031675025075225677,
      "loss": 1.3538,
      "theoretical_loss": 3.4238538883806173,
      "tokens_seen": 2265251840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031673019057171513,
      "loss": 1.3526,
      "theoretical_loss": 3.423845897299963,
      "tokens_seen": 2265317376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031671013039117355,
      "loss": 1.3548,
      "theoretical_loss": 3.423837906515218,
      "tokens_seen": 2265382912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166900702106319,
      "loss": 1.2499,
      "theoretical_loss": 3.423829916026362,
      "tokens_seen": 2265448448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031667001003009027,
      "loss": 1.3064,
      "theoretical_loss": 3.423821925833377,
      "tokens_seen": 2265513984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031664994984954863,
      "loss": 1.2685,
      "theoretical_loss": 3.423813935936242,
      "tokens_seen": 2265579520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316629889669007,
      "loss": 1.3814,
      "theoretical_loss": 3.4238059463349377,
      "tokens_seen": 2265645056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166098294884654,
      "loss": 1.2939,
      "theoretical_loss": 3.4237979570294454,
      "tokens_seen": 2265710592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165897693079238,
      "loss": 1.2421,
      "theoretical_loss": 3.423789968019745,
      "tokens_seen": 2265776128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031656970912738214,
      "loss": 1.3809,
      "theoretical_loss": 3.4237819793058173,
      "tokens_seen": 2265841664
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165496489468405,
      "loss": 1.4027,
      "theoretical_loss": 3.4237739908876423,
      "tokens_seen": 2265907200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165295887662989,
      "loss": 1.2975,
      "theoretical_loss": 3.423766002765201,
      "tokens_seen": 2265972736
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 825438,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3243449926376343,
      "objective/train/theoretical_loss": 3.423758014938474,
      "objective/train/tokens_used": 636497376,
      "theoretical_loss": 3.423758014938474,
      "tokens_seen": 2266038272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165095285857573,
      "loss": 1.2886,
      "theoretical_loss": 3.423758014938474,
      "tokens_seen": 2266038272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031648946840521564,
      "loss": 1.3231,
      "theoretical_loss": 3.423750027407441,
      "tokens_seen": 2266103808
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316469408224674,
      "loss": 1.2709,
      "theoretical_loss": 3.4237420401720833,
      "tokens_seen": 2266169344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031644934804413237,
      "loss": 1.2743,
      "theoretical_loss": 3.423734053232381,
      "tokens_seen": 2266234880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003164292878635908,
      "loss": 1.2997,
      "theoretical_loss": 3.4237260665883147,
      "tokens_seen": 2266300416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031640922768304914,
      "loss": 1.5299,
      "theoretical_loss": 3.423718080239865,
      "tokens_seen": 2266365952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163891675025075,
      "loss": 1.349,
      "theoretical_loss": 3.423710094187012,
      "tokens_seen": 2266431488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031636910732196587,
      "loss": 1.2911,
      "theoretical_loss": 3.423702108429737,
      "tokens_seen": 2266497024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163490471414243,
      "loss": 1.3395,
      "theoretical_loss": 3.42369412296802,
      "tokens_seen": 2266562560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031632898696088265,
      "loss": 1.3541,
      "theoretical_loss": 3.4236861378018415,
      "tokens_seen": 2266628096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316308926780341,
      "loss": 1.4256,
      "theoretical_loss": 3.423678152931182,
      "tokens_seen": 2266693632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031628886659979937,
      "loss": 1.4446,
      "theoretical_loss": 3.4236701683560224,
      "tokens_seen": 2266759168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031626880641925773,
      "loss": 1.309,
      "theoretical_loss": 3.423662184076343,
      "tokens_seen": 2266824704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031624874623871615,
      "loss": 1.3731,
      "theoretical_loss": 3.4236542000921237,
      "tokens_seen": 2266890240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162286860581745,
      "loss": 1.2647,
      "theoretical_loss": 3.4236462164033457,
      "tokens_seen": 2266955776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162086258776329,
      "loss": 1.3859,
      "theoretical_loss": 3.42363823300999,
      "tokens_seen": 2267021312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031618856569709124,
      "loss": 1.3212,
      "theoretical_loss": 3.423630249912036,
      "tokens_seen": 2267086848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031616850551654965,
      "loss": 1.3167,
      "theoretical_loss": 3.423622267109465,
      "tokens_seen": 2267152384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316148445336008,
      "loss": 1.3949,
      "theoretical_loss": 3.423614284602257,
      "tokens_seen": 2267217920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003161283851554664,
      "loss": 1.293,
      "theoretical_loss": 3.423606302390393,
      "tokens_seen": 2267283456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031610832497492474,
      "loss": 1.3901,
      "theoretical_loss": 3.423598320473854,
      "tokens_seen": 2267348992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160882647943831,
      "loss": 1.3159,
      "theoretical_loss": 3.423590338852619,
      "tokens_seen": 2267414528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160682046138415,
      "loss": 1.2792,
      "theoretical_loss": 3.42358235752667,
      "tokens_seen": 2267480064
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031604814443329994,
      "loss": 1.2882,
      "theoretical_loss": 3.4235743764959867,
      "tokens_seen": 2267545600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160280842527583,
      "loss": 1.3013,
      "theoretical_loss": 3.42356639576055,
      "tokens_seen": 2267611136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031600802407221666,
      "loss": 1.3033,
      "theoretical_loss": 3.4235584153203407,
      "tokens_seen": 2267676672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003159879638916751,
      "loss": 1.2979,
      "theoretical_loss": 3.423550435175339,
      "tokens_seen": 2267742208
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031596790371113344,
      "loss": 1.2566,
      "theoretical_loss": 3.423542455325525,
      "tokens_seen": 2267807744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003159478435305918,
      "loss": 1.3325,
      "theoretical_loss": 3.42353447577088,
      "tokens_seen": 2267873280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031592778335005016,
      "loss": 1.2413,
      "theoretical_loss": 3.4235264965113843,
      "tokens_seen": 2267938816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003159077231695086,
      "loss": 1.4014,
      "theoretical_loss": 3.4235185175470186,
      "tokens_seen": 2268004352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031588766298896694,
      "loss": 1.4239,
      "theoretical_loss": 3.423510538877763,
      "tokens_seen": 2268069888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003158676028084253,
      "loss": 1.416,
      "theoretical_loss": 3.4235025605035987,
      "tokens_seen": 2268135424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031584754262788367,
      "loss": 1.4103,
      "theoretical_loss": 3.4234945824245058,
      "tokens_seen": 2268200960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031582748244734203,
      "loss": 1.2994,
      "theoretical_loss": 3.423486604640465,
      "tokens_seen": 2268266496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031580742226680044,
      "loss": 1.255,
      "theoretical_loss": 3.4234786271514563,
      "tokens_seen": 2268332032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003157873620862588,
      "loss": 1.362,
      "theoretical_loss": 3.423470649957461,
      "tokens_seen": 2268397568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031576730190571717,
      "loss": 1.3785,
      "theoretical_loss": 3.4234626730584603,
      "tokens_seen": 2268463104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031574724172517553,
      "loss": 1.2947,
      "theoretical_loss": 3.4234546964544332,
      "tokens_seen": 2268528640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031572718154463395,
      "loss": 1.2728,
      "theoretical_loss": 3.4234467201453613,
      "tokens_seen": 2268594176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003157071213640923,
      "loss": 1.2468,
      "theoretical_loss": 3.4234387441312246,
      "tokens_seen": 2268659712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031568706118355067,
      "loss": 1.4021,
      "theoretical_loss": 3.423430768412004,
      "tokens_seen": 2268725248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031566700100300903,
      "loss": 1.273,
      "theoretical_loss": 3.4234227929876804,
      "tokens_seen": 2268790784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003156469408224674,
      "loss": 1.3642,
      "theoretical_loss": 3.4234148178582338,
      "tokens_seen": 2268856320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003156268806419258,
      "loss": 1.3198,
      "theoretical_loss": 3.4234068430236446,
      "tokens_seen": 2268921856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003156068204613842,
      "loss": 1.3041,
      "theoretical_loss": 3.4233988684838943,
      "tokens_seen": 2268987392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031558676028084254,
      "loss": 1.3536,
      "theoretical_loss": 3.423390894238963,
      "tokens_seen": 2269052928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155667001003009,
      "loss": 1.2933,
      "theoretical_loss": 3.423382920288831,
      "tokens_seen": 2269118464
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155466399197593,
      "loss": 1.1906,
      "theoretical_loss": 3.423374946633479,
      "tokens_seen": 2269184000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155265797392177,
      "loss": 1.2789,
      "theoretical_loss": 3.423366973272888,
      "tokens_seen": 2269249536
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 826474,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2791855335235596,
      "objective/train/theoretical_loss": 3.4233590002070384,
      "objective/train/tokens_used": 639774176,
      "theoretical_loss": 3.4233590002070384,
      "tokens_seen": 2269315072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031550651955867604,
      "loss": 1.3293,
      "theoretical_loss": 3.4233590002070384,
      "tokens_seen": 2269315072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154864593781344,
      "loss": 1.3329,
      "theoretical_loss": 3.4233510274359107,
      "tokens_seen": 2269380608
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031546639919759277,
      "loss": 1.3198,
      "theoretical_loss": 3.4233430549594854,
      "tokens_seen": 2269446144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154463390170512,
      "loss": 1.2804,
      "theoretical_loss": 3.423335082777743,
      "tokens_seen": 2269511680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031542627883650954,
      "loss": 1.3649,
      "theoretical_loss": 3.4233271108906647,
      "tokens_seen": 2269577216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154062186559679,
      "loss": 1.2806,
      "theoretical_loss": 3.4233191392982305,
      "tokens_seen": 2269642752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031538615847542627,
      "loss": 1.2899,
      "theoretical_loss": 3.423311168000421,
      "tokens_seen": 2269708288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003153660982948847,
      "loss": 1.3163,
      "theoretical_loss": 3.4233031969972174,
      "tokens_seen": 2269773824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031534603811434305,
      "loss": 1.2895,
      "theoretical_loss": 3.4232952262886,
      "tokens_seen": 2269839360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003153259779338014,
      "loss": 1.3795,
      "theoretical_loss": 3.423287255874549,
      "tokens_seen": 2269904896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031530591775325977,
      "loss": 1.3685,
      "theoretical_loss": 3.423279285755045,
      "tokens_seen": 2269970432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031528585757271813,
      "loss": 1.3008,
      "theoretical_loss": 3.4232713159300694,
      "tokens_seen": 2270035968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031526579739217655,
      "loss": 1.1915,
      "theoretical_loss": 3.4232633463996023,
      "tokens_seen": 2270101504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003152457372116349,
      "loss": 1.2787,
      "theoretical_loss": 3.4232553771636245,
      "tokens_seen": 2270167040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003152256770310933,
      "loss": 1.2514,
      "theoretical_loss": 3.4232474082221165,
      "tokens_seen": 2270232576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031520561685055164,
      "loss": 1.2126,
      "theoretical_loss": 3.4232394395750587,
      "tokens_seen": 2270298112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031518555667001005,
      "loss": 1.3392,
      "theoretical_loss": 3.4232314712224317,
      "tokens_seen": 2270363648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151654964894684,
      "loss": 1.2786,
      "theoretical_loss": 3.423223503164217,
      "tokens_seen": 2270429184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151454363089268,
      "loss": 1.2944,
      "theoretical_loss": 3.423215535400394,
      "tokens_seen": 2270494720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031512537612838514,
      "loss": 1.3857,
      "theoretical_loss": 3.423207567930944,
      "tokens_seen": 2270560256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151053159478435,
      "loss": 1.2976,
      "theoretical_loss": 3.423199600755848,
      "tokens_seen": 2270625792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150852557673019,
      "loss": 1.3315,
      "theoretical_loss": 3.423191633875086,
      "tokens_seen": 2270691328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150651955867603,
      "loss": 1.3245,
      "theoretical_loss": 3.4231836672886384,
      "tokens_seen": 2270756864
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031504513540621864,
      "loss": 1.3049,
      "theoretical_loss": 3.4231757009964863,
      "tokens_seen": 2270822400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000315025075225677,
      "loss": 1.3287,
      "theoretical_loss": 3.4231677349986107,
      "tokens_seen": 2270887936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150050150451354,
      "loss": 1.1428,
      "theoretical_loss": 3.4231597692949913,
      "tokens_seen": 2270953472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149849548645938,
      "loss": 1.2547,
      "theoretical_loss": 3.4231518038856095,
      "tokens_seen": 2271019008
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031496489468405215,
      "loss": 1.3618,
      "theoretical_loss": 3.4231438387704456,
      "tokens_seen": 2271084544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149448345035105,
      "loss": 1.3367,
      "theoretical_loss": 3.42313587394948,
      "tokens_seen": 2271150080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149247743229689,
      "loss": 1.451,
      "theoretical_loss": 3.4231279094226945,
      "tokens_seen": 2271215616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149047141424273,
      "loss": 1.3541,
      "theoretical_loss": 3.4231199451900682,
      "tokens_seen": 2271281152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031488465396188565,
      "loss": 1.2798,
      "theoretical_loss": 3.4231119812515827,
      "tokens_seen": 2271346688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314864593781344,
      "loss": 1.2459,
      "theoretical_loss": 3.4231040176072187,
      "tokens_seen": 2271412224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003148445336008024,
      "loss": 1.4042,
      "theoretical_loss": 3.4230960542569564,
      "tokens_seen": 2271477760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003148244734202608,
      "loss": 1.2704,
      "theoretical_loss": 3.4230880912007766,
      "tokens_seen": 2271543296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031480441323971915,
      "loss": 1.2606,
      "theoretical_loss": 3.4230801284386594,
      "tokens_seen": 2271608832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147843530591775,
      "loss": 1.2651,
      "theoretical_loss": 3.423072165970587,
      "tokens_seen": 2271674368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147642928786359,
      "loss": 1.3158,
      "theoretical_loss": 3.4230642037965384,
      "tokens_seen": 2271739904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147442326980943,
      "loss": 1.3353,
      "theoretical_loss": 3.423056241916495,
      "tokens_seen": 2271805440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031472417251755266,
      "loss": 1.2985,
      "theoretical_loss": 3.4230482803304376,
      "tokens_seen": 2271870976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314704112337011,
      "loss": 1.3788,
      "theoretical_loss": 3.423040319038347,
      "tokens_seen": 2271936512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146840521564694,
      "loss": 1.2801,
      "theoretical_loss": 3.423032358040203,
      "tokens_seen": 2272002048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031466399197592774,
      "loss": 1.2869,
      "theoretical_loss": 3.423024397335987,
      "tokens_seen": 2272067584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031464393179538616,
      "loss": 1.3256,
      "theoretical_loss": 3.4230164369256797,
      "tokens_seen": 2272133120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146238716148445,
      "loss": 1.385,
      "theoretical_loss": 3.4230084768092612,
      "tokens_seen": 2272198656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146038114343029,
      "loss": 1.2548,
      "theoretical_loss": 3.423000516986713,
      "tokens_seen": 2272264192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031458375125376125,
      "loss": 1.3888,
      "theoretical_loss": 3.4229925574580147,
      "tokens_seen": 2272329728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145636910732197,
      "loss": 1.3687,
      "theoretical_loss": 3.422984598223148,
      "tokens_seen": 2272395264
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145436308926781,
      "loss": 1.2808,
      "theoretical_loss": 3.4229766392820933,
      "tokens_seen": 2272460800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031452357071213644,
      "loss": 1.228,
      "theoretical_loss": 3.4229686806348307,
      "tokens_seen": 2272526336
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 827716,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3444836139678955,
      "objective/train/theoretical_loss": 3.4229607222813416,
      "objective/train/tokens_used": 643050976,
      "theoretical_loss": 3.4229607222813416,
      "tokens_seen": 2272591872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145035105315948,
      "loss": 1.3083,
      "theoretical_loss": 3.4229607222813416,
      "tokens_seen": 2272591872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031448345035105317,
      "loss": 1.3633,
      "theoretical_loss": 3.4229527642216064,
      "tokens_seen": 2272657408
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144633901705116,
      "loss": 1.3269,
      "theoretical_loss": 3.4229448064556056,
      "tokens_seen": 2272722944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031444332998996994,
      "loss": 1.408,
      "theoretical_loss": 3.42293684898332,
      "tokens_seen": 2272788480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144232698094283,
      "loss": 1.2728,
      "theoretical_loss": 3.422928891804731,
      "tokens_seen": 2272854016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031440320962888667,
      "loss": 1.2904,
      "theoretical_loss": 3.4229209349198184,
      "tokens_seen": 2272919552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143831494483451,
      "loss": 1.3569,
      "theoretical_loss": 3.422912978328563,
      "tokens_seen": 2272985088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031436308926780345,
      "loss": 1.2309,
      "theoretical_loss": 3.4229050220309456,
      "tokens_seen": 2273050624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143430290872618,
      "loss": 1.3575,
      "theoretical_loss": 3.4228970660269473,
      "tokens_seen": 2273116160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031432296890672017,
      "loss": 1.3659,
      "theoretical_loss": 3.422889110316548,
      "tokens_seen": 2273181696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031430290872617853,
      "loss": 1.2539,
      "theoretical_loss": 3.422881154899729,
      "tokens_seen": 2273247232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031428284854563695,
      "loss": 1.3712,
      "theoretical_loss": 3.422873199776471,
      "tokens_seen": 2273312768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142627883650953,
      "loss": 1.2843,
      "theoretical_loss": 3.4228652449467547,
      "tokens_seen": 2273378304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142427281845537,
      "loss": 1.2933,
      "theoretical_loss": 3.4228572904105605,
      "tokens_seen": 2273443840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031422266800401204,
      "loss": 1.3838,
      "theoretical_loss": 3.4228493361678693,
      "tokens_seen": 2273509376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031420260782347045,
      "loss": 1.3896,
      "theoretical_loss": 3.4228413822186616,
      "tokens_seen": 2273574912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141825476429288,
      "loss": 1.3701,
      "theoretical_loss": 3.4228334285629187,
      "tokens_seen": 2273640448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141624874623872,
      "loss": 1.3951,
      "theoretical_loss": 3.4228254752006206,
      "tokens_seen": 2273705984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031414242728184554,
      "loss": 1.4413,
      "theoretical_loss": 3.4228175221317483,
      "tokens_seen": 2273771520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141223671013039,
      "loss": 1.3033,
      "theoretical_loss": 3.422809569356283,
      "tokens_seen": 2273837056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141023069207623,
      "loss": 1.3422,
      "theoretical_loss": 3.4228016168742044,
      "tokens_seen": 2273902592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140822467402207,
      "loss": 1.3284,
      "theoretical_loss": 3.422793664685494,
      "tokens_seen": 2273968128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031406218655967904,
      "loss": 1.3892,
      "theoretical_loss": 3.4227857127901324,
      "tokens_seen": 2274033664
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140421263791374,
      "loss": 1.2324,
      "theoretical_loss": 3.4227777611881,
      "tokens_seen": 2274099200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140220661985958,
      "loss": 1.3167,
      "theoretical_loss": 3.4227698098793784,
      "tokens_seen": 2274164736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140020060180542,
      "loss": 1.4027,
      "theoretical_loss": 3.422761858863947,
      "tokens_seen": 2274230272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031398194583751255,
      "loss": 1.3724,
      "theoretical_loss": 3.4227539081417877,
      "tokens_seen": 2274295808
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139618856569709,
      "loss": 1.376,
      "theoretical_loss": 3.4227459577128805,
      "tokens_seen": 2274361344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139418254764293,
      "loss": 1.3556,
      "theoretical_loss": 3.4227380075772067,
      "tokens_seen": 2274426880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139217652958877,
      "loss": 1.4243,
      "theoretical_loss": 3.4227300577347464,
      "tokens_seen": 2274492416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031390170511534605,
      "loss": 1.4111,
      "theoretical_loss": 3.422722108185481,
      "tokens_seen": 2274557952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138816449348044,
      "loss": 1.3285,
      "theoretical_loss": 3.4227141589293906,
      "tokens_seen": 2274623488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138615847542628,
      "loss": 1.2785,
      "theoretical_loss": 3.4227062099664565,
      "tokens_seen": 2274689024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138415245737212,
      "loss": 1.3156,
      "theoretical_loss": 3.4226982612966586,
      "tokens_seen": 2274754560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031382146439317955,
      "loss": 1.1823,
      "theoretical_loss": 3.4226903129199786,
      "tokens_seen": 2274820096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138014042126379,
      "loss": 1.275,
      "theoretical_loss": 3.422682364836397,
      "tokens_seen": 2274885632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137813440320963,
      "loss": 1.2978,
      "theoretical_loss": 3.4226744170458945,
      "tokens_seen": 2274951168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137612838515547,
      "loss": 1.3652,
      "theoretical_loss": 3.4226664695484517,
      "tokens_seen": 2275016704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031374122367101306,
      "loss": 1.3871,
      "theoretical_loss": 3.4226585223440495,
      "tokens_seen": 2275082240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137211634904714,
      "loss": 1.347,
      "theoretical_loss": 3.4226505754326686,
      "tokens_seen": 2275147776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137011033099298,
      "loss": 1.3375,
      "theoretical_loss": 3.4226426288142893,
      "tokens_seen": 2275213312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031368104312938814,
      "loss": 1.3626,
      "theoretical_loss": 3.4226346824888934,
      "tokens_seen": 2275278848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031366098294884656,
      "loss": 1.2423,
      "theoretical_loss": 3.422626736456461,
      "tokens_seen": 2275344384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136409227683049,
      "loss": 1.312,
      "theoretical_loss": 3.4226187907169723,
      "tokens_seen": 2275409920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136208625877633,
      "loss": 1.3829,
      "theoretical_loss": 3.4226108452704094,
      "tokens_seen": 2275475456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031360080240722165,
      "loss": 1.2626,
      "theoretical_loss": 3.422602900116752,
      "tokens_seen": 2275540992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031358074222668006,
      "loss": 1.3898,
      "theoretical_loss": 3.422594955255981,
      "tokens_seen": 2275606528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135606820461384,
      "loss": 1.3207,
      "theoretical_loss": 3.422587010688078,
      "tokens_seen": 2275672064
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135406218655968,
      "loss": 1.4256,
      "theoretical_loss": 3.422579066413023,
      "tokens_seen": 2275737600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031352056168505515,
      "loss": 1.3792,
      "theoretical_loss": 3.4225711224307966,
      "tokens_seen": 2275803136
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 828910,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2416036128997803,
      "objective/train/theoretical_loss": 3.42256317874138,
      "objective/train/tokens_used": 646327776,
      "theoretical_loss": 3.42256317874138,
      "tokens_seen": 2275868672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135005015045135,
      "loss": 1.3088,
      "theoretical_loss": 3.42256317874138,
      "tokens_seen": 2275868672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031348044132397193,
      "loss": 1.4286,
      "theoretical_loss": 3.422555235344754,
      "tokens_seen": 2275934208
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134603811434303,
      "loss": 1.1684,
      "theoretical_loss": 3.4225472922408993,
      "tokens_seen": 2275999744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031344032096288865,
      "loss": 1.3829,
      "theoretical_loss": 3.4225393494297967,
      "tokens_seen": 2276065280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313420260782347,
      "loss": 1.3313,
      "theoretical_loss": 3.4225314069114265,
      "tokens_seen": 2276130816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031340020060180543,
      "loss": 1.2778,
      "theoretical_loss": 3.4225234646857703,
      "tokens_seen": 2276196352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133801404212638,
      "loss": 1.378,
      "theoretical_loss": 3.4225155227528083,
      "tokens_seen": 2276261888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031336008024072216,
      "loss": 1.3645,
      "theoretical_loss": 3.4225075811125216,
      "tokens_seen": 2276327424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133400200601805,
      "loss": 1.3042,
      "theoretical_loss": 3.422499639764891,
      "tokens_seen": 2276392960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133199598796389,
      "loss": 1.3678,
      "theoretical_loss": 3.422491698709897,
      "tokens_seen": 2276458496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132998996990973,
      "loss": 1.2056,
      "theoretical_loss": 3.4224837579475205,
      "tokens_seen": 2276524032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031327983951855566,
      "loss": 1.2938,
      "theoretical_loss": 3.422475817477742,
      "tokens_seen": 2276589568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313259779338014,
      "loss": 1.4211,
      "theoretical_loss": 3.422467877300543,
      "tokens_seen": 2276655104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132397191574724,
      "loss": 1.2548,
      "theoretical_loss": 3.422459937415904,
      "tokens_seen": 2276720640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132196589769308,
      "loss": 1.3966,
      "theoretical_loss": 3.4224519978238055,
      "tokens_seen": 2276786176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031319959879638916,
      "loss": 1.3433,
      "theoretical_loss": 3.422444058524229,
      "tokens_seen": 2276851712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003131795386158475,
      "loss": 1.2742,
      "theoretical_loss": 3.4224361195171547,
      "tokens_seen": 2276917248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003131594784353059,
      "loss": 1.3375,
      "theoretical_loss": 3.422428180802563,
      "tokens_seen": 2276982784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031313941825476425,
      "loss": 1.2373,
      "theoretical_loss": 3.422420242380436,
      "tokens_seen": 2277048320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031311935807422267,
      "loss": 1.252,
      "theoretical_loss": 3.422412304250753,
      "tokens_seen": 2277113856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031309929789368103,
      "loss": 1.3981,
      "theoretical_loss": 3.422404366413496,
      "tokens_seen": 2277179392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130792377131394,
      "loss": 1.3659,
      "theoretical_loss": 3.4223964288686455,
      "tokens_seen": 2277244928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130591775325978,
      "loss": 1.3125,
      "theoretical_loss": 3.422388491616182,
      "tokens_seen": 2277310464
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130391173520562,
      "loss": 1.249,
      "theoretical_loss": 3.422380554656087,
      "tokens_seen": 2277376000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130190571715146,
      "loss": 1.413,
      "theoretical_loss": 3.42237261798834,
      "tokens_seen": 2277441536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031299899699097295,
      "loss": 1.3406,
      "theoretical_loss": 3.422364681612923,
      "tokens_seen": 2277507072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003129789368104313,
      "loss": 1.3259,
      "theoretical_loss": 3.4223567455298167,
      "tokens_seen": 2277572608
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003129588766298897,
      "loss": 1.3627,
      "theoretical_loss": 3.4223488097390016,
      "tokens_seen": 2277638144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003129388164493481,
      "loss": 1.3205,
      "theoretical_loss": 3.422340874240458,
      "tokens_seen": 2277703680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031291875626880645,
      "loss": 1.4216,
      "theoretical_loss": 3.422332939034168,
      "tokens_seen": 2277769216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128986960882648,
      "loss": 1.3796,
      "theoretical_loss": 3.4223250041201116,
      "tokens_seen": 2277834752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128786359077232,
      "loss": 1.2808,
      "theoretical_loss": 3.42231706949827,
      "tokens_seen": 2277900288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128585757271816,
      "loss": 1.3852,
      "theoretical_loss": 3.4223091351686237,
      "tokens_seen": 2277965824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031283851554663995,
      "loss": 1.3091,
      "theoretical_loss": 3.4223012011311535,
      "tokens_seen": 2278031360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128184553660983,
      "loss": 1.3835,
      "theoretical_loss": 3.4222932673858404,
      "tokens_seen": 2278096896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127983951855567,
      "loss": 1.2637,
      "theoretical_loss": 3.4222853339326655,
      "tokens_seen": 2278162432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127783350050151,
      "loss": 1.3812,
      "theoretical_loss": 3.4222774007716095,
      "tokens_seen": 2278227968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031275827482447346,
      "loss": 1.238,
      "theoretical_loss": 3.4222694679026526,
      "tokens_seen": 2278293504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127382146439318,
      "loss": 1.2695,
      "theoretical_loss": 3.4222615353257764,
      "tokens_seen": 2278359040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127181544633902,
      "loss": 1.3677,
      "theoretical_loss": 3.422253603040962,
      "tokens_seen": 2278424576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031269809428284854,
      "loss": 1.3223,
      "theoretical_loss": 3.422245671048189,
      "tokens_seen": 2278490112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031267803410230696,
      "loss": 1.4013,
      "theoretical_loss": 3.4222377393474392,
      "tokens_seen": 2278555648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003126579739217653,
      "loss": 1.4126,
      "theoretical_loss": 3.422229807938694,
      "tokens_seen": 2278621184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003126379137412237,
      "loss": 1.4429,
      "theoretical_loss": 3.422221876821933,
      "tokens_seen": 2278686720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031261785356068205,
      "loss": 1.2284,
      "theoretical_loss": 3.422213945997137,
      "tokens_seen": 2278752256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031259779338014046,
      "loss": 1.2272,
      "theoretical_loss": 3.422206015464288,
      "tokens_seen": 2278817792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125777331995988,
      "loss": 1.3482,
      "theoretical_loss": 3.422198085223366,
      "tokens_seen": 2278883328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125576730190572,
      "loss": 1.2693,
      "theoretical_loss": 3.422190155274352,
      "tokens_seen": 2278948864
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031253761283851555,
      "loss": 1.2568,
      "theoretical_loss": 3.422182225617228,
      "tokens_seen": 2279014400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125175526579739,
      "loss": 1.3014,
      "theoretical_loss": 3.422174296251973,
      "tokens_seen": 2279079936
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 830070,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4052246809005737,
      "objective/train/theoretical_loss": 3.4221663671785687,
      "objective/train/tokens_used": 649604576,
      "theoretical_loss": 3.4221663671785687,
      "tokens_seen": 2279145472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031249749247743233,
      "loss": 1.348,
      "theoretical_loss": 3.4221663671785687,
      "tokens_seen": 2279145472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003124774322968907,
      "loss": 1.399,
      "theoretical_loss": 3.4221584383969965,
      "tokens_seen": 2279211008
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031245737211634905,
      "loss": 1.2513,
      "theoretical_loss": 3.422150509907236,
      "tokens_seen": 2279276544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003124373119358074,
      "loss": 1.2671,
      "theoretical_loss": 3.422142581709269,
      "tokens_seen": 2279342080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031241725175526583,
      "loss": 1.2308,
      "theoretical_loss": 3.422134653803077,
      "tokens_seen": 2279407616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123971915747242,
      "loss": 1.2963,
      "theoretical_loss": 3.4221267261886394,
      "tokens_seen": 2279473152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031237713139418256,
      "loss": 1.4848,
      "theoretical_loss": 3.422118798865938,
      "tokens_seen": 2279538688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123570712136409,
      "loss": 1.3358,
      "theoretical_loss": 3.422110871834953,
      "tokens_seen": 2279604224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123370110330993,
      "loss": 1.3508,
      "theoretical_loss": 3.422102945095666,
      "tokens_seen": 2279669760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123169508525577,
      "loss": 1.2728,
      "theoretical_loss": 3.422095018648058,
      "tokens_seen": 2279735296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031229689067201606,
      "loss": 1.3979,
      "theoretical_loss": 3.4220870924921085,
      "tokens_seen": 2279800832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122768304914744,
      "loss": 1.4036,
      "theoretical_loss": 3.4220791666278,
      "tokens_seen": 2279866368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122567703109328,
      "loss": 1.3083,
      "theoretical_loss": 3.4220712410551126,
      "tokens_seen": 2279931904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122367101303912,
      "loss": 1.3162,
      "theoretical_loss": 3.4220633157740274,
      "tokens_seen": 2279997440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031221664994984956,
      "loss": 1.2242,
      "theoretical_loss": 3.422055390784525,
      "tokens_seen": 2280062976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121965897693079,
      "loss": 1.3682,
      "theoretical_loss": 3.422047466086587,
      "tokens_seen": 2280128512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121765295887663,
      "loss": 1.2556,
      "theoretical_loss": 3.4220395416801934,
      "tokens_seen": 2280194048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031215646940822465,
      "loss": 1.2739,
      "theoretical_loss": 3.4220316175653256,
      "tokens_seen": 2280259584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031213640922768307,
      "loss": 1.3798,
      "theoretical_loss": 3.4220236937419646,
      "tokens_seen": 2280325120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031211634904714143,
      "loss": 1.2226,
      "theoretical_loss": 3.422015770210091,
      "tokens_seen": 2280390656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003120962888665998,
      "loss": 1.3171,
      "theoretical_loss": 3.4220078469696853,
      "tokens_seen": 2280456192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031207622868605815,
      "loss": 1.4636,
      "theoretical_loss": 3.4219999240207297,
      "tokens_seen": 2280521728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031205616850551657,
      "loss": 1.3077,
      "theoretical_loss": 3.4219920013632037,
      "tokens_seen": 2280587264
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031203610832497493,
      "loss": 1.3283,
      "theoretical_loss": 3.421984078997089,
      "tokens_seen": 2280652800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003120160481444333,
      "loss": 1.262,
      "theoretical_loss": 3.421976156922366,
      "tokens_seen": 2280718336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031199598796389166,
      "loss": 1.349,
      "theoretical_loss": 3.4219682351390164,
      "tokens_seen": 2280783872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031197592778335,
      "loss": 1.3937,
      "theoretical_loss": 3.4219603136470207,
      "tokens_seen": 2280849408
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031195586760280843,
      "loss": 1.3059,
      "theoretical_loss": 3.4219523924463595,
      "tokens_seen": 2280914944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003119358074222668,
      "loss": 1.2466,
      "theoretical_loss": 3.4219444715370138,
      "tokens_seen": 2280980480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031191574724172516,
      "loss": 1.3576,
      "theoretical_loss": 3.4219365509189648,
      "tokens_seen": 2281046016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118956870611835,
      "loss": 1.3172,
      "theoretical_loss": 3.4219286305921934,
      "tokens_seen": 2281111552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031187562688064194,
      "loss": 1.3591,
      "theoretical_loss": 3.4219207105566802,
      "tokens_seen": 2281177088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118555667001003,
      "loss": 1.4031,
      "theoretical_loss": 3.421912790812407,
      "tokens_seen": 2281242624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031183550651955866,
      "loss": 1.2355,
      "theoretical_loss": 3.421904871359353,
      "tokens_seen": 2281308160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000311815446339017,
      "loss": 1.3552,
      "theoretical_loss": 3.421896952197501,
      "tokens_seen": 2281373696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117953861584754,
      "loss": 1.2175,
      "theoretical_loss": 3.4218890333268313,
      "tokens_seen": 2281439232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117753259779338,
      "loss": 1.2584,
      "theoretical_loss": 3.421881114747324,
      "tokens_seen": 2281504768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031175526579739217,
      "loss": 1.2597,
      "theoretical_loss": 3.421873196458961,
      "tokens_seen": 2281570304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031173520561685053,
      "loss": 1.3005,
      "theoretical_loss": 3.421865278461723,
      "tokens_seen": 2281635840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117151454363089,
      "loss": 1.3096,
      "theoretical_loss": 3.4218573607555904,
      "tokens_seen": 2281701376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116950852557673,
      "loss": 1.4209,
      "theoretical_loss": 3.421849443340545,
      "tokens_seen": 2281766912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031167502507522567,
      "loss": 1.3921,
      "theoretical_loss": 3.4218415262165673,
      "tokens_seen": 2281832448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031165496489468403,
      "loss": 1.3128,
      "theoretical_loss": 3.421833609383638,
      "tokens_seen": 2281897984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116349047141424,
      "loss": 1.3118,
      "theoretical_loss": 3.4218256928417388,
      "tokens_seen": 2281963520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116148445336008,
      "loss": 1.3656,
      "theoretical_loss": 3.4218177765908497,
      "tokens_seen": 2282029056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031159478435305917,
      "loss": 1.3162,
      "theoretical_loss": 3.421809860630952,
      "tokens_seen": 2282094592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031157472417251753,
      "loss": 1.3124,
      "theoretical_loss": 3.421801944962027,
      "tokens_seen": 2282160128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031155466399197595,
      "loss": 1.1769,
      "theoretical_loss": 3.4217940295840554,
      "tokens_seen": 2282225664
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003115346038114343,
      "loss": 1.2758,
      "theoretical_loss": 3.421786114497018,
      "tokens_seen": 2282291200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031151454363089273,
      "loss": 1.3223,
      "theoretical_loss": 3.4217781997008965,
      "tokens_seen": 2282356736
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 831223,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.272815465927124,
      "objective/train/theoretical_loss": 3.4217702851956706,
      "objective/train/tokens_used": 652881376,
      "theoretical_loss": 3.4217702851956706,
      "tokens_seen": 2282422272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114944834503511,
      "loss": 1.3403,
      "theoretical_loss": 3.4217702851956706,
      "tokens_seen": 2282422272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031147442326980945,
      "loss": 1.2478,
      "theoretical_loss": 3.421762370981322,
      "tokens_seen": 2282487808
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114543630892678,
      "loss": 1.3738,
      "theoretical_loss": 3.4217544570578315,
      "tokens_seen": 2282553344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031143430290872623,
      "loss": 1.2698,
      "theoretical_loss": 3.4217465434251806,
      "tokens_seen": 2282618880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114142427281846,
      "loss": 1.3156,
      "theoretical_loss": 3.4217386300833494,
      "tokens_seen": 2282684416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031139418254764296,
      "loss": 1.2899,
      "theoretical_loss": 3.421730717032319,
      "tokens_seen": 2282749952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113741223671013,
      "loss": 1.4426,
      "theoretical_loss": 3.421722804272071,
      "tokens_seen": 2282815488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113540621865597,
      "loss": 1.4258,
      "theoretical_loss": 3.421714891802586,
      "tokens_seen": 2282881024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113340020060181,
      "loss": 1.2838,
      "theoretical_loss": 3.421706979623845,
      "tokens_seen": 2282946560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031131394182547646,
      "loss": 1.3136,
      "theoretical_loss": 3.4216990677358288,
      "tokens_seen": 2283012096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112938816449348,
      "loss": 1.2838,
      "theoretical_loss": 3.421691156138518,
      "tokens_seen": 2283077632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112738214643932,
      "loss": 1.3447,
      "theoretical_loss": 3.421683244831895,
      "tokens_seen": 2283143168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112537612838516,
      "loss": 1.2075,
      "theoretical_loss": 3.4216753338159394,
      "tokens_seen": 2283208704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031123370110330996,
      "loss": 1.2258,
      "theoretical_loss": 3.4216674230906325,
      "tokens_seen": 2283274240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112136409227683,
      "loss": 1.2774,
      "theoretical_loss": 3.4216595126559555,
      "tokens_seen": 2283339776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111935807422267,
      "loss": 1.2819,
      "theoretical_loss": 3.4216516025118895,
      "tokens_seen": 2283405312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031117352056168505,
      "loss": 1.4573,
      "theoretical_loss": 3.421643692658415,
      "tokens_seen": 2283470848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031115346038114347,
      "loss": 1.3288,
      "theoretical_loss": 3.421635783095513,
      "tokens_seen": 2283536384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031113340020060183,
      "loss": 1.4551,
      "theoretical_loss": 3.4216278738231654,
      "tokens_seen": 2283601920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111133400200602,
      "loss": 1.2476,
      "theoretical_loss": 3.4216199648413523,
      "tokens_seen": 2283667456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031109327983951855,
      "loss": 1.3921,
      "theoretical_loss": 3.421612056150055,
      "tokens_seen": 2283732992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031107321965897697,
      "loss": 1.2785,
      "theoretical_loss": 3.421604147749254,
      "tokens_seen": 2283798528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031105315947843533,
      "loss": 1.3602,
      "theoretical_loss": 3.421596239638931,
      "tokens_seen": 2283864064
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003110330992978937,
      "loss": 1.4148,
      "theoretical_loss": 3.4215883318190667,
      "tokens_seen": 2283929600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031101303911735206,
      "loss": 1.316,
      "theoretical_loss": 3.421580424289642,
      "tokens_seen": 2283995136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109929789368104,
      "loss": 1.3387,
      "theoretical_loss": 3.421572517050638,
      "tokens_seen": 2284060672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031097291875626883,
      "loss": 1.2477,
      "theoretical_loss": 3.4215646101020356,
      "tokens_seen": 2284126208
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109528585757272,
      "loss": 1.2883,
      "theoretical_loss": 3.4215567034438163,
      "tokens_seen": 2284191744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031093279839518556,
      "loss": 1.257,
      "theoretical_loss": 3.4215487970759604,
      "tokens_seen": 2284257280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109127382146439,
      "loss": 1.3497,
      "theoretical_loss": 3.4215408909984495,
      "tokens_seen": 2284322816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031089267803410234,
      "loss": 1.2615,
      "theoretical_loss": 3.4215329852112637,
      "tokens_seen": 2284388352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108726178535607,
      "loss": 1.2576,
      "theoretical_loss": 3.4215250797143852,
      "tokens_seen": 2284453888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031085255767301906,
      "loss": 1.3441,
      "theoretical_loss": 3.4215171745077946,
      "tokens_seen": 2284519424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108324974924774,
      "loss": 1.3487,
      "theoretical_loss": 3.421509269591472,
      "tokens_seen": 2284584960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108124373119358,
      "loss": 1.2536,
      "theoretical_loss": 3.4215013649654,
      "tokens_seen": 2284650496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107923771313942,
      "loss": 1.281,
      "theoretical_loss": 3.4214934606295584,
      "tokens_seen": 2284716032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031077231695085257,
      "loss": 1.3197,
      "theoretical_loss": 3.4214855565839284,
      "tokens_seen": 2284781568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031075225677031093,
      "loss": 1.3433,
      "theoretical_loss": 3.421477652828491,
      "tokens_seen": 2284847104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107321965897693,
      "loss": 1.3833,
      "theoretical_loss": 3.4214697493632285,
      "tokens_seen": 2284912640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107121364092277,
      "loss": 1.3009,
      "theoretical_loss": 3.4214618461881203,
      "tokens_seen": 2284978176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031069207622868607,
      "loss": 1.2542,
      "theoretical_loss": 3.4214539433031477,
      "tokens_seen": 2285043712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031067201604814443,
      "loss": 1.3658,
      "theoretical_loss": 3.4214460407082923,
      "tokens_seen": 2285109248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003106519558676028,
      "loss": 1.3465,
      "theoretical_loss": 3.421438138403535,
      "tokens_seen": 2285174784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003106318956870612,
      "loss": 1.3107,
      "theoretical_loss": 3.4214302363888565,
      "tokens_seen": 2285240320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031061183550651957,
      "loss": 1.4092,
      "theoretical_loss": 3.4214223346642383,
      "tokens_seen": 2285305856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031059177532597793,
      "loss": 1.3314,
      "theoretical_loss": 3.421414433229661,
      "tokens_seen": 2285371392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003105717151454363,
      "loss": 1.3858,
      "theoretical_loss": 3.4214065320851055,
      "tokens_seen": 2285436928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031055165496489466,
      "loss": 1.258,
      "theoretical_loss": 3.4213986312305535,
      "tokens_seen": 2285502464
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003105315947843531,
      "loss": 1.2349,
      "theoretical_loss": 3.4213907306659856,
      "tokens_seen": 2285568000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031051153460381144,
      "loss": 1.2747,
      "theoretical_loss": 3.4213828303913827,
      "tokens_seen": 2285633536
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 832464,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2004680633544922,
      "objective/train/theoretical_loss": 3.4213749304067265,
      "objective/train/tokens_used": 656158176,
      "theoretical_loss": 3.4213749304067265,
      "tokens_seen": 2285699072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104914744232698,
      "loss": 1.2988,
      "theoretical_loss": 3.4213749304067265,
      "tokens_seen": 2285699072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031047141424272816,
      "loss": 1.3727,
      "theoretical_loss": 3.4213670307119974,
      "tokens_seen": 2285764608
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104513540621866,
      "loss": 1.2657,
      "theoretical_loss": 3.4213591313071765,
      "tokens_seen": 2285830144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031043129388164494,
      "loss": 1.3149,
      "theoretical_loss": 3.421351232192245,
      "tokens_seen": 2285895680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104112337011033,
      "loss": 1.2488,
      "theoretical_loss": 3.4213433333671843,
      "tokens_seen": 2285961216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031039117352056167,
      "loss": 1.3831,
      "theoretical_loss": 3.421335434831975,
      "tokens_seen": 2286026752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031037111334002003,
      "loss": 1.3024,
      "theoretical_loss": 3.421327536586598,
      "tokens_seen": 2286092288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031035105315947844,
      "loss": 1.2364,
      "theoretical_loss": 3.4213196386310347,
      "tokens_seen": 2286157824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003103309929789368,
      "loss": 1.2459,
      "theoretical_loss": 3.4213117409652662,
      "tokens_seen": 2286223360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031031093279839517,
      "loss": 1.4962,
      "theoretical_loss": 3.4213038435892735,
      "tokens_seen": 2286288896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031029087261785353,
      "loss": 1.2534,
      "theoretical_loss": 3.4212959465030375,
      "tokens_seen": 2286354432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031027081243731195,
      "loss": 1.3141,
      "theoretical_loss": 3.4212880497065394,
      "tokens_seen": 2286419968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003102507522567703,
      "loss": 1.3077,
      "theoretical_loss": 3.4212801531997603,
      "tokens_seen": 2286485504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031023069207622867,
      "loss": 1.3553,
      "theoretical_loss": 3.4212722569826814,
      "tokens_seen": 2286551040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031021063189568703,
      "loss": 1.2816,
      "theoretical_loss": 3.421264361055283,
      "tokens_seen": 2286616576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101905717151454,
      "loss": 1.3052,
      "theoretical_loss": 3.4212564654175472,
      "tokens_seen": 2286682112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101705115346038,
      "loss": 1.312,
      "theoretical_loss": 3.421248570069454,
      "tokens_seen": 2286747648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101504513540622,
      "loss": 1.4145,
      "theoretical_loss": 3.421240675010986,
      "tokens_seen": 2286813184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031013039117352054,
      "loss": 1.3362,
      "theoretical_loss": 3.4212327802421227,
      "tokens_seen": 2286878720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101103309929789,
      "loss": 1.4111,
      "theoretical_loss": 3.421224885762846,
      "tokens_seen": 2286944256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100902708124373,
      "loss": 1.3658,
      "theoretical_loss": 3.4212169915731367,
      "tokens_seen": 2287009792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100702106318957,
      "loss": 1.3521,
      "theoretical_loss": 3.4212090976729765,
      "tokens_seen": 2287075328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100501504513541,
      "loss": 1.3366,
      "theoretical_loss": 3.4212012040623456,
      "tokens_seen": 2287140864
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031003009027081246,
      "loss": 1.2581,
      "theoretical_loss": 3.4211933107412253,
      "tokens_seen": 2287206400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100100300902708,
      "loss": 1.3943,
      "theoretical_loss": 3.421185417709597,
      "tokens_seen": 2287271936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030998996990972924,
      "loss": 1.4666,
      "theoretical_loss": 3.421177524967442,
      "tokens_seen": 2287337472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003099699097291876,
      "loss": 1.275,
      "theoretical_loss": 3.4211696325147405,
      "tokens_seen": 2287403008
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030994984954864596,
      "loss": 1.2751,
      "theoretical_loss": 3.4211617403514745,
      "tokens_seen": 2287468544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003099297893681043,
      "loss": 1.2505,
      "theoretical_loss": 3.4211538484776245,
      "tokens_seen": 2287534080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030990972918756274,
      "loss": 1.3728,
      "theoretical_loss": 3.421145956893172,
      "tokens_seen": 2287599616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003098896690070211,
      "loss": 1.2774,
      "theoretical_loss": 3.4211380655980976,
      "tokens_seen": 2287665152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030986960882647946,
      "loss": 1.3289,
      "theoretical_loss": 3.421130174592383,
      "tokens_seen": 2287730688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003098495486459378,
      "loss": 1.4023,
      "theoretical_loss": 3.421122283876009,
      "tokens_seen": 2287796224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003098294884653962,
      "loss": 1.4101,
      "theoretical_loss": 3.4211143934489567,
      "tokens_seen": 2287861760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003098094282848546,
      "loss": 1.3196,
      "theoretical_loss": 3.421106503311207,
      "tokens_seen": 2287927296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030978936810431297,
      "loss": 1.4507,
      "theoretical_loss": 3.4210986134627417,
      "tokens_seen": 2287992832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030976930792377133,
      "loss": 1.3148,
      "theoretical_loss": 3.421090723903541,
      "tokens_seen": 2288058368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097492477432297,
      "loss": 1.3067,
      "theoretical_loss": 3.4210828346335864,
      "tokens_seen": 2288123904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097291875626881,
      "loss": 1.436,
      "theoretical_loss": 3.4210749456528595,
      "tokens_seen": 2288189440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030970912738214647,
      "loss": 1.1947,
      "theoretical_loss": 3.421067056961341,
      "tokens_seen": 2288254976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030968906720160483,
      "loss": 1.3105,
      "theoretical_loss": 3.4210591685590113,
      "tokens_seen": 2288320512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096690070210632,
      "loss": 1.3058,
      "theoretical_loss": 3.4210512804458526,
      "tokens_seen": 2288386048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096489468405216,
      "loss": 1.4547,
      "theoretical_loss": 3.4210433926218458,
      "tokens_seen": 2288451584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030962888665997997,
      "loss": 1.2596,
      "theoretical_loss": 3.4210355050869716,
      "tokens_seen": 2288517120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030960882647943833,
      "loss": 1.3046,
      "theoretical_loss": 3.4210276178412116,
      "tokens_seen": 2288582656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095887662988967,
      "loss": 1.3757,
      "theoretical_loss": 3.421019730884546,
      "tokens_seen": 2288648192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030956870611835506,
      "loss": 1.295,
      "theoretical_loss": 3.4210118442169573,
      "tokens_seen": 2288713728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095486459378135,
      "loss": 1.2446,
      "theoretical_loss": 3.4210039578384257,
      "tokens_seen": 2288779264
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030952858575727184,
      "loss": 1.2992,
      "theoretical_loss": 3.4209960717489327,
      "tokens_seen": 2288844800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095085255767302,
      "loss": 1.2972,
      "theoretical_loss": 3.420988185948459,
      "tokens_seen": 2288910336
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 833545,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3949838876724243,
      "objective/train/theoretical_loss": 3.4209803004369865,
      "objective/train/tokens_used": 659434976,
      "theoretical_loss": 3.4209803004369865,
      "tokens_seen": 2288975872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030948846539618856,
      "loss": 1.3159,
      "theoretical_loss": 3.4209803004369865,
      "tokens_seen": 2288975872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000309468405215647,
      "loss": 1.3452,
      "theoretical_loss": 3.4209724152144956,
      "tokens_seen": 2289041408
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030944834503510534,
      "loss": 1.3737,
      "theoretical_loss": 3.420964530280968,
      "tokens_seen": 2289106944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094282848545637,
      "loss": 1.2262,
      "theoretical_loss": 3.4209566456363842,
      "tokens_seen": 2289172480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030940822467402207,
      "loss": 1.3255,
      "theoretical_loss": 3.4209487612807257,
      "tokens_seen": 2289238016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030938816449348043,
      "loss": 1.2949,
      "theoretical_loss": 3.4209408772139738,
      "tokens_seen": 2289303552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030936810431293884,
      "loss": 1.3541,
      "theoretical_loss": 3.4209329934361095,
      "tokens_seen": 2289369088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003093480441323972,
      "loss": 1.3476,
      "theoretical_loss": 3.4209251099471136,
      "tokens_seen": 2289434624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030932798395185557,
      "loss": 1.4121,
      "theoretical_loss": 3.4209172267469676,
      "tokens_seen": 2289500160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030930792377131393,
      "loss": 1.347,
      "theoretical_loss": 3.420909343835653,
      "tokens_seen": 2289565696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030928786359077235,
      "loss": 1.2828,
      "theoretical_loss": 3.42090146121315,
      "tokens_seen": 2289631232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003092678034102307,
      "loss": 1.3231,
      "theoretical_loss": 3.420893578879441,
      "tokens_seen": 2289696768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030924774322968907,
      "loss": 1.361,
      "theoretical_loss": 3.420885696834506,
      "tokens_seen": 2289762304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030922768304914743,
      "loss": 1.2923,
      "theoretical_loss": 3.4208778150783266,
      "tokens_seen": 2289827840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003092076228686058,
      "loss": 1.2441,
      "theoretical_loss": 3.4208699336108843,
      "tokens_seen": 2289893376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091875626880642,
      "loss": 1.3607,
      "theoretical_loss": 3.4208620524321596,
      "tokens_seen": 2289958912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091675025075226,
      "loss": 1.4097,
      "theoretical_loss": 3.420854171542134,
      "tokens_seen": 2290024448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030914744232698094,
      "loss": 1.2348,
      "theoretical_loss": 3.420846290940789,
      "tokens_seen": 2290089984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091273821464393,
      "loss": 1.34,
      "theoretical_loss": 3.420838410628105,
      "tokens_seen": 2290155520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091073219658977,
      "loss": 1.3663,
      "theoretical_loss": 3.420830530604064,
      "tokens_seen": 2290221056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090872617853561,
      "loss": 1.2582,
      "theoretical_loss": 3.4208226508686463,
      "tokens_seen": 2290286592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030906720160481444,
      "loss": 1.251,
      "theoretical_loss": 3.420814771421834,
      "tokens_seen": 2290352128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090471414242728,
      "loss": 1.2802,
      "theoretical_loss": 3.4208068922636077,
      "tokens_seen": 2290417664
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030902708124373116,
      "loss": 1.2286,
      "theoretical_loss": 3.4207990133939488,
      "tokens_seen": 2290483200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090070210631896,
      "loss": 1.3673,
      "theoretical_loss": 3.4207911348128377,
      "tokens_seen": 2290548736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030898696088264794,
      "loss": 1.3027,
      "theoretical_loss": 3.4207832565202567,
      "tokens_seen": 2290614272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003089669007021063,
      "loss": 1.3198,
      "theoretical_loss": 3.4207753785161867,
      "tokens_seen": 2290679808
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030894684052156467,
      "loss": 1.386,
      "theoretical_loss": 3.4207675008006087,
      "tokens_seen": 2290745344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003089267803410231,
      "loss": 1.2098,
      "theoretical_loss": 3.4207596233735034,
      "tokens_seen": 2290810880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030890672016048145,
      "loss": 1.3366,
      "theoretical_loss": 3.4207517462348527,
      "tokens_seen": 2290876416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088866599799398,
      "loss": 1.3594,
      "theoretical_loss": 3.420743869384638,
      "tokens_seen": 2290941952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030886659979939817,
      "loss": 1.389,
      "theoretical_loss": 3.4207359928228396,
      "tokens_seen": 2291007488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030884653961885653,
      "loss": 1.2543,
      "theoretical_loss": 3.420728116549439,
      "tokens_seen": 2291073024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030882647943831495,
      "loss": 1.2422,
      "theoretical_loss": 3.4207202405644175,
      "tokens_seen": 2291138560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088064192577733,
      "loss": 1.3826,
      "theoretical_loss": 3.4207123648677564,
      "tokens_seen": 2291204096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087863590772317,
      "loss": 1.3423,
      "theoretical_loss": 3.420704489459437,
      "tokens_seen": 2291269632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030876629889669004,
      "loss": 1.2938,
      "theoretical_loss": 3.42069661433944,
      "tokens_seen": 2291335168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030874623871614845,
      "loss": 1.3459,
      "theoretical_loss": 3.420688739507747,
      "tokens_seen": 2291400704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087261785356068,
      "loss": 1.3615,
      "theoretical_loss": 3.420680864964339,
      "tokens_seen": 2291466240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087061183550652,
      "loss": 1.2909,
      "theoretical_loss": 3.4206729907091975,
      "tokens_seen": 2291531776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030868605817452354,
      "loss": 1.294,
      "theoretical_loss": 3.4206651167423034,
      "tokens_seen": 2291597312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086659979939819,
      "loss": 1.3171,
      "theoretical_loss": 3.420657243063638,
      "tokens_seen": 2291662848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086459378134403,
      "loss": 1.3536,
      "theoretical_loss": 3.420649369673183,
      "tokens_seen": 2291728384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086258776328987,
      "loss": 1.3219,
      "theoretical_loss": 3.4206414965709184,
      "tokens_seen": 2291793920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030860581745235704,
      "loss": 1.3707,
      "theoretical_loss": 3.4206336237568267,
      "tokens_seen": 2291859456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085857572718154,
      "loss": 1.3637,
      "theoretical_loss": 3.420625751230888,
      "tokens_seen": 2291924992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085656970912738,
      "loss": 1.3018,
      "theoretical_loss": 3.4206178789930846,
      "tokens_seen": 2291990528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030854563691073224,
      "loss": 1.356,
      "theoretical_loss": 3.4206100070433965,
      "tokens_seen": 2292056064
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085255767301906,
      "loss": 1.2556,
      "theoretical_loss": 3.420602135381806,
      "tokens_seen": 2292121600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030850551654964896,
      "loss": 1.2167,
      "theoretical_loss": 3.420594264008294,
      "tokens_seen": 2292187136
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 834793,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2405096292495728,
      "objective/train/theoretical_loss": 3.4205863929228415,
      "objective/train/tokens_used": 662711776,
      "theoretical_loss": 3.4205863929228415,
      "tokens_seen": 2292252672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003084854563691074,
      "loss": 1.2682,
      "theoretical_loss": 3.4205863929228415,
      "tokens_seen": 2292252672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030846539618856574,
      "loss": 1.3506,
      "theoretical_loss": 3.42057852212543,
      "tokens_seen": 2292318208
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003084453360080241,
      "loss": 1.1343,
      "theoretical_loss": 3.4205706516160403,
      "tokens_seen": 2292383744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030842527582748247,
      "loss": 1.2668,
      "theoretical_loss": 3.420562781394654,
      "tokens_seen": 2292449280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030840521564694083,
      "loss": 1.3169,
      "theoretical_loss": 3.4205549114612523,
      "tokens_seen": 2292514816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030838515546639924,
      "loss": 1.3265,
      "theoretical_loss": 3.4205470418158166,
      "tokens_seen": 2292580352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003083650952858576,
      "loss": 1.3332,
      "theoretical_loss": 3.4205391724583274,
      "tokens_seen": 2292645888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030834503510531597,
      "loss": 1.382,
      "theoretical_loss": 3.420531303388767,
      "tokens_seen": 2292711424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030832497492477433,
      "loss": 1.3819,
      "theoretical_loss": 3.420523434607116,
      "tokens_seen": 2292776960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030830491474423275,
      "loss": 1.3875,
      "theoretical_loss": 3.4205155661133553,
      "tokens_seen": 2292842496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082848545636911,
      "loss": 1.3447,
      "theoretical_loss": 3.4205076979074667,
      "tokens_seen": 2292908032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030826479438314947,
      "loss": 1.2513,
      "theoretical_loss": 3.420499829989431,
      "tokens_seen": 2292973568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030824473420260783,
      "loss": 1.256,
      "theoretical_loss": 3.42049196235923,
      "tokens_seen": 2293039104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082246740220662,
      "loss": 1.2473,
      "theoretical_loss": 3.4204840950168447,
      "tokens_seen": 2293104640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082046138415246,
      "loss": 1.2046,
      "theoretical_loss": 3.420476227962256,
      "tokens_seen": 2293170176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308184553660983,
      "loss": 1.3442,
      "theoretical_loss": 3.420468361195446,
      "tokens_seen": 2293235712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030816449348044134,
      "loss": 1.3664,
      "theoretical_loss": 3.4204604947163952,
      "tokens_seen": 2293301248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081444332998997,
      "loss": 1.2128,
      "theoretical_loss": 3.4204526285250854,
      "tokens_seen": 2293366784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081243731193581,
      "loss": 1.3575,
      "theoretical_loss": 3.420444762621497,
      "tokens_seen": 2293432320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081043129388165,
      "loss": 1.3835,
      "theoretical_loss": 3.4204368970056116,
      "tokens_seen": 2293497856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030808425275827484,
      "loss": 1.4044,
      "theoretical_loss": 3.4204290316774113,
      "tokens_seen": 2293563392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003080641925777332,
      "loss": 1.2343,
      "theoretical_loss": 3.420421166636876,
      "tokens_seen": 2293628928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030804413239719157,
      "loss": 1.325,
      "theoretical_loss": 3.420413301883988,
      "tokens_seen": 2293694464
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030802407221665,
      "loss": 1.1399,
      "theoretical_loss": 3.4204054374187285,
      "tokens_seen": 2293760000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030800401203610834,
      "loss": 1.3272,
      "theoretical_loss": 3.420397573241078,
      "tokens_seen": 2293825536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079839518555667,
      "loss": 1.3813,
      "theoretical_loss": 3.4203897093510185,
      "tokens_seen": 2293891072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030796389167502507,
      "loss": 1.3791,
      "theoretical_loss": 3.4203818457485307,
      "tokens_seen": 2293956608
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079438314944835,
      "loss": 1.3351,
      "theoretical_loss": 3.4203739824335964,
      "tokens_seen": 2294022144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030792377131394185,
      "loss": 1.3899,
      "theoretical_loss": 3.4203661194061965,
      "tokens_seen": 2294087680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079037111334002,
      "loss": 1.3845,
      "theoretical_loss": 3.420358256666313,
      "tokens_seen": 2294153216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030788365095285857,
      "loss": 1.3103,
      "theoretical_loss": 3.420350394213926,
      "tokens_seen": 2294218752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030786359077231693,
      "loss": 1.2906,
      "theoretical_loss": 3.4203425320490175,
      "tokens_seen": 2294284288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030784353059177535,
      "loss": 1.2971,
      "theoretical_loss": 3.4203346701715684,
      "tokens_seen": 2294349824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078234704112337,
      "loss": 1.3238,
      "theoretical_loss": 3.4203268085815606,
      "tokens_seen": 2294415360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078034102306921,
      "loss": 1.3012,
      "theoretical_loss": 3.420318947278975,
      "tokens_seen": 2294480896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030778335005015044,
      "loss": 1.3194,
      "theoretical_loss": 3.4203110862637924,
      "tokens_seen": 2294546432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030776328986960885,
      "loss": 1.4411,
      "theoretical_loss": 3.420303225535995,
      "tokens_seen": 2294611968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077432296890672,
      "loss": 1.1563,
      "theoretical_loss": 3.4202953650955634,
      "tokens_seen": 2294677504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077231695085256,
      "loss": 1.3424,
      "theoretical_loss": 3.4202875049424795,
      "tokens_seen": 2294743040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030770310932798394,
      "loss": 1.3091,
      "theoretical_loss": 3.420279645076724,
      "tokens_seen": 2294808576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076830491474423,
      "loss": 1.4602,
      "theoretical_loss": 3.4202717854982785,
      "tokens_seen": 2294874112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076629889669007,
      "loss": 1.3731,
      "theoretical_loss": 3.420263926207124,
      "tokens_seen": 2294939648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076429287863591,
      "loss": 1.3165,
      "theoretical_loss": 3.4202560672032423,
      "tokens_seen": 2295005184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030762286860581744,
      "loss": 1.285,
      "theoretical_loss": 3.420248208486614,
      "tokens_seen": 2295070720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076028084252758,
      "loss": 1.3232,
      "theoretical_loss": 3.420240350057221,
      "tokens_seen": 2295136256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075827482447342,
      "loss": 1.1694,
      "theoretical_loss": 3.4202324919150446,
      "tokens_seen": 2295201792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075626880641926,
      "loss": 1.3174,
      "theoretical_loss": 3.420224634060066,
      "tokens_seen": 2295267328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030754262788365095,
      "loss": 1.2465,
      "theoretical_loss": 3.420216776492266,
      "tokens_seen": 2295332864
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075225677031093,
      "loss": 1.3986,
      "theoretical_loss": 3.420208919211627,
      "tokens_seen": 2295398400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075025075225677,
      "loss": 1.3003,
      "theoretical_loss": 3.420201062218129,
      "tokens_seen": 2295463936
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 836058,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1093416213989258,
      "objective/train/theoretical_loss": 3.420193205511754,
      "objective/train/tokens_used": 665988576,
      "theoretical_loss": 3.420193205511754,
      "tokens_seen": 2295529472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074824473420261,
      "loss": 1.1661,
      "theoretical_loss": 3.420193205511754,
      "tokens_seen": 2295529472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030746238716148445,
      "loss": 1.3943,
      "theoretical_loss": 3.4201853490924834,
      "tokens_seen": 2295595008
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074423269809428,
      "loss": 1.3037,
      "theoretical_loss": 3.420177492960298,
      "tokens_seen": 2295660544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074222668004012,
      "loss": 1.2376,
      "theoretical_loss": 3.42016963711518,
      "tokens_seen": 2295726080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074022066198596,
      "loss": 1.3062,
      "theoretical_loss": 3.42016178155711,
      "tokens_seen": 2295791616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030738214643931795,
      "loss": 1.3919,
      "theoretical_loss": 3.4201539262860696,
      "tokens_seen": 2295857152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073620862587763,
      "loss": 1.3921,
      "theoretical_loss": 3.4201460713020397,
      "tokens_seen": 2295922688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073420260782347,
      "loss": 1.371,
      "theoretical_loss": 3.420138216605002,
      "tokens_seen": 2295988224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073219658976931,
      "loss": 1.3823,
      "theoretical_loss": 3.4201303621949384,
      "tokens_seen": 2296053760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030730190571715146,
      "loss": 1.3091,
      "theoretical_loss": 3.420122508071829,
      "tokens_seen": 2296119296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072818455366098,
      "loss": 1.3259,
      "theoretical_loss": 3.4201146542356557,
      "tokens_seen": 2296184832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072617853560682,
      "loss": 1.3361,
      "theoretical_loss": 3.4201068006864,
      "tokens_seen": 2296250368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030724172517552654,
      "loss": 1.4059,
      "theoretical_loss": 3.420098947424043,
      "tokens_seen": 2296315904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030722166499498496,
      "loss": 1.2967,
      "theoretical_loss": 3.4200910944485665,
      "tokens_seen": 2296381440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072016048144433,
      "loss": 1.324,
      "theoretical_loss": 3.4200832417599507,
      "tokens_seen": 2296446976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071815446339017,
      "loss": 1.4418,
      "theoretical_loss": 3.4200753893581783,
      "tokens_seen": 2296512512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030716148445336005,
      "loss": 1.4553,
      "theoretical_loss": 3.42006753724323,
      "tokens_seen": 2296578048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030714142427281846,
      "loss": 1.3193,
      "theoretical_loss": 3.4200596854150866,
      "tokens_seen": 2296643584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071213640922768,
      "loss": 1.3432,
      "theoretical_loss": 3.4200518338737305,
      "tokens_seen": 2296709120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071013039117352,
      "loss": 1.3294,
      "theoretical_loss": 3.420043982619142,
      "tokens_seen": 2296774656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030708124373119355,
      "loss": 1.2788,
      "theoretical_loss": 3.4200361316513037,
      "tokens_seen": 2296840192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070611835506519,
      "loss": 1.3076,
      "theoretical_loss": 3.420028280970196,
      "tokens_seen": 2296905728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070411233701104,
      "loss": 1.3058,
      "theoretical_loss": 3.4200204305758004,
      "tokens_seen": 2296971264
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030702106318956874,
      "loss": 1.3349,
      "theoretical_loss": 3.4200125804680983,
      "tokens_seen": 2297036800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070010030090271,
      "loss": 1.3601,
      "theoretical_loss": 3.420004730647071,
      "tokens_seen": 2297102336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030698094282848547,
      "loss": 1.3333,
      "theoretical_loss": 3.4199968811127004,
      "tokens_seen": 2297167872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003069608826479439,
      "loss": 1.2505,
      "theoretical_loss": 3.419989031864967,
      "tokens_seen": 2297233408
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030694082246740225,
      "loss": 1.3008,
      "theoretical_loss": 3.419981182903852,
      "tokens_seen": 2297298944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003069207622868606,
      "loss": 1.2611,
      "theoretical_loss": 3.419973334229338,
      "tokens_seen": 2297364480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030690070210631897,
      "loss": 1.3647,
      "theoretical_loss": 3.4199654858414057,
      "tokens_seen": 2297430016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030688064192577733,
      "loss": 1.2623,
      "theoretical_loss": 3.419957637740036,
      "tokens_seen": 2297495552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030686058174523575,
      "loss": 1.3993,
      "theoretical_loss": 3.419949789925211,
      "tokens_seen": 2297561088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068405215646941,
      "loss": 1.266,
      "theoretical_loss": 3.4199419423969117,
      "tokens_seen": 2297626624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068204613841525,
      "loss": 1.2297,
      "theoretical_loss": 3.4199340951551194,
      "tokens_seen": 2297692160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030680040120361084,
      "loss": 1.3165,
      "theoretical_loss": 3.4199262481998156,
      "tokens_seen": 2297757696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030678034102306925,
      "loss": 1.3014,
      "theoretical_loss": 3.419918401530982,
      "tokens_seen": 2297823232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003067602808425276,
      "loss": 1.2619,
      "theoretical_loss": 3.419910555148599,
      "tokens_seen": 2297888768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306740220661986,
      "loss": 1.3285,
      "theoretical_loss": 3.419902709052649,
      "tokens_seen": 2297954304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030672016048144434,
      "loss": 1.3551,
      "theoretical_loss": 3.4198948632431128,
      "tokens_seen": 2298019840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003067001003009027,
      "loss": 1.4281,
      "theoretical_loss": 3.419887017719972,
      "tokens_seen": 2298085376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066800401203611,
      "loss": 1.3251,
      "theoretical_loss": 3.419879172483208,
      "tokens_seen": 2298150912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066599799398195,
      "loss": 1.3003,
      "theoretical_loss": 3.419871327532802,
      "tokens_seen": 2298216448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030663991975927784,
      "loss": 1.3669,
      "theoretical_loss": 3.4198634828687355,
      "tokens_seen": 2298281984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066198595787362,
      "loss": 1.3332,
      "theoretical_loss": 3.41985563849099,
      "tokens_seen": 2298347520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065997993981946,
      "loss": 1.2826,
      "theoretical_loss": 3.4198477943995464,
      "tokens_seen": 2298413056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306579739217653,
      "loss": 1.3073,
      "theoretical_loss": 3.419839950594387,
      "tokens_seen": 2298478592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030655967903711135,
      "loss": 1.2733,
      "theoretical_loss": 3.419832107075492,
      "tokens_seen": 2298544128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065396188565697,
      "loss": 1.4615,
      "theoretical_loss": 3.419824263842844,
      "tokens_seen": 2298609664
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065195586760281,
      "loss": 1.3059,
      "theoretical_loss": 3.4198164208964235,
      "tokens_seen": 2298675200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064994984954865,
      "loss": 1.2271,
      "theoretical_loss": 3.4198085782362124,
      "tokens_seen": 2298740736
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 837138,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1962976455688477,
      "objective/train/theoretical_loss": 3.419800735862192,
      "objective/train/tokens_used": 669265376,
      "theoretical_loss": 3.419800735862192,
      "tokens_seen": 2298806272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030647943831494485,
      "loss": 1.1987,
      "theoretical_loss": 3.419800735862192,
      "tokens_seen": 2298806272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064593781344032,
      "loss": 1.4421,
      "theoretical_loss": 3.4197928937743436,
      "tokens_seen": 2298871808
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064393179538616,
      "loss": 1.4064,
      "theoretical_loss": 3.4197850519726485,
      "tokens_seen": 2298937344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030641925777332,
      "loss": 1.3172,
      "theoretical_loss": 3.419777210457088,
      "tokens_seen": 2299002880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030639919759277835,
      "loss": 1.1757,
      "theoretical_loss": 3.4197693692276436,
      "tokens_seen": 2299068416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063791374122367,
      "loss": 1.2895,
      "theoretical_loss": 3.4197615282842975,
      "tokens_seen": 2299133952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063590772316951,
      "loss": 1.3535,
      "theoretical_loss": 3.41975368762703,
      "tokens_seen": 2299199488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063390170511535,
      "loss": 1.2948,
      "theoretical_loss": 3.4197458472558226,
      "tokens_seen": 2299265024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030631895687061186,
      "loss": 1.3044,
      "theoretical_loss": 3.4197380071706576,
      "tokens_seen": 2299330560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062988966900702,
      "loss": 1.3765,
      "theoretical_loss": 3.4197301673715157,
      "tokens_seen": 2299396096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062788365095286,
      "loss": 1.2523,
      "theoretical_loss": 3.419722327858379,
      "tokens_seen": 2299461632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030625877632898694,
      "loss": 1.2807,
      "theoretical_loss": 3.4197144886312274,
      "tokens_seen": 2299527168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030623871614844536,
      "loss": 1.3466,
      "theoretical_loss": 3.4197066496900437,
      "tokens_seen": 2299592704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062186559679037,
      "loss": 1.2634,
      "theoretical_loss": 3.419698811034809,
      "tokens_seen": 2299658240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061985957873621,
      "loss": 1.2867,
      "theoretical_loss": 3.4196909726655047,
      "tokens_seen": 2299723776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030617853560682045,
      "loss": 1.2816,
      "theoretical_loss": 3.4196831345821117,
      "tokens_seen": 2299789312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030615847542627886,
      "loss": 1.293,
      "theoretical_loss": 3.4196752967846127,
      "tokens_seen": 2299854848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061384152457372,
      "loss": 1.3621,
      "theoretical_loss": 3.4196674592729877,
      "tokens_seen": 2299920384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061183550651956,
      "loss": 1.3215,
      "theoretical_loss": 3.419659622047219,
      "tokens_seen": 2299985920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030609829488465395,
      "loss": 1.4327,
      "theoretical_loss": 3.4196517851072876,
      "tokens_seen": 2300051456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003060782347041123,
      "loss": 1.3545,
      "theoretical_loss": 3.4196439484531753,
      "tokens_seen": 2300116992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030605817452357073,
      "loss": 1.4895,
      "theoretical_loss": 3.419636112084863,
      "tokens_seen": 2300182528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003060381143430291,
      "loss": 1.2612,
      "theoretical_loss": 3.4196282760023324,
      "tokens_seen": 2300248064
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030601805416248745,
      "loss": 1.2977,
      "theoretical_loss": 3.4196204402055654,
      "tokens_seen": 2300313600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003059979939819458,
      "loss": 1.3587,
      "theoretical_loss": 3.419612604694543,
      "tokens_seen": 2300379136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030597793380140423,
      "loss": 1.3322,
      "theoretical_loss": 3.4196047694692466,
      "tokens_seen": 2300444672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003059578736208626,
      "loss": 1.3294,
      "theoretical_loss": 3.419596934529658,
      "tokens_seen": 2300510208
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030593781344032096,
      "loss": 1.3496,
      "theoretical_loss": 3.419589099875758,
      "tokens_seen": 2300575744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003059177532597793,
      "loss": 1.2959,
      "theoretical_loss": 3.4195812655075284,
      "tokens_seen": 2300641280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058976930792377,
      "loss": 1.3593,
      "theoretical_loss": 3.4195734314249506,
      "tokens_seen": 2300706816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058776328986961,
      "loss": 1.273,
      "theoretical_loss": 3.4195655976280066,
      "tokens_seen": 2300772352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030585757271815446,
      "loss": 1.3414,
      "theoretical_loss": 3.419557764116677,
      "tokens_seen": 2300837888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058375125376128,
      "loss": 1.4565,
      "theoretical_loss": 3.419549930890944,
      "tokens_seen": 2300903424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058174523570712,
      "loss": 1.2996,
      "theoretical_loss": 3.419542097950788,
      "tokens_seen": 2300968960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057973921765296,
      "loss": 1.3554,
      "theoretical_loss": 3.4195342652961918,
      "tokens_seen": 2301034496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030577733199598796,
      "loss": 1.4252,
      "theoretical_loss": 3.4195264329271353,
      "tokens_seen": 2301100032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057572718154463,
      "loss": 1.3005,
      "theoretical_loss": 3.4195186008436016,
      "tokens_seen": 2301165568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057372116349047,
      "loss": 1.278,
      "theoretical_loss": 3.419510769045571,
      "tokens_seen": 2301231104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030571715145436305,
      "loss": 1.3614,
      "theoretical_loss": 3.4195029375330255,
      "tokens_seen": 2301296640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030569709127382147,
      "loss": 1.4513,
      "theoretical_loss": 3.4194951063059467,
      "tokens_seen": 2301362176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030567703109327983,
      "loss": 1.3966,
      "theoretical_loss": 3.419487275364315,
      "tokens_seen": 2301427712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003056569709127382,
      "loss": 1.3468,
      "theoretical_loss": 3.4194794447081134,
      "tokens_seen": 2301493248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030563691073219655,
      "loss": 1.2386,
      "theoretical_loss": 3.4194716143373225,
      "tokens_seen": 2301558784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030561685055165497,
      "loss": 1.2769,
      "theoretical_loss": 3.4194637842519238,
      "tokens_seen": 2301624320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030559679037111333,
      "loss": 1.3143,
      "theoretical_loss": 3.4194559544518985,
      "tokens_seen": 2301689856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055767301905717,
      "loss": 1.3054,
      "theoretical_loss": 3.419448124937229,
      "tokens_seen": 2301755392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030555667001003006,
      "loss": 1.1853,
      "theoretical_loss": 3.419440295707896,
      "tokens_seen": 2301820928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055366098294885,
      "loss": 1.2656,
      "theoretical_loss": 3.4194324667638814,
      "tokens_seen": 2301886464
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055165496489469,
      "loss": 1.4154,
      "theoretical_loss": 3.4194246381051663,
      "tokens_seen": 2301952000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030549648946840525,
      "loss": 1.3523,
      "theoretical_loss": 3.419416809731732,
      "tokens_seen": 2302017536
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 838318,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2945561408996582,
      "objective/train/theoretical_loss": 3.419408981643561,
      "objective/train/tokens_used": 672542176,
      "theoretical_loss": 3.419408981643561,
      "tokens_seen": 2302083072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054764292878636,
      "loss": 1.3049,
      "theoretical_loss": 3.419408981643561,
      "tokens_seen": 2302083072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000305456369107322,
      "loss": 1.3339,
      "theoretical_loss": 3.4194011538406333,
      "tokens_seen": 2302148608
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054363089267804,
      "loss": 1.3502,
      "theoretical_loss": 3.4193933263229317,
      "tokens_seen": 2302214144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030541624874623875,
      "loss": 1.3428,
      "theoretical_loss": 3.4193854990904375,
      "tokens_seen": 2302279680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053961885656971,
      "loss": 1.3427,
      "theoretical_loss": 3.4193776721431313,
      "tokens_seen": 2302345216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053761283851555,
      "loss": 1.281,
      "theoretical_loss": 3.4193698454809955,
      "tokens_seen": 2302410752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053560682046139,
      "loss": 1.3703,
      "theoretical_loss": 3.4193620191040113,
      "tokens_seen": 2302476288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030533600802407226,
      "loss": 1.3498,
      "theoretical_loss": 3.4193541930121603,
      "tokens_seen": 2302541824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053159478435306,
      "loss": 1.2243,
      "theoretical_loss": 3.4193463672054234,
      "tokens_seen": 2302607360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000305295887662989,
      "loss": 1.1954,
      "theoretical_loss": 3.419338541683783,
      "tokens_seen": 2302672896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030527582748244734,
      "loss": 1.3808,
      "theoretical_loss": 3.41933071644722,
      "tokens_seen": 2302738432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030525576730190576,
      "loss": 1.3674,
      "theoretical_loss": 3.4193228914957157,
      "tokens_seen": 2302803968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052357071213641,
      "loss": 1.3431,
      "theoretical_loss": 3.4193150668292525,
      "tokens_seen": 2302869504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052156469408225,
      "loss": 1.3908,
      "theoretical_loss": 3.419307242447811,
      "tokens_seen": 2302935040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030519558676028085,
      "loss": 1.3625,
      "theoretical_loss": 3.4192994183513736,
      "tokens_seen": 2303000576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030517552657973926,
      "loss": 1.375,
      "theoretical_loss": 3.4192915945399207,
      "tokens_seen": 2303066112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003051554663991976,
      "loss": 1.2921,
      "theoretical_loss": 3.4192837710134345,
      "tokens_seen": 2303131648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000305135406218656,
      "loss": 1.3388,
      "theoretical_loss": 3.419275947771897,
      "tokens_seen": 2303197184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030511534603811435,
      "loss": 1.3836,
      "theoretical_loss": 3.4192681248152885,
      "tokens_seen": 2303262720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050952858575727,
      "loss": 1.2521,
      "theoretical_loss": 3.4192603021435914,
      "tokens_seen": 2303328256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030507522567703113,
      "loss": 1.2955,
      "theoretical_loss": 3.419252479756787,
      "tokens_seen": 2303393792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050551654964895,
      "loss": 1.4169,
      "theoretical_loss": 3.4192446576548567,
      "tokens_seen": 2303459328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030503510531594785,
      "loss": 1.3302,
      "theoretical_loss": 3.419236835837782,
      "tokens_seen": 2303524864
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050150451354062,
      "loss": 1.2711,
      "theoretical_loss": 3.419229014305545,
      "tokens_seen": 2303590400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030499498495486463,
      "loss": 1.2155,
      "theoretical_loss": 3.419221193058126,
      "tokens_seen": 2303655936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000304974924774323,
      "loss": 1.3041,
      "theoretical_loss": 3.4192133720955074,
      "tokens_seen": 2303721472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030495486459378136,
      "loss": 1.3629,
      "theoretical_loss": 3.4192055514176714,
      "tokens_seen": 2303787008
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003049348044132397,
      "loss": 1.2605,
      "theoretical_loss": 3.419197731024598,
      "tokens_seen": 2303852544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003049147442326981,
      "loss": 1.3385,
      "theoretical_loss": 3.4191899109162693,
      "tokens_seen": 2303918080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048946840521565,
      "loss": 1.4245,
      "theoretical_loss": 3.4191820910926674,
      "tokens_seen": 2303983616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030487462387161486,
      "loss": 1.3832,
      "theoretical_loss": 3.419174271553773,
      "tokens_seen": 2304049152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048545636910732,
      "loss": 1.3049,
      "theoretical_loss": 3.419166452299569,
      "tokens_seen": 2304114688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048345035105316,
      "loss": 1.313,
      "theoretical_loss": 3.4191586333300354,
      "tokens_seen": 2304180224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030481444332999,
      "loss": 1.2704,
      "theoretical_loss": 3.419150814645154,
      "tokens_seen": 2304245760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030479438314944836,
      "loss": 1.3347,
      "theoretical_loss": 3.4191429962449074,
      "tokens_seen": 2304311296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047743229689067,
      "loss": 1.3249,
      "theoretical_loss": 3.419135178129276,
      "tokens_seen": 2304376832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047542627883651,
      "loss": 1.3527,
      "theoretical_loss": 3.419127360298242,
      "tokens_seen": 2304442368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030473420260782345,
      "loss": 1.4751,
      "theoretical_loss": 3.4191195427517864,
      "tokens_seen": 2304507904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030471414242728187,
      "loss": 1.1912,
      "theoretical_loss": 3.4191117254898913,
      "tokens_seen": 2304573440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030469408224674023,
      "loss": 1.3793,
      "theoretical_loss": 3.419103908512538,
      "tokens_seen": 2304638976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003046740220661986,
      "loss": 1.4098,
      "theoretical_loss": 3.419096091819708,
      "tokens_seen": 2304704512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030465396188565695,
      "loss": 1.306,
      "theoretical_loss": 3.419088275411383,
      "tokens_seen": 2304770048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030463390170511537,
      "loss": 1.3689,
      "theoretical_loss": 3.4190804592875446,
      "tokens_seen": 2304835584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030461384152457373,
      "loss": 1.2591,
      "theoretical_loss": 3.4190726434481737,
      "tokens_seen": 2304901120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045937813440321,
      "loss": 1.2996,
      "theoretical_loss": 3.419064827893253,
      "tokens_seen": 2304966656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030457372116349046,
      "loss": 1.3937,
      "theoretical_loss": 3.4190570126227633,
      "tokens_seen": 2305032192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045536609829488,
      "loss": 1.2631,
      "theoretical_loss": 3.4190491976366864,
      "tokens_seen": 2305097728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030453360080240723,
      "loss": 1.2879,
      "theoretical_loss": 3.4190413829350037,
      "tokens_seen": 2305163264
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045135406218656,
      "loss": 1.2513,
      "theoretical_loss": 3.4190335685176967,
      "tokens_seen": 2305228800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030449348044132396,
      "loss": 1.2981,
      "theoretical_loss": 3.419025754384747,
      "tokens_seen": 2305294336
    },
    {
      "debugging/Compilability": 1.0,
      "debugging/distinct-1-grams": 0.753016582273028,
      "debugging/entropy-1-grams": 5.229850717926142,
      "debugging/length": 439.2307692307692,
      "debugging/num_segments": 13,
      "debugging/score": 0.013377460944710597,
      "debugging/score_std": 0.02051348399645926,
      "epoch": 0.4,
      "objective/train/docs_used": 839437,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3634417057037354,
      "objective/train/theoretical_loss": 3.4190179405361367,
      "objective/train/tokens_used": 675818976,
      "theoretical_loss": 3.4190179405361367,
      "tokens_seen": 2305359872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044734202607823,
      "loss": 1.4161,
      "theoretical_loss": 3.4190179405361367,
      "tokens_seen": 2305359872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030445336008024074,
      "loss": 1.3611,
      "theoretical_loss": 3.4190101269718465,
      "tokens_seen": 2305425408
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044332998996991,
      "loss": 1.3643,
      "theoretical_loss": 3.419002313691859,
      "tokens_seen": 2305490944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030441323971915746,
      "loss": 1.4038,
      "theoretical_loss": 3.418994500696155,
      "tokens_seen": 2305556480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043931795386158,
      "loss": 1.2971,
      "theoretical_loss": 3.4189866879847157,
      "tokens_seen": 2305622016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043731193580742,
      "loss": 1.3797,
      "theoretical_loss": 3.4189788755575243,
      "tokens_seen": 2305687552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043530591775326,
      "loss": 1.4518,
      "theoretical_loss": 3.4189710634145607,
      "tokens_seen": 2305753088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030433299899699097,
      "loss": 1.375,
      "theoretical_loss": 3.418963251555807,
      "tokens_seen": 2305818624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030431293881644933,
      "loss": 1.352,
      "theoretical_loss": 3.418955439981245,
      "tokens_seen": 2305884160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042928786359077,
      "loss": 1.2615,
      "theoretical_loss": 3.4189476286908564,
      "tokens_seen": 2305949696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042728184553661,
      "loss": 1.2764,
      "theoretical_loss": 3.418939817684622,
      "tokens_seen": 2306015232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030425275827482447,
      "loss": 1.3208,
      "theoretical_loss": 3.4189320069625246,
      "tokens_seen": 2306080768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030423269809428283,
      "loss": 1.403,
      "theoretical_loss": 3.418924196524545,
      "tokens_seen": 2306146304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042126379137412,
      "loss": 1.2701,
      "theoretical_loss": 3.418916386370665,
      "tokens_seen": 2306211840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041925777331996,
      "loss": 1.2588,
      "theoretical_loss": 3.418908576500866,
      "tokens_seen": 2306277376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030417251755265797,
      "loss": 1.3732,
      "theoretical_loss": 3.41890076691513,
      "tokens_seen": 2306342912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030415245737211633,
      "loss": 1.3369,
      "theoretical_loss": 3.418892957613438,
      "tokens_seen": 2306408448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041323971915747,
      "loss": 1.3356,
      "theoretical_loss": 3.418885148595772,
      "tokens_seen": 2306473984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030411233701103306,
      "loss": 1.2508,
      "theoretical_loss": 3.4188773398621137,
      "tokens_seen": 2306539520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003040922768304915,
      "loss": 1.2693,
      "theoretical_loss": 3.4188695314124447,
      "tokens_seen": 2306605056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030407221664994984,
      "loss": 1.2421,
      "theoretical_loss": 3.4188617232467458,
      "tokens_seen": 2306670592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030405215646940825,
      "loss": 1.2955,
      "theoretical_loss": 3.4188539153649993,
      "tokens_seen": 2306736128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003040320962888666,
      "loss": 1.3536,
      "theoretical_loss": 3.4188461077671874,
      "tokens_seen": 2306801664
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030401203610832503,
      "loss": 1.3138,
      "theoretical_loss": 3.4188383004532903,
      "tokens_seen": 2306867200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003039919759277834,
      "loss": 1.3446,
      "theoretical_loss": 3.418830493423291,
      "tokens_seen": 2306932736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030397191574724176,
      "loss": 1.379,
      "theoretical_loss": 3.41882268667717,
      "tokens_seen": 2306998272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003039518555667001,
      "loss": 1.4427,
      "theoretical_loss": 3.4188148802149096,
      "tokens_seen": 2307063808
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003039317953861585,
      "loss": 1.3429,
      "theoretical_loss": 3.4188070740364913,
      "tokens_seen": 2307129344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003039117352056169,
      "loss": 1.5064,
      "theoretical_loss": 3.418799268141896,
      "tokens_seen": 2307194880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030389167502507526,
      "loss": 1.373,
      "theoretical_loss": 3.418791462531107,
      "tokens_seen": 2307260416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038716148445336,
      "loss": 1.3274,
      "theoretical_loss": 3.418783657204104,
      "tokens_seen": 2307325952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303851554663992,
      "loss": 1.307,
      "theoretical_loss": 3.41877585216087,
      "tokens_seen": 2307391488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038314944834504,
      "loss": 1.2884,
      "theoretical_loss": 3.418768047401386,
      "tokens_seen": 2307457024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030381143430290876,
      "loss": 1.3686,
      "theoretical_loss": 3.418760242925633,
      "tokens_seen": 2307522560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037913741223671,
      "loss": 1.2705,
      "theoretical_loss": 3.4187524387335944,
      "tokens_seen": 2307588096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037713139418255,
      "loss": 1.4601,
      "theoretical_loss": 3.4187446348252504,
      "tokens_seen": 2307653632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030375125376128385,
      "loss": 1.2987,
      "theoretical_loss": 3.418736831200583,
      "tokens_seen": 2307719168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030373119358074227,
      "loss": 1.3119,
      "theoretical_loss": 3.4187290278595737,
      "tokens_seen": 2307784704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030371113340020063,
      "loss": 1.3028,
      "theoretical_loss": 3.4187212248022045,
      "tokens_seen": 2307850240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303691073219659,
      "loss": 1.3852,
      "theoretical_loss": 3.4187134220284565,
      "tokens_seen": 2307915776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030367101303911735,
      "loss": 1.2782,
      "theoretical_loss": 3.418705619538312,
      "tokens_seen": 2307981312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030365095285857577,
      "loss": 1.3356,
      "theoretical_loss": 3.4186978173317524,
      "tokens_seen": 2308046848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030363089267803413,
      "loss": 1.3757,
      "theoretical_loss": 3.418690015408759,
      "tokens_seen": 2308112384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003036108324974925,
      "loss": 1.3665,
      "theoretical_loss": 3.418682213769314,
      "tokens_seen": 2308177920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030359077231695086,
      "loss": 1.3203,
      "theoretical_loss": 3.418674412413398,
      "tokens_seen": 2308243456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003035707121364092,
      "loss": 1.2473,
      "theoretical_loss": 3.418666611340994,
      "tokens_seen": 2308308992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030355065195586763,
      "loss": 1.4159,
      "theoretical_loss": 3.418658810552083,
      "tokens_seen": 2308374528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303530591775326,
      "loss": 1.4028,
      "theoretical_loss": 3.4186510100466463,
      "tokens_seen": 2308440064
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030351053159478436,
      "loss": 1.4188,
      "theoretical_loss": 3.418643209824666,
      "tokens_seen": 2308505600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034904714142427,
      "loss": 1.3897,
      "theoretical_loss": 3.418635409886124,
      "tokens_seen": 2308571136
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 840762,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2229909896850586,
      "objective/train/theoretical_loss": 3.4186276102310016,
      "objective/train/tokens_used": 679095776,
      "theoretical_loss": 3.4186276102310016,
      "tokens_seen": 2308636672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030347041123370114,
      "loss": 1.3159,
      "theoretical_loss": 3.4186276102310016,
      "tokens_seen": 2308636672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034503510531595,
      "loss": 1.2617,
      "theoretical_loss": 3.41861981085928,
      "tokens_seen": 2308702208
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030343029087261786,
      "loss": 1.3382,
      "theoretical_loss": 3.418612011770942,
      "tokens_seen": 2308767744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034102306920762,
      "loss": 1.3178,
      "theoretical_loss": 3.418604212965968,
      "tokens_seen": 2308833280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033901705115346,
      "loss": 1.3188,
      "theoretical_loss": 3.4185964144443406,
      "tokens_seen": 2308898816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303370110330993,
      "loss": 1.3796,
      "theoretical_loss": 3.418588616206041,
      "tokens_seen": 2308964352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030335005015045137,
      "loss": 1.2994,
      "theoretical_loss": 3.4185808182510513,
      "tokens_seen": 2309029888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030332998996990973,
      "loss": 1.2656,
      "theoretical_loss": 3.4185730205793528,
      "tokens_seen": 2309095424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033099297893681,
      "loss": 1.2961,
      "theoretical_loss": 3.418565223190927,
      "tokens_seen": 2309160960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032898696088265,
      "loss": 1.3741,
      "theoretical_loss": 3.418557426085756,
      "tokens_seen": 2309226496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030326980942828487,
      "loss": 1.2641,
      "theoretical_loss": 3.418549629263821,
      "tokens_seen": 2309292032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030324974924774323,
      "loss": 1.3316,
      "theoretical_loss": 3.4185418327251043,
      "tokens_seen": 2309357568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032296890672016,
      "loss": 1.2497,
      "theoretical_loss": 3.4185340364695866,
      "tokens_seen": 2309423104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030320962888666,
      "loss": 1.3562,
      "theoretical_loss": 3.418526240497251,
      "tokens_seen": 2309488640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030318956870611837,
      "loss": 1.3177,
      "theoretical_loss": 3.418518444808078,
      "tokens_seen": 2309554176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030316950852557673,
      "loss": 1.4123,
      "theoretical_loss": 3.4185106494020494,
      "tokens_seen": 2309619712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031494483450351,
      "loss": 1.4084,
      "theoretical_loss": 3.4185028542791476,
      "tokens_seen": 2309685248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030312938816449346,
      "loss": 1.3101,
      "theoretical_loss": 3.4184950594393535,
      "tokens_seen": 2309750784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031093279839519,
      "loss": 1.3506,
      "theoretical_loss": 3.4184872648826494,
      "tokens_seen": 2309816320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030308926780341024,
      "loss": 1.3299,
      "theoretical_loss": 3.4184794706090162,
      "tokens_seen": 2309881856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003030692076228686,
      "loss": 1.4423,
      "theoretical_loss": 3.4184716766184366,
      "tokens_seen": 2309947392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030304914744232696,
      "loss": 1.4018,
      "theoretical_loss": 3.4184638829108915,
      "tokens_seen": 2310012928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003030290872617854,
      "loss": 1.381,
      "theoretical_loss": 3.418456089486363,
      "tokens_seen": 2310078464
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030300902708124374,
      "loss": 1.3378,
      "theoretical_loss": 3.418448296344833,
      "tokens_seen": 2310144000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029889669007021,
      "loss": 1.34,
      "theoretical_loss": 3.4184405034862824,
      "tokens_seen": 2310209536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030296890672016046,
      "loss": 1.2839,
      "theoretical_loss": 3.418432710910693,
      "tokens_seen": 2310275072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029488465396188,
      "loss": 1.3221,
      "theoretical_loss": 3.4184249186180478,
      "tokens_seen": 2310340608
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030292878635907724,
      "loss": 1.3141,
      "theoretical_loss": 3.418417126608327,
      "tokens_seen": 2310406144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029087261785356,
      "loss": 1.3623,
      "theoretical_loss": 3.4184093348815128,
      "tokens_seen": 2310471680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030288866599799397,
      "loss": 1.25,
      "theoretical_loss": 3.418401543437587,
      "tokens_seen": 2310537216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030286860581745233,
      "loss": 1.4052,
      "theoretical_loss": 3.4183937522765317,
      "tokens_seen": 2310602752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030284854563691075,
      "loss": 1.3798,
      "theoretical_loss": 3.418385961398328,
      "tokens_seen": 2310668288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028284854563691,
      "loss": 1.2929,
      "theoretical_loss": 3.4183781708029572,
      "tokens_seen": 2310733824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030280842527582747,
      "loss": 1.3549,
      "theoretical_loss": 3.418370380490402,
      "tokens_seen": 2310799360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030278836509528583,
      "loss": 1.2175,
      "theoretical_loss": 3.418362590460644,
      "tokens_seen": 2310864896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027683049147442,
      "loss": 1.3518,
      "theoretical_loss": 3.418354800713664,
      "tokens_seen": 2310930432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027482447342026,
      "loss": 1.3995,
      "theoretical_loss": 3.4183470112494447,
      "tokens_seen": 2310995968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302728184553661,
      "loss": 1.3963,
      "theoretical_loss": 3.4183392220679676,
      "tokens_seen": 2311061504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030270812437311934,
      "loss": 1.3102,
      "theoretical_loss": 3.418331433169214,
      "tokens_seen": 2311127040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026880641925777,
      "loss": 1.3193,
      "theoretical_loss": 3.4183236445531664,
      "tokens_seen": 2311192576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026680040120361,
      "loss": 1.3907,
      "theoretical_loss": 3.4183158562198055,
      "tokens_seen": 2311258112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026479438314945,
      "loss": 1.368,
      "theoretical_loss": 3.4183080681691136,
      "tokens_seen": 2311323648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030262788365095284,
      "loss": 1.2964,
      "theoretical_loss": 3.4183002804010725,
      "tokens_seen": 2311389184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026078234704112,
      "loss": 1.2629,
      "theoretical_loss": 3.418292492915664,
      "tokens_seen": 2311454720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030258776328986956,
      "loss": 1.4242,
      "theoretical_loss": 3.4182847057128694,
      "tokens_seen": 2311520256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302567703109328,
      "loss": 1.3809,
      "theoretical_loss": 3.418276918792671,
      "tokens_seen": 2311585792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025476429287864,
      "loss": 1.2449,
      "theoretical_loss": 3.4182691321550496,
      "tokens_seen": 2311651328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030252758274824476,
      "loss": 1.2899,
      "theoretical_loss": 3.418261345799988,
      "tokens_seen": 2311716864
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025075225677031,
      "loss": 1.3976,
      "theoretical_loss": 3.418253559727467,
      "tokens_seen": 2311782400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030248746238716154,
      "loss": 1.2174,
      "theoretical_loss": 3.418245773937469,
      "tokens_seen": 2311847936
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 841951,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3188968896865845,
      "objective/train/theoretical_loss": 3.4182379884299756,
      "objective/train/tokens_used": 682372576,
      "theoretical_loss": 3.4182379884299756,
      "tokens_seen": 2311913472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003024674022066199,
      "loss": 1.316,
      "theoretical_loss": 3.4182379884299756,
      "tokens_seen": 2311913472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030244734202607826,
      "loss": 1.2904,
      "theoretical_loss": 3.4182302032049687,
      "tokens_seen": 2311979008
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003024272818455366,
      "loss": 1.3186,
      "theoretical_loss": 3.4182224182624297,
      "tokens_seen": 2312044544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302407221664995,
      "loss": 1.3953,
      "theoretical_loss": 3.4182146336023402,
      "tokens_seen": 2312110080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023871614844534,
      "loss": 1.3158,
      "theoretical_loss": 3.4182068492246827,
      "tokens_seen": 2312175616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030236710130391177,
      "loss": 1.3113,
      "theoretical_loss": 3.4181990651294383,
      "tokens_seen": 2312241152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030234704112337013,
      "loss": 1.3855,
      "theoretical_loss": 3.418191281316589,
      "tokens_seen": 2312306688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023269809428285,
      "loss": 1.2326,
      "theoretical_loss": 3.4181834977861163,
      "tokens_seen": 2312372224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023069207622869,
      "loss": 1.4527,
      "theoretical_loss": 3.418175714538002,
      "tokens_seen": 2312437760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030228686058174527,
      "loss": 1.244,
      "theoretical_loss": 3.418167931572228,
      "tokens_seen": 2312503296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030226680040120363,
      "loss": 1.2672,
      "theoretical_loss": 3.4181601488887763,
      "tokens_seen": 2312568832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302246740220662,
      "loss": 1.3127,
      "theoretical_loss": 3.4181523664876283,
      "tokens_seen": 2312634368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022266800401204,
      "loss": 1.4019,
      "theoretical_loss": 3.4181445843687657,
      "tokens_seen": 2312699904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030220661985957877,
      "loss": 1.2384,
      "theoretical_loss": 3.4181368025321706,
      "tokens_seen": 2312765440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030218655967903713,
      "loss": 1.1765,
      "theoretical_loss": 3.418129020977825,
      "tokens_seen": 2312830976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021664994984955,
      "loss": 1.4129,
      "theoretical_loss": 3.4181212397057097,
      "tokens_seen": 2312896512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030214643931795386,
      "loss": 1.2611,
      "theoretical_loss": 3.4181134587158066,
      "tokens_seen": 2312962048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021263791374123,
      "loss": 1.2861,
      "theoretical_loss": 3.4181056780080987,
      "tokens_seen": 2313027584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030210631895687064,
      "loss": 1.4248,
      "theoretical_loss": 3.418097897582567,
      "tokens_seen": 2313093120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302086258776329,
      "loss": 1.3863,
      "theoretical_loss": 3.418090117439193,
      "tokens_seen": 2313158656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030206619859578736,
      "loss": 1.3337,
      "theoretical_loss": 3.418082337577958,
      "tokens_seen": 2313224192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003020461384152458,
      "loss": 1.3666,
      "theoretical_loss": 3.418074557998845,
      "tokens_seen": 2313289728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030202607823470414,
      "loss": 1.4007,
      "theoretical_loss": 3.4180667787018355,
      "tokens_seen": 2313355264
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003020060180541625,
      "loss": 1.351,
      "theoretical_loss": 3.418058999686911,
      "tokens_seen": 2313420800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030198595787362087,
      "loss": 1.3477,
      "theoretical_loss": 3.418051220954053,
      "tokens_seen": 2313486336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030196589769307923,
      "loss": 1.3481,
      "theoretical_loss": 3.4180434425032438,
      "tokens_seen": 2313551872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030194583751253764,
      "loss": 1.3096,
      "theoretical_loss": 3.418035664334465,
      "tokens_seen": 2313617408
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301925777331996,
      "loss": 1.3581,
      "theoretical_loss": 3.4180278864476983,
      "tokens_seen": 2313682944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030190571715145437,
      "loss": 1.3355,
      "theoretical_loss": 3.4180201088429256,
      "tokens_seen": 2313748480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030188565697091273,
      "loss": 1.3508,
      "theoretical_loss": 3.4180123315201287,
      "tokens_seen": 2313814016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030186559679037115,
      "loss": 1.3134,
      "theoretical_loss": 3.418004554479289,
      "tokens_seen": 2313879552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003018455366098295,
      "loss": 1.3309,
      "theoretical_loss": 3.417996777720389,
      "tokens_seen": 2313945088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030182547642928787,
      "loss": 1.3344,
      "theoretical_loss": 3.41798900124341,
      "tokens_seen": 2314010624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030180541624874623,
      "loss": 1.2217,
      "theoretical_loss": 3.417981225048334,
      "tokens_seen": 2314076160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017853560682046,
      "loss": 1.38,
      "theoretical_loss": 3.417973449135143,
      "tokens_seen": 2314141696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301765295887663,
      "loss": 1.2584,
      "theoretical_loss": 3.4179656735038177,
      "tokens_seen": 2314207232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017452357071214,
      "loss": 1.3821,
      "theoretical_loss": 3.4179578981543415,
      "tokens_seen": 2314272768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030172517552657974,
      "loss": 1.3986,
      "theoretical_loss": 3.417950123086695,
      "tokens_seen": 2314338304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017051153460381,
      "loss": 1.3918,
      "theoretical_loss": 3.41794234830086,
      "tokens_seen": 2314403840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016850551654965,
      "loss": 1.2941,
      "theoretical_loss": 3.4179345737968196,
      "tokens_seen": 2314469376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016649949849549,
      "loss": 1.3583,
      "theoretical_loss": 3.4179267995745546,
      "tokens_seen": 2314534912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030164493480441324,
      "loss": 1.2991,
      "theoretical_loss": 3.4179190256340464,
      "tokens_seen": 2314600448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016248746238716,
      "loss": 1.2718,
      "theoretical_loss": 3.4179112519752777,
      "tokens_seen": 2314665984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030160481444332996,
      "loss": 1.3233,
      "theoretical_loss": 3.41790347859823,
      "tokens_seen": 2314731520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003015847542627884,
      "loss": 1.3227,
      "theoretical_loss": 3.417895705502885,
      "tokens_seen": 2314797056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030156469408224674,
      "loss": 1.3716,
      "theoretical_loss": 3.417887932689225,
      "tokens_seen": 2314862592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003015446339017051,
      "loss": 1.3808,
      "theoretical_loss": 3.417880160157231,
      "tokens_seen": 2314928128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030152457372116347,
      "loss": 1.382,
      "theoretical_loss": 3.417872387906885,
      "tokens_seen": 2314993664
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003015045135406219,
      "loss": 1.2498,
      "theoretical_loss": 3.4178646159381696,
      "tokens_seen": 2315059200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030148445336008025,
      "loss": 1.3623,
      "theoretical_loss": 3.4178568442510655,
      "tokens_seen": 2315124736
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 843188,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.470449686050415,
      "objective/train/theoretical_loss": 3.4178490728455557,
      "objective/train/tokens_used": 685649376,
      "theoretical_loss": 3.4178490728455557,
      "tokens_seen": 2315190272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003014643931795386,
      "loss": 1.411,
      "theoretical_loss": 3.4178490728455557,
      "tokens_seen": 2315190272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030144433299899697,
      "loss": 1.3366,
      "theoretical_loss": 3.417841301721621,
      "tokens_seen": 2315255808
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030142427281845533,
      "loss": 1.3363,
      "theoretical_loss": 3.4178335308792436,
      "tokens_seen": 2315321344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030140421263791375,
      "loss": 1.2642,
      "theoretical_loss": 3.417825760318406,
      "tokens_seen": 2315386880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013841524573721,
      "loss": 1.2778,
      "theoretical_loss": 3.417817990039089,
      "tokens_seen": 2315452416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013640922768305,
      "loss": 1.2598,
      "theoretical_loss": 3.4178102200412748,
      "tokens_seen": 2315517952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030134403209628884,
      "loss": 1.356,
      "theoretical_loss": 3.4178024503249453,
      "tokens_seen": 2315583488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030132397191574725,
      "loss": 1.375,
      "theoretical_loss": 3.4177946808900828,
      "tokens_seen": 2315649024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013039117352056,
      "loss": 1.3849,
      "theoretical_loss": 3.417786911736668,
      "tokens_seen": 2315714560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301283851554664,
      "loss": 1.2668,
      "theoretical_loss": 3.4177791428646835,
      "tokens_seen": 2315780096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030126379137412234,
      "loss": 1.4356,
      "theoretical_loss": 3.4177713742741114,
      "tokens_seen": 2315845632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012437311935807,
      "loss": 1.3242,
      "theoretical_loss": 3.417763605964933,
      "tokens_seen": 2315911168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012236710130391,
      "loss": 1.1946,
      "theoretical_loss": 3.4177558379371304,
      "tokens_seen": 2315976704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012036108324975,
      "loss": 1.3067,
      "theoretical_loss": 3.4177480701906853,
      "tokens_seen": 2316042240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030118355065195584,
      "loss": 1.3339,
      "theoretical_loss": 3.4177403027255795,
      "tokens_seen": 2316107776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011634904714142,
      "loss": 1.358,
      "theoretical_loss": 3.4177325355417953,
      "tokens_seen": 2316173312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011434302908726,
      "loss": 1.3194,
      "theoretical_loss": 3.4177247686393137,
      "tokens_seen": 2316238848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301123370110331,
      "loss": 1.4045,
      "theoretical_loss": 3.417717002018118,
      "tokens_seen": 2316304384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030110330992978935,
      "loss": 1.2267,
      "theoretical_loss": 3.4177092356781884,
      "tokens_seen": 2316369920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003010832497492477,
      "loss": 1.3221,
      "theoretical_loss": 3.417701469619508,
      "tokens_seen": 2316435456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030106318956870607,
      "loss": 1.3191,
      "theoretical_loss": 3.4176937038420574,
      "tokens_seen": 2316500992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030104312938816454,
      "loss": 1.3367,
      "theoretical_loss": 3.41768593834582,
      "tokens_seen": 2316566528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003010230692076229,
      "loss": 1.2884,
      "theoretical_loss": 3.417678173130776,
      "tokens_seen": 2316632064
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030100300902708127,
      "loss": 1.2608,
      "theoretical_loss": 3.417670408196909,
      "tokens_seen": 2316697600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030098294884653963,
      "loss": 1.4313,
      "theoretical_loss": 3.4176626435441992,
      "tokens_seen": 2316763136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030096288866599804,
      "loss": 1.3455,
      "theoretical_loss": 3.41765487917263,
      "tokens_seen": 2316828672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003009428284854564,
      "loss": 1.3167,
      "theoretical_loss": 3.4176471150821826,
      "tokens_seen": 2316894208
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030092276830491477,
      "loss": 1.3157,
      "theoretical_loss": 3.4176393512728382,
      "tokens_seen": 2316959744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030090270812437313,
      "loss": 1.3507,
      "theoretical_loss": 3.4176315877445793,
      "tokens_seen": 2317025280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030088264794383155,
      "loss": 1.3271,
      "theoretical_loss": 3.4176238244973884,
      "tokens_seen": 2317090816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003008625877632899,
      "loss": 1.4025,
      "theoretical_loss": 3.417616061531246,
      "tokens_seen": 2317156352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030084252758274827,
      "loss": 1.285,
      "theoretical_loss": 3.417608298846135,
      "tokens_seen": 2317221888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030082246740220663,
      "loss": 1.3636,
      "theoretical_loss": 3.417600536442037,
      "tokens_seen": 2317287424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000300802407221665,
      "loss": 1.2194,
      "theoretical_loss": 3.4175927743189343,
      "tokens_seen": 2317352960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007823470411234,
      "loss": 1.3609,
      "theoretical_loss": 3.417585012476808,
      "tokens_seen": 2317418496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007622868605818,
      "loss": 1.3221,
      "theoretical_loss": 3.41757725091564,
      "tokens_seen": 2317484032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030074222668004014,
      "loss": 1.3575,
      "theoretical_loss": 3.4175694896354125,
      "tokens_seen": 2317549568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007221664994985,
      "loss": 1.3061,
      "theoretical_loss": 3.417561728636108,
      "tokens_seen": 2317615104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007021063189569,
      "loss": 1.3583,
      "theoretical_loss": 3.4175539679177076,
      "tokens_seen": 2317680640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006820461384153,
      "loss": 1.3085,
      "theoretical_loss": 3.4175462074801928,
      "tokens_seen": 2317746176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030066198595787364,
      "loss": 1.4035,
      "theoretical_loss": 3.4175384473235466,
      "tokens_seen": 2317811712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000300641925777332,
      "loss": 1.4512,
      "theoretical_loss": 3.4175306874477505,
      "tokens_seen": 2317877248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030062186559679036,
      "loss": 1.3859,
      "theoretical_loss": 3.417522927852786,
      "tokens_seen": 2317942784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006018054162488,
      "loss": 1.2733,
      "theoretical_loss": 3.4175151685386354,
      "tokens_seen": 2318008320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030058174523570714,
      "loss": 1.2964,
      "theoretical_loss": 3.4175074095052804,
      "tokens_seen": 2318073856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003005616850551655,
      "loss": 1.3012,
      "theoretical_loss": 3.4174996507527027,
      "tokens_seen": 2318139392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030054162487462387,
      "loss": 1.3892,
      "theoretical_loss": 3.417491892280885,
      "tokens_seen": 2318204928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003005215646940823,
      "loss": 1.3756,
      "theoretical_loss": 3.417484134089808,
      "tokens_seen": 2318270464
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030050150451354065,
      "loss": 1.329,
      "theoretical_loss": 3.417476376179455,
      "tokens_seen": 2318336000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000300481444332999,
      "loss": 1.371,
      "theoretical_loss": 3.4174686185498064,
      "tokens_seen": 2318401536
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 844398,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2924747467041016,
      "objective/train/theoretical_loss": 3.4174608612008455,
      "objective/train/tokens_used": 688926176,
      "theoretical_loss": 3.4174608612008455,
      "tokens_seen": 2318467072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030046138415245737,
      "loss": 1.3545,
      "theoretical_loss": 3.4174608612008455,
      "tokens_seen": 2318467072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030044132397191573,
      "loss": 1.31,
      "theoretical_loss": 3.4174531041325533,
      "tokens_seen": 2318532608
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030042126379137415,
      "loss": 1.4708,
      "theoretical_loss": 3.4174453473449122,
      "tokens_seen": 2318598144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004012036108325,
      "loss": 1.2725,
      "theoretical_loss": 3.4174375908379035,
      "tokens_seen": 2318663680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003811434302909,
      "loss": 1.3348,
      "theoretical_loss": 3.41742983461151,
      "tokens_seen": 2318729216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030036108324974924,
      "loss": 1.2673,
      "theoretical_loss": 3.4174220786657132,
      "tokens_seen": 2318794752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030034102306920765,
      "loss": 1.3144,
      "theoretical_loss": 3.4174143230004947,
      "tokens_seen": 2318860288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000300320962888666,
      "loss": 1.3886,
      "theoretical_loss": 3.4174065676158367,
      "tokens_seen": 2318925824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003009027081244,
      "loss": 1.2858,
      "theoretical_loss": 3.4173988125117214,
      "tokens_seen": 2318991360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030028084252758274,
      "loss": 1.3776,
      "theoretical_loss": 3.41739105768813,
      "tokens_seen": 2319056896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002607823470411,
      "loss": 1.2683,
      "theoretical_loss": 3.417383303145045,
      "tokens_seen": 2319122432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002407221664995,
      "loss": 1.3697,
      "theoretical_loss": 3.4173755488824487,
      "tokens_seen": 2319187968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002206619859579,
      "loss": 1.3529,
      "theoretical_loss": 3.417367794900322,
      "tokens_seen": 2319253504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030020060180541624,
      "loss": 1.2769,
      "theoretical_loss": 3.417360041198647,
      "tokens_seen": 2319319040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001805416248746,
      "loss": 1.271,
      "theoretical_loss": 3.4173522877774065,
      "tokens_seen": 2319384576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000300160481444333,
      "loss": 1.3926,
      "theoretical_loss": 3.417344534636582,
      "tokens_seen": 2319450112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001404212637914,
      "loss": 1.2467,
      "theoretical_loss": 3.417336781776155,
      "tokens_seen": 2319515648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030012036108324975,
      "loss": 1.3824,
      "theoretical_loss": 3.4173290291961083,
      "tokens_seen": 2319581184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001003009027081,
      "loss": 1.3446,
      "theoretical_loss": 3.417321276896423,
      "tokens_seen": 2319646720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030008024072216647,
      "loss": 1.4116,
      "theoretical_loss": 3.4173135248770814,
      "tokens_seen": 2319712256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003000601805416249,
      "loss": 1.2669,
      "theoretical_loss": 3.4173057731380654,
      "tokens_seen": 2319777792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030004012036108325,
      "loss": 1.4348,
      "theoretical_loss": 3.4172980216793567,
      "tokens_seen": 2319843328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003000200601805416,
      "loss": 1.3198,
      "theoretical_loss": 3.417290270500938,
      "tokens_seen": 2319908864
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003,
      "loss": 1.3848,
      "theoretical_loss": 3.4172825196027903,
      "tokens_seen": 2319974400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002999799398194584,
      "loss": 1.4139,
      "theoretical_loss": 3.417274768984896,
      "tokens_seen": 2320039936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029995987963891675,
      "loss": 1.3585,
      "theoretical_loss": 3.4172670186472374,
      "tokens_seen": 2320105472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002999398194583751,
      "loss": 1.312,
      "theoretical_loss": 3.4172592685897962,
      "tokens_seen": 2320171008
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002999197592778335,
      "loss": 1.3837,
      "theoretical_loss": 3.417251518812554,
      "tokens_seen": 2320236544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998996990972919,
      "loss": 1.3585,
      "theoretical_loss": 3.417243769315493,
      "tokens_seen": 2320302080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029987963891675026,
      "loss": 1.4267,
      "theoretical_loss": 3.417236020098595,
      "tokens_seen": 2320367616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998595787362086,
      "loss": 1.1875,
      "theoretical_loss": 3.4172282711618425,
      "tokens_seen": 2320433152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299839518555667,
      "loss": 1.3223,
      "theoretical_loss": 3.417220522505217,
      "tokens_seen": 2320498688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029981945837512534,
      "loss": 1.4236,
      "theoretical_loss": 3.4172127741287004,
      "tokens_seen": 2320564224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029979939819458376,
      "loss": 1.3446,
      "theoretical_loss": 3.4172050260322746,
      "tokens_seen": 2320629760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997793380140421,
      "loss": 1.2907,
      "theoretical_loss": 3.417197278215922,
      "tokens_seen": 2320695296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997592778335005,
      "loss": 1.2979,
      "theoretical_loss": 3.4171895306796243,
      "tokens_seen": 2320760832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029973921765295885,
      "loss": 1.326,
      "theoretical_loss": 3.4171817834233638,
      "tokens_seen": 2320826368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029971915747241726,
      "loss": 1.2653,
      "theoretical_loss": 3.417174036447122,
      "tokens_seen": 2320891904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996990972918756,
      "loss": 1.2683,
      "theoretical_loss": 3.417166289750881,
      "tokens_seen": 2320957440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299679037111334,
      "loss": 1.445,
      "theoretical_loss": 3.417158543334623,
      "tokens_seen": 2321022976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029965897693079235,
      "loss": 1.3592,
      "theoretical_loss": 3.4171507971983295,
      "tokens_seen": 2321088512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996389167502507,
      "loss": 1.2946,
      "theoretical_loss": 3.4171430513419834,
      "tokens_seen": 2321154048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029961885656970913,
      "loss": 1.3561,
      "theoretical_loss": 3.4171353057655653,
      "tokens_seen": 2321219584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995987963891675,
      "loss": 1.422,
      "theoretical_loss": 3.4171275604690585,
      "tokens_seen": 2321285120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029957873620862585,
      "loss": 1.2697,
      "theoretical_loss": 3.417119815452444,
      "tokens_seen": 2321350656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995586760280842,
      "loss": 1.2822,
      "theoretical_loss": 3.4171120707157043,
      "tokens_seen": 2321416192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995386158475427,
      "loss": 1.2487,
      "theoretical_loss": 3.4171043262588214,
      "tokens_seen": 2321481728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029951855566700105,
      "loss": 1.4081,
      "theoretical_loss": 3.4170965820817774,
      "tokens_seen": 2321547264
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002994984954864594,
      "loss": 1.2559,
      "theoretical_loss": 3.417088838184554,
      "tokens_seen": 2321612800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029947843530591777,
      "loss": 1.3436,
      "theoretical_loss": 3.417081094567133,
      "tokens_seen": 2321678336
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 845463,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3060812950134277,
      "objective/train/theoretical_loss": 3.4170733512294964,
      "objective/train/tokens_used": 692202976,
      "theoretical_loss": 3.4170733512294964,
      "tokens_seen": 2321743872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029945837512537613,
      "loss": 1.287,
      "theoretical_loss": 3.4170733512294964,
      "tokens_seen": 2321743872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029943831494483455,
      "loss": 1.226,
      "theoretical_loss": 3.4170656081716273,
      "tokens_seen": 2321809408
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002994182547642929,
      "loss": 1.2837,
      "theoretical_loss": 3.417057865393506,
      "tokens_seen": 2321874944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993981945837513,
      "loss": 1.3648,
      "theoretical_loss": 3.417050122895116,
      "tokens_seen": 2321940480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029937813440320964,
      "loss": 1.2569,
      "theoretical_loss": 3.4170423806764383,
      "tokens_seen": 2322006016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029935807422266805,
      "loss": 1.2698,
      "theoretical_loss": 3.417034638737455,
      "tokens_seen": 2322071552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993380140421264,
      "loss": 1.3053,
      "theoretical_loss": 3.417026897078149,
      "tokens_seen": 2322137088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993179538615848,
      "loss": 1.3126,
      "theoretical_loss": 3.4170191556985015,
      "tokens_seen": 2322202624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029929789368104314,
      "loss": 1.4164,
      "theoretical_loss": 3.4170114145984942,
      "tokens_seen": 2322268160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992778335005015,
      "loss": 1.3429,
      "theoretical_loss": 3.41700367377811,
      "tokens_seen": 2322333696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992577733199599,
      "loss": 1.339,
      "theoretical_loss": 3.41699593323733,
      "tokens_seen": 2322399232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992377131394183,
      "loss": 1.3387,
      "theoretical_loss": 3.416988192976137,
      "tokens_seen": 2322464768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029921765295887664,
      "loss": 1.1831,
      "theoretical_loss": 3.4169804529945127,
      "tokens_seen": 2322530304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299197592778335,
      "loss": 1.3474,
      "theoretical_loss": 3.4169727132924392,
      "tokens_seen": 2322595840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991775325977934,
      "loss": 1.406,
      "theoretical_loss": 3.416964973869898,
      "tokens_seen": 2322661376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991574724172518,
      "loss": 1.3032,
      "theoretical_loss": 3.416957234726872,
      "tokens_seen": 2322726912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029913741223671015,
      "loss": 1.3625,
      "theoretical_loss": 3.4169494958633426,
      "tokens_seen": 2322792448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991173520561685,
      "loss": 1.5022,
      "theoretical_loss": 3.416941757279292,
      "tokens_seen": 2322857984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029909729187562687,
      "loss": 1.4387,
      "theoretical_loss": 3.416934018974702,
      "tokens_seen": 2322923520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002990772316950853,
      "loss": 1.3277,
      "theoretical_loss": 3.4169262809495553,
      "tokens_seen": 2322989056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029905717151454365,
      "loss": 1.488,
      "theoretical_loss": 3.4169185432038334,
      "tokens_seen": 2323054592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299037111334002,
      "loss": 1.4867,
      "theoretical_loss": 3.416910805737518,
      "tokens_seen": 2323120128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002990170511534604,
      "loss": 1.2783,
      "theoretical_loss": 3.4169030685505914,
      "tokens_seen": 2323185664
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989969909729188,
      "loss": 1.3284,
      "theoretical_loss": 3.4168953316430364,
      "tokens_seen": 2323251200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029897693079237715,
      "loss": 1.2179,
      "theoretical_loss": 3.416887595014834,
      "tokens_seen": 2323316736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989568706118355,
      "loss": 1.3086,
      "theoretical_loss": 3.4168798586659666,
      "tokens_seen": 2323382272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989368104312939,
      "loss": 1.2476,
      "theoretical_loss": 3.4168721225964163,
      "tokens_seen": 2323447808
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989167502507523,
      "loss": 1.3361,
      "theoretical_loss": 3.4168643868061648,
      "tokens_seen": 2323513344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029889669007021066,
      "loss": 1.339,
      "theoretical_loss": 3.416856651295195,
      "tokens_seen": 2323578880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298876629889669,
      "loss": 1.2408,
      "theoretical_loss": 3.4168489160634876,
      "tokens_seen": 2323644416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988565697091274,
      "loss": 1.332,
      "theoretical_loss": 3.4168411811110255,
      "tokens_seen": 2323709952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029883650952858574,
      "loss": 1.3044,
      "theoretical_loss": 3.416833446437791,
      "tokens_seen": 2323775488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029881644934804416,
      "loss": 1.3418,
      "theoretical_loss": 3.416825712043766,
      "tokens_seen": 2323841024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987963891675025,
      "loss": 1.3042,
      "theoretical_loss": 3.416817977928932,
      "tokens_seen": 2323906560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987763289869609,
      "loss": 1.344,
      "theoretical_loss": 3.4168102440932717,
      "tokens_seen": 2323972096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029875626880641925,
      "loss": 1.338,
      "theoretical_loss": 3.4168025105367663,
      "tokens_seen": 2324037632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029873620862587766,
      "loss": 1.3089,
      "theoretical_loss": 3.4167947772593985,
      "tokens_seen": 2324103168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298716148445336,
      "loss": 1.3484,
      "theoretical_loss": 3.4167870442611505,
      "tokens_seen": 2324168704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986960882647944,
      "loss": 1.3548,
      "theoretical_loss": 3.416779311542004,
      "tokens_seen": 2324234240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029867602808425275,
      "loss": 1.2672,
      "theoretical_loss": 3.4167715791019413,
      "tokens_seen": 2324299776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986559679037111,
      "loss": 1.3502,
      "theoretical_loss": 3.4167638469409445,
      "tokens_seen": 2324365312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029863590772316953,
      "loss": 1.3384,
      "theoretical_loss": 3.4167561150589947,
      "tokens_seen": 2324430848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986158475426279,
      "loss": 1.2103,
      "theoretical_loss": 3.4167483834560755,
      "tokens_seen": 2324496384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029859578736208625,
      "loss": 1.327,
      "theoretical_loss": 3.4167406521321677,
      "tokens_seen": 2324561920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985757271815446,
      "loss": 1.2901,
      "theoretical_loss": 3.416732921087254,
      "tokens_seen": 2324627456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029855566700100303,
      "loss": 1.2901,
      "theoretical_loss": 3.416725190321317,
      "tokens_seen": 2324692992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985356068204614,
      "loss": 1.4099,
      "theoretical_loss": 3.416717459834337,
      "tokens_seen": 2324758528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029851554663991976,
      "loss": 1.2924,
      "theoretical_loss": 3.416709729626298,
      "tokens_seen": 2324824064
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984954864593781,
      "loss": 1.3037,
      "theoretical_loss": 3.416701999697181,
      "tokens_seen": 2324889600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984754262788365,
      "loss": 1.3325,
      "theoretical_loss": 3.416694270046968,
      "tokens_seen": 2324955136
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 846164,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3871805667877197,
      "objective/train/theoretical_loss": 3.4166865406756415,
      "objective/train/tokens_used": 695479776,
      "theoretical_loss": 3.4166865406756415,
      "tokens_seen": 2325020672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984553660982949,
      "loss": 1.3233,
      "theoretical_loss": 3.4166865406756415,
      "tokens_seen": 2325020672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029843530591775326,
      "loss": 1.4494,
      "theoretical_loss": 3.416678811583184,
      "tokens_seen": 2325086208
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984152457372116,
      "loss": 1.2388,
      "theoretical_loss": 3.4166710827695765,
      "tokens_seen": 2325151744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029839518555667,
      "loss": 1.3266,
      "theoretical_loss": 3.416663354234802,
      "tokens_seen": 2325217280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983751253761284,
      "loss": 1.2436,
      "theoretical_loss": 3.416655625978842,
      "tokens_seen": 2325282816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029835506519558676,
      "loss": 1.2568,
      "theoretical_loss": 3.4166478980016786,
      "tokens_seen": 2325348352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983350050150451,
      "loss": 1.2759,
      "theoretical_loss": 3.4166401703032943,
      "tokens_seen": 2325413888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983149448345035,
      "loss": 1.3686,
      "theoretical_loss": 3.416632442883671,
      "tokens_seen": 2325479424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029829488465396185,
      "loss": 1.2813,
      "theoretical_loss": 3.416624715742791,
      "tokens_seen": 2325544960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029827482447342026,
      "loss": 1.3076,
      "theoretical_loss": 3.416616988880636,
      "tokens_seen": 2325610496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002982547642928786,
      "loss": 1.3844,
      "theoretical_loss": 3.416609262297188,
      "tokens_seen": 2325676032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298234704112337,
      "loss": 1.3185,
      "theoretical_loss": 3.4166015359924296,
      "tokens_seen": 2325741568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029821464393179535,
      "loss": 1.2477,
      "theoretical_loss": 3.416593809966342,
      "tokens_seen": 2325807104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029819458375125377,
      "loss": 1.3701,
      "theoretical_loss": 3.416586084218909,
      "tokens_seen": 2325872640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029817452357071213,
      "loss": 1.3564,
      "theoretical_loss": 3.416578358750111,
      "tokens_seen": 2325938176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981544633901705,
      "loss": 1.3666,
      "theoretical_loss": 3.416570633559931,
      "tokens_seen": 2326003712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029813440320962885,
      "loss": 1.2992,
      "theoretical_loss": 3.41656290864835,
      "tokens_seen": 2326069248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981143430290872,
      "loss": 1.2307,
      "theoretical_loss": 3.416555184015352,
      "tokens_seen": 2326134784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029809428284854563,
      "loss": 1.4657,
      "theoretical_loss": 3.4165474596609173,
      "tokens_seen": 2326200320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298074222668004,
      "loss": 1.3357,
      "theoretical_loss": 3.416539735585029,
      "tokens_seen": 2326265856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029805416248746236,
      "loss": 1.2444,
      "theoretical_loss": 3.416532011787669,
      "tokens_seen": 2326331392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002980341023069208,
      "loss": 1.2933,
      "theoretical_loss": 3.4165242882688194,
      "tokens_seen": 2326396928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002980140421263792,
      "loss": 1.2904,
      "theoretical_loss": 3.4165165650284623,
      "tokens_seen": 2326462464
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029799398194583755,
      "loss": 1.3447,
      "theoretical_loss": 3.4165088420665795,
      "tokens_seen": 2326528000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979739217652959,
      "loss": 1.3363,
      "theoretical_loss": 3.416501119383154,
      "tokens_seen": 2326593536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979538615847543,
      "loss": 1.2475,
      "theoretical_loss": 3.416493396978167,
      "tokens_seen": 2326659072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979338014042127,
      "loss": 1.3534,
      "theoretical_loss": 3.4164856748516006,
      "tokens_seen": 2326724608
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029791374122367106,
      "loss": 1.1943,
      "theoretical_loss": 3.4164779530034375,
      "tokens_seen": 2326790144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978936810431294,
      "loss": 1.3433,
      "theoretical_loss": 3.4164702314336597,
      "tokens_seen": 2326855680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978736208625878,
      "loss": 1.3027,
      "theoretical_loss": 3.416462510142249,
      "tokens_seen": 2326921216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029785356068204614,
      "loss": 1.3255,
      "theoretical_loss": 3.4164547891291877,
      "tokens_seen": 2326986752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029783350050150456,
      "loss": 1.2712,
      "theoretical_loss": 3.4164470683944583,
      "tokens_seen": 2327052288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978134403209629,
      "loss": 1.3842,
      "theoretical_loss": 3.4164393479380424,
      "tokens_seen": 2327117824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977933801404213,
      "loss": 1.3272,
      "theoretical_loss": 3.4164316277599225,
      "tokens_seen": 2327183360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029777331995987965,
      "loss": 1.3929,
      "theoretical_loss": 3.4164239078600804,
      "tokens_seen": 2327248896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029775325977933806,
      "loss": 1.275,
      "theoretical_loss": 3.416416188238498,
      "tokens_seen": 2327314432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977331995987964,
      "loss": 1.3618,
      "theoretical_loss": 3.4164084688951584,
      "tokens_seen": 2327379968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977131394182548,
      "loss": 1.3126,
      "theoretical_loss": 3.416400749830043,
      "tokens_seen": 2327445504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029769307923771315,
      "loss": 1.2169,
      "theoretical_loss": 3.4163930310431336,
      "tokens_seen": 2327511040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976730190571715,
      "loss": 1.3784,
      "theoretical_loss": 3.4163853125344135,
      "tokens_seen": 2327576576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029765295887662993,
      "loss": 1.255,
      "theoretical_loss": 3.4163775943038637,
      "tokens_seen": 2327642112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976328986960883,
      "loss": 1.2552,
      "theoretical_loss": 3.4163698763514674,
      "tokens_seen": 2327707648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029761283851554665,
      "loss": 1.3461,
      "theoretical_loss": 3.4163621586772055,
      "tokens_seen": 2327773184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000297592778335005,
      "loss": 1.3109,
      "theoretical_loss": 3.416354441281061,
      "tokens_seen": 2327838720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029757271815446343,
      "loss": 1.2444,
      "theoretical_loss": 3.416346724163016,
      "tokens_seen": 2327904256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002975526579739218,
      "loss": 1.3298,
      "theoretical_loss": 3.416339007323052,
      "tokens_seen": 2327969792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029753259779338016,
      "loss": 1.3681,
      "theoretical_loss": 3.4163312907611525,
      "tokens_seen": 2328035328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002975125376128385,
      "loss": 1.3533,
      "theoretical_loss": 3.4163235744772984,
      "tokens_seen": 2328100864
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974924774322969,
      "loss": 1.3243,
      "theoretical_loss": 3.416315858471472,
      "tokens_seen": 2328166400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974724172517553,
      "loss": 1.2615,
      "theoretical_loss": 3.4163081427436564,
      "tokens_seen": 2328231936
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 847468,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.23801851272583,
      "objective/train/theoretical_loss": 3.4163004272938324,
      "objective/train/tokens_used": 698756576,
      "theoretical_loss": 3.4163004272938324,
      "tokens_seen": 2328297472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029745235707121366,
      "loss": 1.2983,
      "theoretical_loss": 3.4163004272938324,
      "tokens_seen": 2328297472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000297432296890672,
      "loss": 1.322,
      "theoretical_loss": 3.416292712121983,
      "tokens_seen": 2328363008
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974122367101304,
      "loss": 1.2374,
      "theoretical_loss": 3.4162849972280904,
      "tokens_seen": 2328428544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973921765295888,
      "loss": 1.3498,
      "theoretical_loss": 3.416277282612137,
      "tokens_seen": 2328494080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029737211634904716,
      "loss": 1.3068,
      "theoretical_loss": 3.4162695682741036,
      "tokens_seen": 2328559616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973520561685055,
      "loss": 1.3514,
      "theoretical_loss": 3.4162618542139738,
      "tokens_seen": 2328625152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973319959879639,
      "loss": 1.3814,
      "theoretical_loss": 3.416254140431729,
      "tokens_seen": 2328690688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029731193580742225,
      "loss": 1.3495,
      "theoretical_loss": 3.416246426927352,
      "tokens_seen": 2328756224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029729187562688067,
      "loss": 1.2906,
      "theoretical_loss": 3.416238713700824,
      "tokens_seen": 2328821760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029727181544633903,
      "loss": 1.3702,
      "theoretical_loss": 3.4162310007521284,
      "tokens_seen": 2328887296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002972517552657974,
      "loss": 1.3749,
      "theoretical_loss": 3.4162232880812464,
      "tokens_seen": 2328952832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029723169508525575,
      "loss": 1.3128,
      "theoretical_loss": 3.4162155756881605,
      "tokens_seen": 2329018368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029721163490471417,
      "loss": 1.4558,
      "theoretical_loss": 3.416207863572853,
      "tokens_seen": 2329083904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029719157472417253,
      "loss": 1.3659,
      "theoretical_loss": 3.4162001517353064,
      "tokens_seen": 2329149440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971715145436309,
      "loss": 1.3827,
      "theoretical_loss": 3.416192440175502,
      "tokens_seen": 2329214976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029715145436308926,
      "loss": 1.3374,
      "theoretical_loss": 3.4161847288934224,
      "tokens_seen": 2329280512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971313941825476,
      "loss": 1.3199,
      "theoretical_loss": 3.41617701788905,
      "tokens_seen": 2329346048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029711133400200603,
      "loss": 1.3217,
      "theoretical_loss": 3.4161693071623667,
      "tokens_seen": 2329411584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970912738214644,
      "loss": 1.2676,
      "theoretical_loss": 3.416161596713355,
      "tokens_seen": 2329477120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029707121364092276,
      "loss": 1.2554,
      "theoretical_loss": 3.4161538865419967,
      "tokens_seen": 2329542656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970511534603811,
      "loss": 1.3127,
      "theoretical_loss": 3.416146176648274,
      "tokens_seen": 2329608192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029703109327983954,
      "loss": 1.1493,
      "theoretical_loss": 3.41613846703217,
      "tokens_seen": 2329673728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970110330992979,
      "loss": 1.3455,
      "theoretical_loss": 3.416130757693666,
      "tokens_seen": 2329739264
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029699097291875626,
      "loss": 1.2408,
      "theoretical_loss": 3.416123048632744,
      "tokens_seen": 2329804800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002969709127382146,
      "loss": 1.3025,
      "theoretical_loss": 3.4161153398493864,
      "tokens_seen": 2329870336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296950852557673,
      "loss": 1.3278,
      "theoretical_loss": 3.416107631343576,
      "tokens_seen": 2329935872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002969307923771314,
      "loss": 1.2766,
      "theoretical_loss": 3.4160999231152944,
      "tokens_seen": 2330001408
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029691073219658976,
      "loss": 1.4089,
      "theoretical_loss": 3.416092215164524,
      "tokens_seen": 2330066944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968906720160481,
      "loss": 1.4085,
      "theoretical_loss": 3.416084507491247,
      "tokens_seen": 2330132480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968706118355065,
      "loss": 1.2834,
      "theoretical_loss": 3.4160768000954453,
      "tokens_seen": 2330198016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968505516549649,
      "loss": 1.2586,
      "theoretical_loss": 3.416069092977102,
      "tokens_seen": 2330263552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029683049147442327,
      "loss": 1.3057,
      "theoretical_loss": 3.4160613861361977,
      "tokens_seen": 2330329088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029681043129388163,
      "loss": 1.2852,
      "theoretical_loss": 3.416053679572716,
      "tokens_seen": 2330394624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029679037111334,
      "loss": 1.3628,
      "theoretical_loss": 3.4160459732866393,
      "tokens_seen": 2330460160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967703109327984,
      "loss": 1.3531,
      "theoretical_loss": 3.4160382672779486,
      "tokens_seen": 2330525696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029675025075225677,
      "loss": 1.2805,
      "theoretical_loss": 3.416030561546627,
      "tokens_seen": 2330591232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029673019057171513,
      "loss": 1.2983,
      "theoretical_loss": 3.4160228560926567,
      "tokens_seen": 2330656768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967101303911735,
      "loss": 1.2896,
      "theoretical_loss": 3.4160151509160195,
      "tokens_seen": 2330722304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029669007021063186,
      "loss": 1.4156,
      "theoretical_loss": 3.416007446016698,
      "tokens_seen": 2330787840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966700100300903,
      "loss": 1.3154,
      "theoretical_loss": 3.415999741394674,
      "tokens_seen": 2330853376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029664994984954864,
      "loss": 1.2985,
      "theoretical_loss": 3.4159920370499295,
      "tokens_seen": 2330918912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296629889669007,
      "loss": 1.3793,
      "theoretical_loss": 3.4159843329824477,
      "tokens_seen": 2330984448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029660982948846536,
      "loss": 1.2832,
      "theoretical_loss": 3.41597662919221,
      "tokens_seen": 2331049984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965897693079238,
      "loss": 1.2897,
      "theoretical_loss": 3.4159689256791994,
      "tokens_seen": 2331115520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029656970912738214,
      "loss": 1.2209,
      "theoretical_loss": 3.4159612224433973,
      "tokens_seen": 2331181056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965496489468405,
      "loss": 1.3581,
      "theoretical_loss": 3.4159535194847868,
      "tokens_seen": 2331246592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965295887662989,
      "loss": 1.272,
      "theoretical_loss": 3.4159458168033487,
      "tokens_seen": 2331312128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965095285857573,
      "loss": 1.4042,
      "theoretical_loss": 3.415938114399067,
      "tokens_seen": 2331377664
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964894684052157,
      "loss": 1.319,
      "theoretical_loss": 3.4159304122719227,
      "tokens_seen": 2331443200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029646940822467406,
      "loss": 1.2634,
      "theoretical_loss": 3.4159227104218983,
      "tokens_seen": 2331508736
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 848727,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.312225580215454,
      "objective/train/theoretical_loss": 3.4159150088489767,
      "objective/train/tokens_used": 702033376,
      "theoretical_loss": 3.4159150088489767,
      "tokens_seen": 2331574272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964493480441324,
      "loss": 1.3079,
      "theoretical_loss": 3.4159150088489767,
      "tokens_seen": 2331574272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964292878635908,
      "loss": 1.2808,
      "theoretical_loss": 3.415907307553139,
      "tokens_seen": 2331639808
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964092276830492,
      "loss": 1.4983,
      "theoretical_loss": 3.4158996065343685,
      "tokens_seen": 2331705344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029638916750250756,
      "loss": 1.2482,
      "theoretical_loss": 3.4158919057926473,
      "tokens_seen": 2331770880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963691073219659,
      "loss": 1.362,
      "theoretical_loss": 3.4158842053279566,
      "tokens_seen": 2331836416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963490471414243,
      "loss": 1.2402,
      "theoretical_loss": 3.4158765051402797,
      "tokens_seen": 2331901952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029632898696088265,
      "loss": 1.2778,
      "theoretical_loss": 3.4158688052295987,
      "tokens_seen": 2331967488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029630892678034107,
      "loss": 1.2674,
      "theoretical_loss": 3.4158611055958956,
      "tokens_seen": 2332033024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029628886659979943,
      "loss": 1.3571,
      "theoretical_loss": 3.415853406239153,
      "tokens_seen": 2332098560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002962688064192578,
      "loss": 1.308,
      "theoretical_loss": 3.4158457071593524,
      "tokens_seen": 2332164096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029624874623871615,
      "loss": 1.3195,
      "theoretical_loss": 3.415838008356477,
      "tokens_seen": 2332229632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029622868605817457,
      "loss": 1.3108,
      "theoretical_loss": 3.4158303098305085,
      "tokens_seen": 2332295168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029620862587763293,
      "loss": 1.3413,
      "theoretical_loss": 3.4158226115814294,
      "tokens_seen": 2332360704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002961885656970913,
      "loss": 1.2903,
      "theoretical_loss": 3.4158149136092217,
      "tokens_seen": 2332426240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029616850551654966,
      "loss": 1.3983,
      "theoretical_loss": 3.415807215913868,
      "tokens_seen": 2332491776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296148445336008,
      "loss": 1.2348,
      "theoretical_loss": 3.4157995184953505,
      "tokens_seen": 2332557312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029612838515546643,
      "loss": 1.249,
      "theoretical_loss": 3.415791821353651,
      "tokens_seen": 2332622848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002961083249749248,
      "loss": 1.3836,
      "theoretical_loss": 3.415784124488752,
      "tokens_seen": 2332688384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029608826479438316,
      "loss": 1.3286,
      "theoretical_loss": 3.415776427900636,
      "tokens_seen": 2332753920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002960682046138415,
      "loss": 1.2731,
      "theoretical_loss": 3.415768731589285,
      "tokens_seen": 2332819456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029604814443329994,
      "loss": 1.2743,
      "theoretical_loss": 3.4157610355546817,
      "tokens_seen": 2332884992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002960280842527583,
      "loss": 1.2831,
      "theoretical_loss": 3.415753339796808,
      "tokens_seen": 2332950528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029600802407221666,
      "loss": 1.412,
      "theoretical_loss": 3.4157456443156464,
      "tokens_seen": 2333016064
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000295987963891675,
      "loss": 1.3577,
      "theoretical_loss": 3.415737949111179,
      "tokens_seen": 2333081600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002959679037111334,
      "loss": 1.3327,
      "theoretical_loss": 3.4157302541833885,
      "tokens_seen": 2333147136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002959478435305918,
      "loss": 1.3875,
      "theoretical_loss": 3.4157225595322562,
      "tokens_seen": 2333212672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029592778335005016,
      "loss": 1.307,
      "theoretical_loss": 3.4157148651577653,
      "tokens_seen": 2333278208
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029590772316950853,
      "loss": 1.3359,
      "theoretical_loss": 3.4157071710598976,
      "tokens_seen": 2333343744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958876629889669,
      "loss": 1.324,
      "theoretical_loss": 3.4156994772386358,
      "tokens_seen": 2333409280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958676028084253,
      "loss": 1.3911,
      "theoretical_loss": 3.415691783693962,
      "tokens_seen": 2333474816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029584754262788367,
      "loss": 1.3705,
      "theoretical_loss": 3.4156840904258585,
      "tokens_seen": 2333540352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029582748244734203,
      "loss": 1.2977,
      "theoretical_loss": 3.415676397434307,
      "tokens_seen": 2333605888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958074222668004,
      "loss": 1.4023,
      "theoretical_loss": 3.4156687047192906,
      "tokens_seen": 2333671424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957873620862588,
      "loss": 1.3715,
      "theoretical_loss": 3.4156610122807916,
      "tokens_seen": 2333736960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029576730190571717,
      "loss": 1.3239,
      "theoretical_loss": 3.4156533201187917,
      "tokens_seen": 2333802496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029574724172517553,
      "loss": 1.3011,
      "theoretical_loss": 3.415645628233274,
      "tokens_seen": 2333868032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957271815446339,
      "loss": 1.3678,
      "theoretical_loss": 3.4156379366242193,
      "tokens_seen": 2333933568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029570712136409226,
      "loss": 1.2218,
      "theoretical_loss": 3.415630245291612,
      "tokens_seen": 2333999104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956870611835507,
      "loss": 1.3398,
      "theoretical_loss": 3.4156225542354326,
      "tokens_seen": 2334064640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029566700100300904,
      "loss": 1.2375,
      "theoretical_loss": 3.4156148634556645,
      "tokens_seen": 2334130176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956469408224674,
      "loss": 1.335,
      "theoretical_loss": 3.4156071729522894,
      "tokens_seen": 2334195712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029562688064192576,
      "loss": 1.2781,
      "theoretical_loss": 3.4155994827252902,
      "tokens_seen": 2334261248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956068204613842,
      "loss": 1.2541,
      "theoretical_loss": 3.4155917927746486,
      "tokens_seen": 2334326784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029558676028084254,
      "loss": 1.3723,
      "theoretical_loss": 3.415584103100347,
      "tokens_seen": 2334392320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955667001003009,
      "loss": 1.3626,
      "theoretical_loss": 3.415576413702368,
      "tokens_seen": 2334457856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029554663991975926,
      "loss": 1.2342,
      "theoretical_loss": 3.415568724580694,
      "tokens_seen": 2334523392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955265797392176,
      "loss": 1.4232,
      "theoretical_loss": 3.415561035735307,
      "tokens_seen": 2334588928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029550651955867604,
      "loss": 1.34,
      "theoretical_loss": 3.415553347166189,
      "tokens_seen": 2334654464
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002954864593781344,
      "loss": 1.3426,
      "theoretical_loss": 3.4155456588733233,
      "tokens_seen": 2334720000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029546639919759277,
      "loss": 1.3654,
      "theoretical_loss": 3.4155379708566915,
      "tokens_seen": 2334785536
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 849886,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3592042922973633,
      "objective/train/theoretical_loss": 3.4155302831162757,
      "objective/train/tokens_used": 705310176,
      "theoretical_loss": 3.4155302831162757,
      "tokens_seen": 2334851072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029544633901705113,
      "loss": 1.3117,
      "theoretical_loss": 3.4155302831162757,
      "tokens_seen": 2334851072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029542627883650955,
      "loss": 1.398,
      "theoretical_loss": 3.415522595652059,
      "tokens_seen": 2334916608
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002954062186559679,
      "loss": 1.3174,
      "theoretical_loss": 3.415514908464023,
      "tokens_seen": 2334982144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029538615847542627,
      "loss": 1.2521,
      "theoretical_loss": 3.4155072215521507,
      "tokens_seen": 2335047680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029536609829488463,
      "loss": 1.1583,
      "theoretical_loss": 3.4154995349164237,
      "tokens_seen": 2335113216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000295346038114343,
      "loss": 1.433,
      "theoretical_loss": 3.415491848556825,
      "tokens_seen": 2335178752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953259779338014,
      "loss": 1.4057,
      "theoretical_loss": 3.4154841624733367,
      "tokens_seen": 2335244288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953059177532598,
      "loss": 1.4043,
      "theoretical_loss": 3.415476476665941,
      "tokens_seen": 2335309824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029528585757271814,
      "loss": 1.3841,
      "theoretical_loss": 3.41546879113462,
      "tokens_seen": 2335375360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952657973921765,
      "loss": 1.3145,
      "theoretical_loss": 3.415461105879357,
      "tokens_seen": 2335440896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952457372116349,
      "loss": 1.3383,
      "theoretical_loss": 3.415453420900133,
      "tokens_seen": 2335506432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952256770310933,
      "loss": 1.2907,
      "theoretical_loss": 3.4154457361969315,
      "tokens_seen": 2335571968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029520561685055164,
      "loss": 1.2641,
      "theoretical_loss": 3.4154380517697343,
      "tokens_seen": 2335637504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029518555667001,
      "loss": 1.387,
      "theoretical_loss": 3.4154303676185234,
      "tokens_seen": 2335703040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029516549648946836,
      "loss": 1.2435,
      "theoretical_loss": 3.415422683743282,
      "tokens_seen": 2335768576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951454363089268,
      "loss": 1.338,
      "theoretical_loss": 3.415415000143992,
      "tokens_seen": 2335834112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029512537612838514,
      "loss": 1.3855,
      "theoretical_loss": 3.415407316820635,
      "tokens_seen": 2335899648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951053159478435,
      "loss": 1.3469,
      "theoretical_loss": 3.4153996337731947,
      "tokens_seen": 2335965184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029508525576730187,
      "loss": 1.3466,
      "theoretical_loss": 3.4153919510016526,
      "tokens_seen": 2336030720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950651955867603,
      "loss": 1.3578,
      "theoretical_loss": 3.4153842685059916,
      "tokens_seen": 2336096256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029504513540621865,
      "loss": 1.2296,
      "theoretical_loss": 3.415376586286194,
      "tokens_seen": 2336161792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029502507522567706,
      "loss": 1.28,
      "theoretical_loss": 3.415368904342241,
      "tokens_seen": 2336227328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950050150451354,
      "loss": 1.2837,
      "theoretical_loss": 3.415361222674117,
      "tokens_seen": 2336292864
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002949849548645938,
      "loss": 1.2328,
      "theoretical_loss": 3.415353541281802,
      "tokens_seen": 2336358400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002949648946840522,
      "loss": 1.3,
      "theoretical_loss": 3.4153458601652806,
      "tokens_seen": 2336423936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029494483450351057,
      "loss": 1.3962,
      "theoretical_loss": 3.4153381793245337,
      "tokens_seen": 2336489472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029492477432296893,
      "loss": 1.3448,
      "theoretical_loss": 3.415330498759544,
      "tokens_seen": 2336555008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002949047141424273,
      "loss": 1.3352,
      "theoretical_loss": 3.415322818470294,
      "tokens_seen": 2336620544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948846539618857,
      "loss": 1.3367,
      "theoretical_loss": 3.4153151384567666,
      "tokens_seen": 2336686080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029486459378134407,
      "loss": 1.376,
      "theoretical_loss": 3.415307458718943,
      "tokens_seen": 2336751616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029484453360080243,
      "loss": 1.3997,
      "theoretical_loss": 3.4152997792568063,
      "tokens_seen": 2336817152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948244734202608,
      "loss": 1.2599,
      "theoretical_loss": 3.415292100070339,
      "tokens_seen": 2336882688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948044132397192,
      "loss": 1.3228,
      "theoretical_loss": 3.415284421159523,
      "tokens_seen": 2336948224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029478435305917757,
      "loss": 1.3112,
      "theoretical_loss": 3.415276742524341,
      "tokens_seen": 2337013760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029476429287863593,
      "loss": 1.3172,
      "theoretical_loss": 3.415269064164775,
      "tokens_seen": 2337079296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947442326980943,
      "loss": 1.2969,
      "theoretical_loss": 3.4152613860808083,
      "tokens_seen": 2337144832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029472417251755266,
      "loss": 1.3242,
      "theoretical_loss": 3.415253708272422,
      "tokens_seen": 2337210368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947041123370111,
      "loss": 1.4255,
      "theoretical_loss": 3.4152460307395995,
      "tokens_seen": 2337275904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029468405215646944,
      "loss": 1.311,
      "theoretical_loss": 3.415238353482323,
      "tokens_seen": 2337341440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946639919759278,
      "loss": 1.354,
      "theoretical_loss": 3.415230676500574,
      "tokens_seen": 2337406976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029464393179538616,
      "loss": 1.4006,
      "theoretical_loss": 3.415222999794336,
      "tokens_seen": 2337472512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946238716148446,
      "loss": 1.3561,
      "theoretical_loss": 3.415215323363591,
      "tokens_seen": 2337538048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029460381143430294,
      "loss": 1.2781,
      "theoretical_loss": 3.4152076472083213,
      "tokens_seen": 2337603584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945837512537613,
      "loss": 1.3327,
      "theoretical_loss": 3.4151999713285095,
      "tokens_seen": 2337669120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029456369107321966,
      "loss": 1.1889,
      "theoretical_loss": 3.4151922957241374,
      "tokens_seen": 2337734656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294543630892678,
      "loss": 1.2396,
      "theoretical_loss": 3.415184620395188,
      "tokens_seen": 2337800192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029452357071213644,
      "loss": 1.2906,
      "theoretical_loss": 3.415176945341644,
      "tokens_seen": 2337865728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945035105315948,
      "loss": 1.3469,
      "theoretical_loss": 3.415169270563487,
      "tokens_seen": 2337931264
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029448345035105317,
      "loss": 1.4277,
      "theoretical_loss": 3.4151615960607,
      "tokens_seen": 2337996800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029446339017051153,
      "loss": 1.3448,
      "theoretical_loss": 3.4151539218332645,
      "tokens_seen": 2338062336
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 851124,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.309006690979004,
      "objective/train/theoretical_loss": 3.4151462478811636,
      "objective/train/tokens_used": 708586976,
      "theoretical_loss": 3.4151462478811636,
      "tokens_seen": 2338127872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029444332998996995,
      "loss": 1.3728,
      "theoretical_loss": 3.4151462478811636,
      "tokens_seen": 2338127872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944232698094283,
      "loss": 1.3859,
      "theoretical_loss": 3.4151385742043803,
      "tokens_seen": 2338193408
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029440320962888667,
      "loss": 1.2727,
      "theoretical_loss": 3.415130900802896,
      "tokens_seen": 2338258944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029438314944834503,
      "loss": 1.2918,
      "theoretical_loss": 3.4151232276766934,
      "tokens_seen": 2338324480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943630892678034,
      "loss": 1.3248,
      "theoretical_loss": 3.415115554825755,
      "tokens_seen": 2338390016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943430290872618,
      "loss": 1.1524,
      "theoretical_loss": 3.415107882250063,
      "tokens_seen": 2338455552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943229689067202,
      "loss": 1.2893,
      "theoretical_loss": 3.4151002099496,
      "tokens_seen": 2338521088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029430290872617854,
      "loss": 1.4178,
      "theoretical_loss": 3.4150925379243486,
      "tokens_seen": 2338586624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942828485456369,
      "loss": 1.3956,
      "theoretical_loss": 3.415084866174291,
      "tokens_seen": 2338652160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942627883650953,
      "loss": 1.3561,
      "theoretical_loss": 3.4150771946994096,
      "tokens_seen": 2338717696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942427281845537,
      "loss": 1.338,
      "theoretical_loss": 3.415069523499687,
      "tokens_seen": 2338783232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029422266800401204,
      "loss": 1.2536,
      "theoretical_loss": 3.4150618525751053,
      "tokens_seen": 2338848768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942026078234704,
      "loss": 1.4008,
      "theoretical_loss": 3.4150541819256475,
      "tokens_seen": 2338914304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029418254764292876,
      "loss": 1.2825,
      "theoretical_loss": 3.415046511551295,
      "tokens_seen": 2338979840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002941624874623872,
      "loss": 1.2175,
      "theoretical_loss": 3.415038841452031,
      "tokens_seen": 2339045376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029414242728184554,
      "loss": 1.2598,
      "theoretical_loss": 3.415031171627838,
      "tokens_seen": 2339110912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002941223671013039,
      "loss": 1.2531,
      "theoretical_loss": 3.4150235020786983,
      "tokens_seen": 2339176448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029410230692076227,
      "loss": 1.248,
      "theoretical_loss": 3.4150158328045936,
      "tokens_seen": 2339241984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002940822467402207,
      "loss": 1.3565,
      "theoretical_loss": 3.4150081638055076,
      "tokens_seen": 2339307520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029406218655967905,
      "loss": 1.3412,
      "theoretical_loss": 3.415000495081422,
      "tokens_seen": 2339373056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002940421263791374,
      "loss": 1.3406,
      "theoretical_loss": 3.414992826632319,
      "tokens_seen": 2339438592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029402206619859577,
      "loss": 1.5103,
      "theoretical_loss": 3.4149851584581814,
      "tokens_seen": 2339504128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029400200601805413,
      "loss": 1.3801,
      "theoretical_loss": 3.414977490558992,
      "tokens_seen": 2339569664
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029398194583751255,
      "loss": 1.2911,
      "theoretical_loss": 3.414969822934733,
      "tokens_seen": 2339635200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002939618856569709,
      "loss": 1.3391,
      "theoretical_loss": 3.414962155585386,
      "tokens_seen": 2339700736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002939418254764293,
      "loss": 1.275,
      "theoretical_loss": 3.4149544885109346,
      "tokens_seen": 2339766272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029392176529588764,
      "loss": 1.2562,
      "theoretical_loss": 3.4149468217113608,
      "tokens_seen": 2339831808
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029390170511534605,
      "loss": 1.3187,
      "theoretical_loss": 3.4149391551866466,
      "tokens_seen": 2339897344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938816449348044,
      "loss": 1.3331,
      "theoretical_loss": 3.4149314889367752,
      "tokens_seen": 2339962880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938615847542628,
      "loss": 1.3265,
      "theoretical_loss": 3.414923822961729,
      "tokens_seen": 2340028416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029384152457372114,
      "loss": 1.2927,
      "theoretical_loss": 3.4149161572614894,
      "tokens_seen": 2340093952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938214643931795,
      "loss": 1.3344,
      "theoretical_loss": 3.4149084918360404,
      "tokens_seen": 2340159488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938014042126379,
      "loss": 1.3062,
      "theoretical_loss": 3.414900826685363,
      "tokens_seen": 2340225024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937813440320963,
      "loss": 1.3522,
      "theoretical_loss": 3.414893161809441,
      "tokens_seen": 2340290560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029376128385155464,
      "loss": 1.3538,
      "theoretical_loss": 3.414885497208256,
      "tokens_seen": 2340356096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000293741223671013,
      "loss": 1.3811,
      "theoretical_loss": 3.4148778328817904,
      "tokens_seen": 2340421632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937211634904714,
      "loss": 1.3052,
      "theoretical_loss": 3.414870168830027,
      "tokens_seen": 2340487168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937011033099298,
      "loss": 1.2729,
      "theoretical_loss": 3.414862505052948,
      "tokens_seen": 2340552704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029368104312938815,
      "loss": 1.4063,
      "theoretical_loss": 3.414854841550537,
      "tokens_seen": 2340618240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002936609829488465,
      "loss": 1.2987,
      "theoretical_loss": 3.4148471783227743,
      "tokens_seen": 2340683776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029364092276830487,
      "loss": 1.4054,
      "theoretical_loss": 3.4148395153696445,
      "tokens_seen": 2340749312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002936208625877633,
      "loss": 1.2776,
      "theoretical_loss": 3.4148318526911288,
      "tokens_seen": 2340814848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029360080240722165,
      "loss": 1.3067,
      "theoretical_loss": 3.41482419028721,
      "tokens_seen": 2340880384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029358074222668,
      "loss": 1.346,
      "theoretical_loss": 3.4148165281578704,
      "tokens_seen": 2340945920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935606820461384,
      "loss": 1.3869,
      "theoretical_loss": 3.414808866303093,
      "tokens_seen": 2341011456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029354062186559684,
      "loss": 1.2453,
      "theoretical_loss": 3.4148012047228598,
      "tokens_seen": 2341076992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935205616850552,
      "loss": 1.2647,
      "theoretical_loss": 3.414793543417153,
      "tokens_seen": 2341142528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029350050150451357,
      "loss": 1.3429,
      "theoretical_loss": 3.414785882385956,
      "tokens_seen": 2341208064
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029348044132397193,
      "loss": 1.3021,
      "theoretical_loss": 3.414778221629251,
      "tokens_seen": 2341273600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029346038114343035,
      "loss": 1.357,
      "theoretical_loss": 3.41477056114702,
      "tokens_seen": 2341339136
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 852174,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2691469192504883,
      "objective/train/theoretical_loss": 3.4147629009392455,
      "objective/train/tokens_used": 711863776,
      "theoretical_loss": 3.4147629009392455,
      "tokens_seen": 2341404672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934403209628887,
      "loss": 1.2818,
      "theoretical_loss": 3.4147629009392455,
      "tokens_seen": 2341404672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029342026078234707,
      "loss": 1.3742,
      "theoretical_loss": 3.4147552410059103,
      "tokens_seen": 2341470208
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029340020060180543,
      "loss": 1.2507,
      "theoretical_loss": 3.414747581346997,
      "tokens_seen": 2341535744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933801404212638,
      "loss": 1.3375,
      "theoretical_loss": 3.4147399219624877,
      "tokens_seen": 2341601280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933600802407222,
      "loss": 1.3167,
      "theoretical_loss": 3.4147322628523655,
      "tokens_seen": 2341666816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933400200601806,
      "loss": 1.2899,
      "theoretical_loss": 3.414724604016612,
      "tokens_seen": 2341732352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029331995987963894,
      "loss": 1.3729,
      "theoretical_loss": 3.4147169454552104,
      "tokens_seen": 2341797888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932998996990973,
      "loss": 1.3274,
      "theoretical_loss": 3.414709287168143,
      "tokens_seen": 2341863424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932798395185557,
      "loss": 1.2407,
      "theoretical_loss": 3.4147016291553918,
      "tokens_seen": 2341928960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932597793380141,
      "loss": 1.323,
      "theoretical_loss": 3.4146939714169404,
      "tokens_seen": 2341994496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029323971915747244,
      "loss": 1.3038,
      "theoretical_loss": 3.4146863139527706,
      "tokens_seen": 2342060032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932196589769308,
      "loss": 1.3204,
      "theoretical_loss": 3.414678656762865,
      "tokens_seen": 2342125568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029319959879638916,
      "loss": 1.3188,
      "theoretical_loss": 3.4146709998472056,
      "tokens_seen": 2342191104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931795386158476,
      "loss": 1.4668,
      "theoretical_loss": 3.414663343205776,
      "tokens_seen": 2342256640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029315947843530594,
      "loss": 1.3487,
      "theoretical_loss": 3.4146556868385574,
      "tokens_seen": 2342322176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931394182547643,
      "loss": 1.3487,
      "theoretical_loss": 3.4146480307455334,
      "tokens_seen": 2342387712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029311935807422267,
      "loss": 1.3095,
      "theoretical_loss": 3.414640374926686,
      "tokens_seen": 2342453248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930992978936811,
      "loss": 1.3843,
      "theoretical_loss": 3.4146327193819976,
      "tokens_seen": 2342518784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029307923771313945,
      "loss": 1.2895,
      "theoretical_loss": 3.4146250641114513,
      "tokens_seen": 2342584320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930591775325978,
      "loss": 1.2366,
      "theoretical_loss": 3.414617409115029,
      "tokens_seen": 2342649856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029303911735205617,
      "loss": 1.4446,
      "theoretical_loss": 3.414609754392714,
      "tokens_seen": 2342715392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029301905717151453,
      "loss": 1.4213,
      "theoretical_loss": 3.4146020999444877,
      "tokens_seen": 2342780928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029299899699097295,
      "loss": 1.3561,
      "theoretical_loss": 3.414594445770333,
      "tokens_seen": 2342846464
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002929789368104313,
      "loss": 1.3572,
      "theoretical_loss": 3.4145867918702333,
      "tokens_seen": 2342912000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002929588766298897,
      "loss": 1.2555,
      "theoretical_loss": 3.41457913824417,
      "tokens_seen": 2342977536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029293881644934804,
      "loss": 1.1866,
      "theoretical_loss": 3.414571484892126,
      "tokens_seen": 2343043072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029291875626880645,
      "loss": 1.4035,
      "theoretical_loss": 3.414563831814084,
      "tokens_seen": 2343108608
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928986960882648,
      "loss": 1.395,
      "theoretical_loss": 3.4145561790100265,
      "tokens_seen": 2343174144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928786359077232,
      "loss": 1.3584,
      "theoretical_loss": 3.414548526479936,
      "tokens_seen": 2343239680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029285857572718154,
      "loss": 1.2548,
      "theoretical_loss": 3.414540874223795,
      "tokens_seen": 2343305216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928385155466399,
      "loss": 1.3288,
      "theoretical_loss": 3.4145332222415856,
      "tokens_seen": 2343370752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928184553660983,
      "loss": 1.2832,
      "theoretical_loss": 3.414525570533291,
      "tokens_seen": 2343436288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927983951855567,
      "loss": 1.3175,
      "theoretical_loss": 3.4145179190988935,
      "tokens_seen": 2343501824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029277833500501504,
      "loss": 1.4078,
      "theoretical_loss": 3.4145102679383754,
      "tokens_seen": 2343567360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927582748244734,
      "loss": 1.4082,
      "theoretical_loss": 3.4145026170517196,
      "tokens_seen": 2343632896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927382146439318,
      "loss": 1.3392,
      "theoretical_loss": 3.4144949664389084,
      "tokens_seen": 2343698432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927181544633902,
      "loss": 1.3127,
      "theoretical_loss": 3.4144873160999243,
      "tokens_seen": 2343763968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029269809428284855,
      "loss": 1.3112,
      "theoretical_loss": 3.4144796660347505,
      "tokens_seen": 2343829504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926780341023069,
      "loss": 1.3139,
      "theoretical_loss": 3.4144720162433684,
      "tokens_seen": 2343895040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029265797392176527,
      "loss": 1.3309,
      "theoretical_loss": 3.4144643667257615,
      "tokens_seen": 2343960576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926379137412237,
      "loss": 1.337,
      "theoretical_loss": 3.414456717481912,
      "tokens_seen": 2344026112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029261785356068205,
      "loss": 1.3388,
      "theoretical_loss": 3.4144490685118023,
      "tokens_seen": 2344091648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925977933801404,
      "loss": 1.3224,
      "theoretical_loss": 3.4144414198154154,
      "tokens_seen": 2344157184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925777331995988,
      "loss": 1.2667,
      "theoretical_loss": 3.414433771392733,
      "tokens_seen": 2344222720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925576730190572,
      "loss": 1.3927,
      "theoretical_loss": 3.4144261232437385,
      "tokens_seen": 2344288256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029253761283851555,
      "loss": 1.3305,
      "theoretical_loss": 3.4144184753684144,
      "tokens_seen": 2344353792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925175526579739,
      "loss": 1.2647,
      "theoretical_loss": 3.4144108277667424,
      "tokens_seen": 2344419328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924974924774323,
      "loss": 1.3053,
      "theoretical_loss": 3.4144031804387067,
      "tokens_seen": 2344484864
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924774322968907,
      "loss": 1.2303,
      "theoretical_loss": 3.4143955333842877,
      "tokens_seen": 2344550400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029245737211634906,
      "loss": 1.3762,
      "theoretical_loss": 3.41438788660347,
      "tokens_seen": 2344615936
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 853299,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2306807041168213,
      "objective/train/theoretical_loss": 3.414380240096235,
      "objective/train/tokens_used": 715140576,
      "theoretical_loss": 3.414380240096235,
      "tokens_seen": 2344681472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924373119358074,
      "loss": 1.2449,
      "theoretical_loss": 3.414380240096235,
      "tokens_seen": 2344681472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924172517552658,
      "loss": 1.2801,
      "theoretical_loss": 3.4143725938625655,
      "tokens_seen": 2344747008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029239719157472414,
      "loss": 1.3608,
      "theoretical_loss": 3.414364947902444,
      "tokens_seen": 2344812544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029237713139418256,
      "loss": 1.2909,
      "theoretical_loss": 3.414357302215853,
      "tokens_seen": 2344878080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923570712136409,
      "loss": 1.2538,
      "theoretical_loss": 3.4143496568027754,
      "tokens_seen": 2344943616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923370110330993,
      "loss": 1.4334,
      "theoretical_loss": 3.414342011663194,
      "tokens_seen": 2345009152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029231695085255765,
      "loss": 1.4027,
      "theoretical_loss": 3.4143343667970907,
      "tokens_seen": 2345074688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029229689067201606,
      "loss": 1.3883,
      "theoretical_loss": 3.414326722204448,
      "tokens_seen": 2345140224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922768304914744,
      "loss": 1.2721,
      "theoretical_loss": 3.4143190778852492,
      "tokens_seen": 2345205760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922567703109328,
      "loss": 1.2336,
      "theoretical_loss": 3.414311433839477,
      "tokens_seen": 2345271296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029223671013039115,
      "loss": 1.2408,
      "theoretical_loss": 3.4143037900671125,
      "tokens_seen": 2345336832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922166499498495,
      "loss": 1.3427,
      "theoretical_loss": 3.41429614656814,
      "tokens_seen": 2345402368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921965897693079,
      "loss": 1.4586,
      "theoretical_loss": 3.414288503342541,
      "tokens_seen": 2345467904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921765295887663,
      "loss": 1.3059,
      "theoretical_loss": 3.4142808603902988,
      "tokens_seen": 2345533440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029215646940822465,
      "loss": 1.3491,
      "theoretical_loss": 3.4142732177113952,
      "tokens_seen": 2345598976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292136409227683,
      "loss": 1.2738,
      "theoretical_loss": 3.4142655753058135,
      "tokens_seen": 2345664512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029211634904714143,
      "loss": 1.1786,
      "theoretical_loss": 3.414257933173536,
      "tokens_seen": 2345730048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002920962888665998,
      "loss": 1.347,
      "theoretical_loss": 3.4142502913145454,
      "tokens_seen": 2345795584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029207622868605815,
      "loss": 1.2846,
      "theoretical_loss": 3.414242649728824,
      "tokens_seen": 2345861120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002920561685055165,
      "loss": 1.301,
      "theoretical_loss": 3.4142350084163544,
      "tokens_seen": 2345926656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029203610832497493,
      "loss": 1.3233,
      "theoretical_loss": 3.4142273673771197,
      "tokens_seen": 2345992192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029201604814443335,
      "loss": 1.2772,
      "theoretical_loss": 3.414219726611102,
      "tokens_seen": 2346057728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002919959879638917,
      "loss": 1.3851,
      "theoretical_loss": 3.414212086118284,
      "tokens_seen": 2346123264
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002919759277833501,
      "loss": 1.3292,
      "theoretical_loss": 3.414204445898649,
      "tokens_seen": 2346188800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029195586760280844,
      "loss": 1.2977,
      "theoretical_loss": 3.4141968059521783,
      "tokens_seen": 2346254336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029193580742226685,
      "loss": 1.279,
      "theoretical_loss": 3.4141891662788555,
      "tokens_seen": 2346319872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002919157472417252,
      "loss": 1.3303,
      "theoretical_loss": 3.4141815268786626,
      "tokens_seen": 2346385408
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918956870611836,
      "loss": 1.3512,
      "theoretical_loss": 3.414173887751583,
      "tokens_seen": 2346450944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029187562688064194,
      "loss": 1.3343,
      "theoretical_loss": 3.4141662488975983,
      "tokens_seen": 2346516480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918555667001003,
      "loss": 1.2663,
      "theoretical_loss": 3.414158610316692,
      "tokens_seen": 2346582016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918355065195587,
      "loss": 1.3676,
      "theoretical_loss": 3.414150972008846,
      "tokens_seen": 2346647552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918154463390171,
      "loss": 1.4228,
      "theoretical_loss": 3.4141433339740432,
      "tokens_seen": 2346713088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029179538615847544,
      "loss": 1.2755,
      "theoretical_loss": 3.4141356962122664,
      "tokens_seen": 2346778624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917753259779338,
      "loss": 1.3381,
      "theoretical_loss": 3.4141280587234983,
      "tokens_seen": 2346844160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917552657973922,
      "loss": 1.2599,
      "theoretical_loss": 3.4141204215077208,
      "tokens_seen": 2346909696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917352056168506,
      "loss": 1.3679,
      "theoretical_loss": 3.4141127845649173,
      "tokens_seen": 2346975232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029171514543630895,
      "loss": 1.2848,
      "theoretical_loss": 3.4141051478950706,
      "tokens_seen": 2347040768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916950852557673,
      "loss": 1.356,
      "theoretical_loss": 3.414097511498162,
      "tokens_seen": 2347106304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029167502507522567,
      "loss": 1.3514,
      "theoretical_loss": 3.4140898753741755,
      "tokens_seen": 2347171840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916549648946841,
      "loss": 1.3511,
      "theoretical_loss": 3.414082239523093,
      "tokens_seen": 2347237376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029163490471414245,
      "loss": 1.3514,
      "theoretical_loss": 3.4140746039448975,
      "tokens_seen": 2347302912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916148445336008,
      "loss": 1.3456,
      "theoretical_loss": 3.4140669686395713,
      "tokens_seen": 2347368448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915947843530592,
      "loss": 1.2719,
      "theoretical_loss": 3.4140593336070975,
      "tokens_seen": 2347433984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915747241725176,
      "loss": 1.3268,
      "theoretical_loss": 3.414051698847458,
      "tokens_seen": 2347499520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029155466399197595,
      "loss": 1.2536,
      "theoretical_loss": 3.414044064360636,
      "tokens_seen": 2347565056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915346038114343,
      "loss": 1.4202,
      "theoretical_loss": 3.4140364301466137,
      "tokens_seen": 2347630592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915145436308927,
      "loss": 1.3299,
      "theoretical_loss": 3.4140287962053746,
      "tokens_seen": 2347696128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914944834503511,
      "loss": 1.3332,
      "theoretical_loss": 3.4140211625369004,
      "tokens_seen": 2347761664
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029147442326980946,
      "loss": 1.3181,
      "theoretical_loss": 3.414013529141174,
      "tokens_seen": 2347827200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914543630892678,
      "loss": 1.271,
      "theoretical_loss": 3.4140058960181787,
      "tokens_seen": 2347892736
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 854534,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3929357528686523,
      "objective/train/theoretical_loss": 3.4139982631678962,
      "objective/train/tokens_used": 718417376,
      "theoretical_loss": 3.4139982631678962,
      "tokens_seen": 2347958272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914343029087262,
      "loss": 1.3423,
      "theoretical_loss": 3.4139982631678962,
      "tokens_seen": 2347958272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029141424272818454,
      "loss": 1.3504,
      "theoretical_loss": 3.41399063059031,
      "tokens_seen": 2348023808
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029139418254764296,
      "loss": 1.3376,
      "theoretical_loss": 3.4139829982854017,
      "tokens_seen": 2348089344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913741223671013,
      "loss": 1.2886,
      "theoretical_loss": 3.4139753662531547,
      "tokens_seen": 2348154880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913540621865597,
      "loss": 1.1705,
      "theoretical_loss": 3.4139677344935517,
      "tokens_seen": 2348220416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029133400200601805,
      "loss": 1.2306,
      "theoretical_loss": 3.4139601030065747,
      "tokens_seen": 2348285952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029131394182547646,
      "loss": 1.2209,
      "theoretical_loss": 3.413952471792207,
      "tokens_seen": 2348351488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912938816449348,
      "loss": 1.3183,
      "theoretical_loss": 3.4139448408504314,
      "tokens_seen": 2348417024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912738214643932,
      "loss": 1.3027,
      "theoretical_loss": 3.4139372101812295,
      "tokens_seen": 2348482560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029125376128385155,
      "loss": 1.4091,
      "theoretical_loss": 3.4139295797845852,
      "tokens_seen": 2348548096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912337011033099,
      "loss": 1.2429,
      "theoretical_loss": 3.4139219496604802,
      "tokens_seen": 2348613632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029121364092276833,
      "loss": 1.4514,
      "theoretical_loss": 3.413914319808898,
      "tokens_seen": 2348679168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911935807422267,
      "loss": 1.3416,
      "theoretical_loss": 3.4139066902298207,
      "tokens_seen": 2348744704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029117352056168505,
      "loss": 1.1761,
      "theoretical_loss": 3.413899060923231,
      "tokens_seen": 2348810240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911534603811434,
      "loss": 1.2629,
      "theoretical_loss": 3.4138914318891116,
      "tokens_seen": 2348875776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029113340020060183,
      "loss": 1.4038,
      "theoretical_loss": 3.4138838031274457,
      "tokens_seen": 2348941312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911133400200602,
      "loss": 1.2231,
      "theoretical_loss": 3.413876174638215,
      "tokens_seen": 2349006848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029109327983951855,
      "loss": 1.41,
      "theoretical_loss": 3.413868546421403,
      "tokens_seen": 2349072384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910732196589769,
      "loss": 1.2798,
      "theoretical_loss": 3.413860918476992,
      "tokens_seen": 2349137920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910531594784353,
      "loss": 1.3151,
      "theoretical_loss": 3.4138532908049646,
      "tokens_seen": 2349203456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910330992978937,
      "loss": 1.389,
      "theoretical_loss": 3.4138456634053034,
      "tokens_seen": 2349268992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029101303911735206,
      "loss": 1.378,
      "theoretical_loss": 3.413838036277992,
      "tokens_seen": 2349334528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909929789368104,
      "loss": 1.2532,
      "theoretical_loss": 3.4138304094230114,
      "tokens_seen": 2349400064
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909729187562688,
      "loss": 1.315,
      "theoretical_loss": 3.413822782840346,
      "tokens_seen": 2349465600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909528585757272,
      "loss": 1.375,
      "theoretical_loss": 3.4138151565299775,
      "tokens_seen": 2349531136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029093279839518556,
      "loss": 1.3639,
      "theoretical_loss": 3.4138075304918885,
      "tokens_seen": 2349596672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909127382146439,
      "loss": 1.2942,
      "theoretical_loss": 3.413799904726062,
      "tokens_seen": 2349662208
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908926780341023,
      "loss": 1.2694,
      "theoretical_loss": 3.413792279232481,
      "tokens_seen": 2349727744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029087261785356065,
      "loss": 1.2612,
      "theoretical_loss": 3.4137846540111276,
      "tokens_seen": 2349793280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029085255767301906,
      "loss": 1.3823,
      "theoretical_loss": 3.413777029061985,
      "tokens_seen": 2349858816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908324974924774,
      "loss": 1.3357,
      "theoretical_loss": 3.4137694043850355,
      "tokens_seen": 2349924352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908124373119358,
      "loss": 1.3077,
      "theoretical_loss": 3.4137617799802618,
      "tokens_seen": 2349989888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029079237713139415,
      "loss": 1.3298,
      "theoretical_loss": 3.413754155847647,
      "tokens_seen": 2350055424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029077231695085257,
      "loss": 1.3167,
      "theoretical_loss": 3.413746531987173,
      "tokens_seen": 2350120960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029075225677031093,
      "loss": 1.3132,
      "theoretical_loss": 3.4137389083988237,
      "tokens_seen": 2350186496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907321965897693,
      "loss": 1.4474,
      "theoretical_loss": 3.413731285082581,
      "tokens_seen": 2350252032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029071213640922765,
      "loss": 1.2245,
      "theoretical_loss": 3.4137236620384277,
      "tokens_seen": 2350317568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000290692076228686,
      "loss": 1.324,
      "theoretical_loss": 3.413716039266346,
      "tokens_seen": 2350383104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029067201604814443,
      "loss": 1.3917,
      "theoretical_loss": 3.4137084167663194,
      "tokens_seen": 2350448640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002906519558676028,
      "loss": 1.3806,
      "theoretical_loss": 3.4137007945383306,
      "tokens_seen": 2350514176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029063189568706116,
      "loss": 1.311,
      "theoretical_loss": 3.4136931725823616,
      "tokens_seen": 2350579712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002906118355065195,
      "loss": 1.3925,
      "theoretical_loss": 3.413685550898396,
      "tokens_seen": 2350645248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029059177532597794,
      "loss": 1.4006,
      "theoretical_loss": 3.4136779294864157,
      "tokens_seen": 2350710784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905717151454363,
      "loss": 1.3693,
      "theoretical_loss": 3.413670308346404,
      "tokens_seen": 2350776320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029055165496489466,
      "loss": 1.3237,
      "theoretical_loss": 3.413662687478343,
      "tokens_seen": 2350841856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905315947843531,
      "loss": 1.4029,
      "theoretical_loss": 3.413655066882216,
      "tokens_seen": 2350907392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905115346038115,
      "loss": 1.3074,
      "theoretical_loss": 3.413647446558006,
      "tokens_seen": 2350972928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029049147442326986,
      "loss": 1.2797,
      "theoretical_loss": 3.4136398265056944,
      "tokens_seen": 2351038464
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002904714142427282,
      "loss": 1.3426,
      "theoretical_loss": 3.413632206725265,
      "tokens_seen": 2351104000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002904513540621866,
      "loss": 1.3192,
      "theoretical_loss": 3.4136245872167006,
      "tokens_seen": 2351169536
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 855824,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.311852216720581,
      "objective/train/theoretical_loss": 3.4136169679799835,
      "objective/train/tokens_used": 721694176,
      "theoretical_loss": 3.4136169679799835,
      "tokens_seen": 2351235072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029043129388164494,
      "loss": 1.3349,
      "theoretical_loss": 3.4136169679799835,
      "tokens_seen": 2351235072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029041123370110336,
      "loss": 1.3626,
      "theoretical_loss": 3.413609349015096,
      "tokens_seen": 2351300608
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903911735205617,
      "loss": 1.3525,
      "theoretical_loss": 3.4136017303220223,
      "tokens_seen": 2351366144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903711133400201,
      "loss": 1.4367,
      "theoretical_loss": 3.4135941119007436,
      "tokens_seen": 2351431680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029035105315947845,
      "loss": 1.2782,
      "theoretical_loss": 3.413586493751243,
      "tokens_seen": 2351497216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029033099297893686,
      "loss": 1.3611,
      "theoretical_loss": 3.4135788758735037,
      "tokens_seen": 2351562752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903109327983952,
      "loss": 1.1881,
      "theoretical_loss": 3.413571258267508,
      "tokens_seen": 2351628288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902908726178536,
      "loss": 1.3276,
      "theoretical_loss": 3.413563640933239,
      "tokens_seen": 2351693824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029027081243731195,
      "loss": 1.3406,
      "theoretical_loss": 3.413556023870679,
      "tokens_seen": 2351759360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902507522567703,
      "loss": 1.2682,
      "theoretical_loss": 3.413548407079811,
      "tokens_seen": 2351824896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029023069207622873,
      "loss": 1.2737,
      "theoretical_loss": 3.413540790560618,
      "tokens_seen": 2351890432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902106318956871,
      "loss": 1.3827,
      "theoretical_loss": 3.4135331743130823,
      "tokens_seen": 2351955968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029019057171514545,
      "loss": 1.3987,
      "theoretical_loss": 3.4135255583371866,
      "tokens_seen": 2352021504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901705115346038,
      "loss": 1.2845,
      "theoretical_loss": 3.413517942632914,
      "tokens_seen": 2352087040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029015045135406223,
      "loss": 1.3382,
      "theoretical_loss": 3.413510327200247,
      "tokens_seen": 2352152576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901303911735206,
      "loss": 1.357,
      "theoretical_loss": 3.4135027120391683,
      "tokens_seen": 2352218112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029011033099297896,
      "loss": 1.3524,
      "theoretical_loss": 3.413495097149661,
      "tokens_seen": 2352283648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900902708124373,
      "loss": 1.3853,
      "theoretical_loss": 3.413487482531708,
      "tokens_seen": 2352349184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900702106318957,
      "loss": 1.4829,
      "theoretical_loss": 3.413479868185291,
      "tokens_seen": 2352414720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900501504513541,
      "loss": 1.3206,
      "theoretical_loss": 3.4134722541103937,
      "tokens_seen": 2352480256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029003009027081246,
      "loss": 1.3706,
      "theoretical_loss": 3.4134646403069984,
      "tokens_seen": 2352545792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900100300902708,
      "loss": 1.3216,
      "theoretical_loss": 3.413457026775088,
      "tokens_seen": 2352611328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899899699097292,
      "loss": 1.3078,
      "theoretical_loss": 3.4134494135146456,
      "tokens_seen": 2352676864
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899699097291876,
      "loss": 1.2753,
      "theoretical_loss": 3.4134418005256535,
      "tokens_seen": 2352742400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028994984954864596,
      "loss": 1.3278,
      "theoretical_loss": 3.4134341878080945,
      "tokens_seen": 2352807936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899297893681043,
      "loss": 1.3278,
      "theoretical_loss": 3.4134265753619513,
      "tokens_seen": 2352873472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899097291875627,
      "loss": 1.2864,
      "theoretical_loss": 3.4134189631872074,
      "tokens_seen": 2352939008
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028988966900702105,
      "loss": 1.3791,
      "theoretical_loss": 3.4134113512838447,
      "tokens_seen": 2353004544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028986960882647946,
      "loss": 1.2683,
      "theoretical_loss": 3.413403739651846,
      "tokens_seen": 2353070080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898495486459378,
      "loss": 1.2813,
      "theoretical_loss": 3.4133961282911947,
      "tokens_seen": 2353135616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898294884653962,
      "loss": 1.2796,
      "theoretical_loss": 3.4133885172018728,
      "tokens_seen": 2353201152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028980942828485455,
      "loss": 1.3621,
      "theoretical_loss": 3.413380906383864,
      "tokens_seen": 2353266688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028978936810431297,
      "loss": 1.3153,
      "theoretical_loss": 3.4133732958371503,
      "tokens_seen": 2353332224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028976930792377133,
      "loss": 1.294,
      "theoretical_loss": 3.4133656855617147,
      "tokens_seen": 2353397760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897492477432297,
      "loss": 1.3717,
      "theoretical_loss": 3.4133580755575403,
      "tokens_seen": 2353463296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028972918756268805,
      "loss": 1.3535,
      "theoretical_loss": 3.413350465824609,
      "tokens_seen": 2353528832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897091273821464,
      "loss": 1.331,
      "theoretical_loss": 3.4133428563629047,
      "tokens_seen": 2353594368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028968906720160483,
      "loss": 1.3378,
      "theoretical_loss": 3.4133352471724097,
      "tokens_seen": 2353659904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896690070210632,
      "loss": 1.3394,
      "theoretical_loss": 3.413327638253106,
      "tokens_seen": 2353725440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028964894684052156,
      "loss": 1.3659,
      "theoretical_loss": 3.4133200296049777,
      "tokens_seen": 2353790976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896288866599799,
      "loss": 1.3217,
      "theoretical_loss": 3.4133124212280066,
      "tokens_seen": 2353856512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028960882647943834,
      "loss": 1.327,
      "theoretical_loss": 3.4133048131221764,
      "tokens_seen": 2353922048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895887662988967,
      "loss": 1.2518,
      "theoretical_loss": 3.413297205287469,
      "tokens_seen": 2353987584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028956870611835506,
      "loss": 1.2116,
      "theoretical_loss": 3.4132895977238675,
      "tokens_seen": 2354053120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895486459378134,
      "loss": 1.3201,
      "theoretical_loss": 3.413281990431355,
      "tokens_seen": 2354118656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895285857572718,
      "loss": 1.3274,
      "theoretical_loss": 3.413274383409914,
      "tokens_seen": 2354184192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895085255767302,
      "loss": 1.2534,
      "theoretical_loss": 3.413266776659527,
      "tokens_seen": 2354249728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028948846539618856,
      "loss": 1.3595,
      "theoretical_loss": 3.4132591701801775,
      "tokens_seen": 2354315264
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894684052156469,
      "loss": 1.3126,
      "theoretical_loss": 3.413251563971848,
      "tokens_seen": 2354380800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894483450351053,
      "loss": 1.3224,
      "theoretical_loss": 3.413243958034521,
      "tokens_seen": 2354446336
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 856929,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.329018235206604,
      "objective/train/theoretical_loss": 3.4132363523681795,
      "objective/train/tokens_used": 724970976,
      "theoretical_loss": 3.4132363523681795,
      "tokens_seen": 2354511872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894282848545637,
      "loss": 1.323,
      "theoretical_loss": 3.4132363523681795,
      "tokens_seen": 2354511872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028940822467402207,
      "loss": 1.366,
      "theoretical_loss": 3.413228746972807,
      "tokens_seen": 2354577408
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028938816449348043,
      "loss": 1.3172,
      "theoretical_loss": 3.4132211418483847,
      "tokens_seen": 2354642944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893681043129388,
      "loss": 1.227,
      "theoretical_loss": 3.413213536994897,
      "tokens_seen": 2354708480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893480441323972,
      "loss": 1.3933,
      "theoretical_loss": 3.4132059324123256,
      "tokens_seen": 2354774016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028932798395185557,
      "loss": 1.2554,
      "theoretical_loss": 3.4131983281006537,
      "tokens_seen": 2354839552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028930792377131393,
      "loss": 1.2502,
      "theoretical_loss": 3.4131907240598647,
      "tokens_seen": 2354905088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892878635907723,
      "loss": 1.2954,
      "theoretical_loss": 3.4131831202899408,
      "tokens_seen": 2354970624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028926780341023066,
      "loss": 1.352,
      "theoretical_loss": 3.4131755167908646,
      "tokens_seen": 2355036160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892477432296891,
      "loss": 1.2774,
      "theoretical_loss": 3.4131679135626194,
      "tokens_seen": 2355101696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028922768304914744,
      "loss": 1.3855,
      "theoretical_loss": 3.4131603106051878,
      "tokens_seen": 2355167232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892076228686058,
      "loss": 1.3631,
      "theoretical_loss": 3.4131527079185524,
      "tokens_seen": 2355232768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028918756268806416,
      "loss": 1.2602,
      "theoretical_loss": 3.413145105502697,
      "tokens_seen": 2355298304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891675025075226,
      "loss": 1.2464,
      "theoretical_loss": 3.413137503357603,
      "tokens_seen": 2355363840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028914744232698094,
      "loss": 1.3946,
      "theoretical_loss": 3.4131299014832543,
      "tokens_seen": 2355429376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891273821464393,
      "loss": 1.3346,
      "theoretical_loss": 3.413122299879633,
      "tokens_seen": 2355494912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028910732196589766,
      "loss": 1.3224,
      "theoretical_loss": 3.4131146985467224,
      "tokens_seen": 2355560448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289087261785356,
      "loss": 1.3819,
      "theoretical_loss": 3.4131070974845055,
      "tokens_seen": 2355625984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028906720160481444,
      "loss": 1.244,
      "theoretical_loss": 3.4130994966929644,
      "tokens_seen": 2355691520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890471414242728,
      "loss": 1.2415,
      "theoretical_loss": 3.413091896172083,
      "tokens_seen": 2355757056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890270812437312,
      "loss": 1.3256,
      "theoretical_loss": 3.4130842959218426,
      "tokens_seen": 2355822592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890070210631896,
      "loss": 1.394,
      "theoretical_loss": 3.4130766959422276,
      "tokens_seen": 2355888128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288986960882648,
      "loss": 1.2872,
      "theoretical_loss": 3.4130690962332197,
      "tokens_seen": 2355953664
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028896690070210636,
      "loss": 1.3187,
      "theoretical_loss": 3.4130614967948025,
      "tokens_seen": 2356019200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889468405215647,
      "loss": 1.3057,
      "theoretical_loss": 3.4130538976269587,
      "tokens_seen": 2356084736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889267803410231,
      "loss": 1.2971,
      "theoretical_loss": 3.413046298729671,
      "tokens_seen": 2356150272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028890672016048145,
      "loss": 1.3528,
      "theoretical_loss": 3.4130387001029217,
      "tokens_seen": 2356215808
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028888665997993986,
      "loss": 1.2862,
      "theoretical_loss": 3.4130311017466943,
      "tokens_seen": 2356281344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028886659979939823,
      "loss": 1.3092,
      "theoretical_loss": 3.4130235036609715,
      "tokens_seen": 2356346880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002888465396188566,
      "loss": 1.2889,
      "theoretical_loss": 3.4130159058457368,
      "tokens_seen": 2356412416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028882647943831495,
      "loss": 1.357,
      "theoretical_loss": 3.413008308300972,
      "tokens_seen": 2356477952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028880641925777337,
      "loss": 1.2521,
      "theoretical_loss": 3.41300071102666,
      "tokens_seen": 2356543488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028878635907723173,
      "loss": 1.3158,
      "theoretical_loss": 3.4129931140227843,
      "tokens_seen": 2356609024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887662988966901,
      "loss": 1.3073,
      "theoretical_loss": 3.4129855172893278,
      "tokens_seen": 2356674560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028874623871614845,
      "loss": 1.2432,
      "theoretical_loss": 3.4129779208262723,
      "tokens_seen": 2356740096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887261785356068,
      "loss": 1.3838,
      "theoretical_loss": 3.412970324633602,
      "tokens_seen": 2356805632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028870611835506523,
      "loss": 1.3482,
      "theoretical_loss": 3.4129627287112987,
      "tokens_seen": 2356871168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886860581745236,
      "loss": 1.325,
      "theoretical_loss": 3.412955133059346,
      "tokens_seen": 2356936704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028866599799398196,
      "loss": 1.2589,
      "theoretical_loss": 3.4129475376777263,
      "tokens_seen": 2357002240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886459378134403,
      "loss": 1.3432,
      "theoretical_loss": 3.412939942566423,
      "tokens_seen": 2357067776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028862587763289874,
      "loss": 1.2777,
      "theoretical_loss": 3.412932347725418,
      "tokens_seen": 2357133312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886058174523571,
      "loss": 1.252,
      "theoretical_loss": 3.412924753154695,
      "tokens_seen": 2357198848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028858575727181546,
      "loss": 1.3248,
      "theoretical_loss": 3.4129171588542366,
      "tokens_seen": 2357264384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885656970912738,
      "loss": 1.3245,
      "theoretical_loss": 3.4129095648240257,
      "tokens_seen": 2357329920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885456369107322,
      "loss": 1.4255,
      "theoretical_loss": 3.412901971064045,
      "tokens_seen": 2357395456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885255767301906,
      "loss": 1.419,
      "theoretical_loss": 3.4128943775742773,
      "tokens_seen": 2357460992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028850551654964896,
      "loss": 1.3005,
      "theoretical_loss": 3.4128867843547064,
      "tokens_seen": 2357526528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884854563691073,
      "loss": 1.318,
      "theoretical_loss": 3.412879191405314,
      "tokens_seen": 2357592064
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884653961885657,
      "loss": 1.319,
      "theoretical_loss": 3.412871598726084,
      "tokens_seen": 2357657600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884453360080241,
      "loss": 1.2678,
      "theoretical_loss": 3.4128640063169975,
      "tokens_seen": 2357723136
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 858137,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2626562118530273,
      "objective/train/theoretical_loss": 3.41285641417804,
      "objective/train/tokens_used": 728247776,
      "theoretical_loss": 3.41285641417804,
      "tokens_seen": 2357788672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028842527582748247,
      "loss": 1.2446,
      "theoretical_loss": 3.41285641417804,
      "tokens_seen": 2357788672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028840521564694083,
      "loss": 1.2647,
      "theoretical_loss": 3.412848822309192,
      "tokens_seen": 2357854208
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883851554663992,
      "loss": 1.2956,
      "theoretical_loss": 3.4128412307104377,
      "tokens_seen": 2357919744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883650952858576,
      "loss": 1.386,
      "theoretical_loss": 3.41283363938176,
      "tokens_seen": 2357985280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028834503510531597,
      "loss": 1.2778,
      "theoretical_loss": 3.412826048323141,
      "tokens_seen": 2358050816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028832497492477433,
      "loss": 1.261,
      "theoretical_loss": 3.412818457534564,
      "tokens_seen": 2358116352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883049147442327,
      "loss": 1.3761,
      "theoretical_loss": 3.412810867016012,
      "tokens_seen": 2358181888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028828485456369106,
      "loss": 1.3664,
      "theoretical_loss": 3.4128032767674683,
      "tokens_seen": 2358247424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002882647943831495,
      "loss": 1.3789,
      "theoretical_loss": 3.4127956867889147,
      "tokens_seen": 2358312960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028824473420260784,
      "loss": 1.3448,
      "theoretical_loss": 3.412788097080335,
      "tokens_seen": 2358378496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002882246740220662,
      "loss": 1.2886,
      "theoretical_loss": 3.4127805076417115,
      "tokens_seen": 2358444032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028820461384152456,
      "loss": 1.3138,
      "theoretical_loss": 3.4127729184730278,
      "tokens_seen": 2358509568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288184553660983,
      "loss": 1.2613,
      "theoretical_loss": 3.412765329574266,
      "tokens_seen": 2358575104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028816449348044134,
      "loss": 1.342,
      "theoretical_loss": 3.4127577409454095,
      "tokens_seen": 2358640640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881444332998997,
      "loss": 1.205,
      "theoretical_loss": 3.412750152586441,
      "tokens_seen": 2358706176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028812437311935806,
      "loss": 1.3256,
      "theoretical_loss": 3.4127425644973437,
      "tokens_seen": 2358771712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881043129388164,
      "loss": 1.4523,
      "theoretical_loss": 3.4127349766781,
      "tokens_seen": 2358837248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028808425275827484,
      "loss": 1.3606,
      "theoretical_loss": 3.4127273891286936,
      "tokens_seen": 2358902784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002880641925777332,
      "loss": 1.2846,
      "theoretical_loss": 3.4127198018491067,
      "tokens_seen": 2358968320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028804413239719157,
      "loss": 1.3933,
      "theoretical_loss": 3.4127122148393223,
      "tokens_seen": 2359033856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028802407221664993,
      "loss": 1.3951,
      "theoretical_loss": 3.4127046280993234,
      "tokens_seen": 2359099392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028800401203610835,
      "loss": 1.2757,
      "theoretical_loss": 3.4126970416290927,
      "tokens_seen": 2359164928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879839518555667,
      "loss": 1.3559,
      "theoretical_loss": 3.412689455428614,
      "tokens_seen": 2359230464
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028796389167502507,
      "loss": 1.3227,
      "theoretical_loss": 3.412681869497869,
      "tokens_seen": 2359296000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028794383149448343,
      "loss": 1.2318,
      "theoretical_loss": 3.4126742838368416,
      "tokens_seen": 2359361536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879237713139418,
      "loss": 1.2728,
      "theoretical_loss": 3.412666698445514,
      "tokens_seen": 2359427072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879037111334002,
      "loss": 1.3249,
      "theoretical_loss": 3.4126591133238695,
      "tokens_seen": 2359492608
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878836509528586,
      "loss": 1.3723,
      "theoretical_loss": 3.412651528471891,
      "tokens_seen": 2359558144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028786359077231694,
      "loss": 1.2988,
      "theoretical_loss": 3.4126439438895613,
      "tokens_seen": 2359623680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878435305917753,
      "loss": 1.2978,
      "theoretical_loss": 3.4126363595768634,
      "tokens_seen": 2359689216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878234704112337,
      "loss": 1.2914,
      "theoretical_loss": 3.41262877553378,
      "tokens_seen": 2359754752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878034102306921,
      "loss": 1.4378,
      "theoretical_loss": 3.412621191760295,
      "tokens_seen": 2359820288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028778335005015044,
      "loss": 1.32,
      "theoretical_loss": 3.41261360825639,
      "tokens_seen": 2359885824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877632898696088,
      "loss": 1.2968,
      "theoretical_loss": 3.4126060250220487,
      "tokens_seen": 2359951360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028774322968906716,
      "loss": 1.3152,
      "theoretical_loss": 3.412598442057254,
      "tokens_seen": 2360016896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877231695085256,
      "loss": 1.2884,
      "theoretical_loss": 3.412590859361988,
      "tokens_seen": 2360082432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028770310932798394,
      "loss": 1.3062,
      "theoretical_loss": 3.4125832769362345,
      "tokens_seen": 2360147968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876830491474423,
      "loss": 1.3748,
      "theoretical_loss": 3.412575694779977,
      "tokens_seen": 2360213504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028766298896690067,
      "loss": 1.2681,
      "theoretical_loss": 3.4125681128931973,
      "tokens_seen": 2360279040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876429287863591,
      "loss": 1.2125,
      "theoretical_loss": 3.4125605312758784,
      "tokens_seen": 2360344576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028762286860581745,
      "loss": 1.2953,
      "theoretical_loss": 3.4125529499280036,
      "tokens_seen": 2360410112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876028084252758,
      "loss": 1.3048,
      "theoretical_loss": 3.412545368849556,
      "tokens_seen": 2360475648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028758274824473417,
      "loss": 1.3188,
      "theoretical_loss": 3.4125377880405185,
      "tokens_seen": 2360541184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028756268806419253,
      "loss": 1.3008,
      "theoretical_loss": 3.4125302075008737,
      "tokens_seen": 2360606720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028754262788365095,
      "loss": 1.3116,
      "theoretical_loss": 3.412522627230605,
      "tokens_seen": 2360672256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028752256770310936,
      "loss": 1.3642,
      "theoretical_loss": 3.412515047229695,
      "tokens_seen": 2360737792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002875025075225677,
      "loss": 1.2148,
      "theoretical_loss": 3.4125074674981266,
      "tokens_seen": 2360803328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874824473420261,
      "loss": 1.3244,
      "theoretical_loss": 3.412499888035883,
      "tokens_seen": 2360868864
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874623871614845,
      "loss": 1.3952,
      "theoretical_loss": 3.412492308842947,
      "tokens_seen": 2360934400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028744232698094287,
      "loss": 1.3934,
      "theoretical_loss": 3.412484729919302,
      "tokens_seen": 2360999936
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 859369,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2805547714233398,
      "objective/train/theoretical_loss": 3.41247715126493,
      "objective/train/tokens_used": 731524576,
      "theoretical_loss": 3.41247715126493,
      "tokens_seen": 2361065472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028742226680040123,
      "loss": 1.241,
      "theoretical_loss": 3.41247715126493,
      "tokens_seen": 2361065472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874022066198596,
      "loss": 1.3405,
      "theoretical_loss": 3.412469572879815,
      "tokens_seen": 2361131008
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287382146439318,
      "loss": 1.3676,
      "theoretical_loss": 3.412461994763939,
      "tokens_seen": 2361196544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028736208625877637,
      "loss": 1.3654,
      "theoretical_loss": 3.412454416917286,
      "tokens_seen": 2361262080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028734202607823473,
      "loss": 1.229,
      "theoretical_loss": 3.412446839339838,
      "tokens_seen": 2361327616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873219658976931,
      "loss": 1.2543,
      "theoretical_loss": 3.4124392620315787,
      "tokens_seen": 2361393152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028730190571715146,
      "loss": 1.2465,
      "theoretical_loss": 3.4124316849924905,
      "tokens_seen": 2361458688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872818455366099,
      "loss": 1.353,
      "theoretical_loss": 3.412424108222557,
      "tokens_seen": 2361524224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028726178535606824,
      "loss": 1.3166,
      "theoretical_loss": 3.4124165317217603,
      "tokens_seen": 2361589760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872417251755266,
      "loss": 1.2422,
      "theoretical_loss": 3.4124089554900845,
      "tokens_seen": 2361655296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028722166499498496,
      "loss": 1.2223,
      "theoretical_loss": 3.4124013795275117,
      "tokens_seen": 2361720832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872016048144434,
      "loss": 1.2965,
      "theoretical_loss": 3.412393803834025,
      "tokens_seen": 2361786368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028718154463390174,
      "loss": 1.2567,
      "theoretical_loss": 3.4123862284096074,
      "tokens_seen": 2361851904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002871614844533601,
      "loss": 1.3686,
      "theoretical_loss": 3.4123786532542417,
      "tokens_seen": 2361917440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028714142427281846,
      "loss": 1.3152,
      "theoretical_loss": 3.4123710783679115,
      "tokens_seen": 2361982976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002871213640922768,
      "loss": 1.3515,
      "theoretical_loss": 3.4123635037505995,
      "tokens_seen": 2362048512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028710130391173524,
      "loss": 1.3211,
      "theoretical_loss": 3.4123559294022883,
      "tokens_seen": 2362114048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002870812437311936,
      "loss": 1.3728,
      "theoretical_loss": 3.412348355322962,
      "tokens_seen": 2362179584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028706118355065197,
      "loss": 1.3758,
      "theoretical_loss": 3.4123407815126017,
      "tokens_seen": 2362245120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028704112337011033,
      "loss": 1.1617,
      "theoretical_loss": 3.412333207971192,
      "tokens_seen": 2362310656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028702106318956875,
      "loss": 1.3347,
      "theoretical_loss": 3.4123256346987154,
      "tokens_seen": 2362376192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002870010030090271,
      "loss": 1.4538,
      "theoretical_loss": 3.4123180616951547,
      "tokens_seen": 2362441728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028698094282848547,
      "loss": 1.2873,
      "theoretical_loss": 3.412310488960493,
      "tokens_seen": 2362507264
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028696088264794383,
      "loss": 1.3706,
      "theoretical_loss": 3.4123029164947134,
      "tokens_seen": 2362572800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002869408224674022,
      "loss": 1.2547,
      "theoretical_loss": 3.412295344297799,
      "tokens_seen": 2362638336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002869207622868606,
      "loss": 1.2416,
      "theoretical_loss": 3.4122877723697327,
      "tokens_seen": 2362703872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000286900702106319,
      "loss": 1.3231,
      "theoretical_loss": 3.412280200710497,
      "tokens_seen": 2362769408
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028688064192577734,
      "loss": 1.3412,
      "theoretical_loss": 3.4122726293200754,
      "tokens_seen": 2362834944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868605817452357,
      "loss": 1.4128,
      "theoretical_loss": 3.412265058198451,
      "tokens_seen": 2362900480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868405215646941,
      "loss": 1.3395,
      "theoretical_loss": 3.4122574873456064,
      "tokens_seen": 2362966016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868204613841525,
      "loss": 1.3915,
      "theoretical_loss": 3.412249916761525,
      "tokens_seen": 2363031552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028680040120361084,
      "loss": 1.3195,
      "theoretical_loss": 3.41224234644619,
      "tokens_seen": 2363097088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867803410230692,
      "loss": 1.2825,
      "theoretical_loss": 3.4122347763995835,
      "tokens_seen": 2363162624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028676028084252756,
      "loss": 1.3586,
      "theoretical_loss": 3.412227206621689,
      "tokens_seen": 2363228160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000286740220661986,
      "loss": 1.2506,
      "theoretical_loss": 3.4122196371124898,
      "tokens_seen": 2363293696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028672016048144434,
      "loss": 1.2937,
      "theoretical_loss": 3.4122120678719687,
      "tokens_seen": 2363359232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867001003009027,
      "loss": 1.3666,
      "theoretical_loss": 3.4122044989001084,
      "tokens_seen": 2363424768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028668004012036107,
      "loss": 1.4238,
      "theoretical_loss": 3.4121969301968926,
      "tokens_seen": 2363490304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866599799398195,
      "loss": 1.3419,
      "theoretical_loss": 3.4121893617623034,
      "tokens_seen": 2363555840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028663991975927785,
      "loss": 1.2763,
      "theoretical_loss": 3.4121817935963246,
      "tokens_seen": 2363621376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866198595787362,
      "loss": 1.2385,
      "theoretical_loss": 3.412174225698939,
      "tokens_seen": 2363686912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028659979939819457,
      "loss": 1.3064,
      "theoretical_loss": 3.41216665807013,
      "tokens_seen": 2363752448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028657973921765293,
      "loss": 1.4066,
      "theoretical_loss": 3.4121590907098795,
      "tokens_seen": 2363817984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028655967903711135,
      "loss": 1.2873,
      "theoretical_loss": 3.4121515236181716,
      "tokens_seen": 2363883520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002865396188565697,
      "loss": 1.3557,
      "theoretical_loss": 3.4121439567949885,
      "tokens_seen": 2363949056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002865195586760281,
      "loss": 1.3409,
      "theoretical_loss": 3.412136390240314,
      "tokens_seen": 2364014592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028649949849548644,
      "loss": 1.2567,
      "theoretical_loss": 3.4121288239541308,
      "tokens_seen": 2364080128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028647943831494485,
      "loss": 1.2833,
      "theoretical_loss": 3.412121257936422,
      "tokens_seen": 2364145664
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864593781344032,
      "loss": 1.2844,
      "theoretical_loss": 3.4121136921871704,
      "tokens_seen": 2364211200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864393179538616,
      "loss": 1.3008,
      "theoretical_loss": 3.4121061267063593,
      "tokens_seen": 2364276736
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 860660,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3390905857086182,
      "objective/train/theoretical_loss": 3.4120985614939716,
      "objective/train/tokens_used": 734801376,
      "theoretical_loss": 3.4120985614939716,
      "tokens_seen": 2364342272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028641925777331994,
      "loss": 1.2884,
      "theoretical_loss": 3.4120985614939716,
      "tokens_seen": 2364342272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863991975927783,
      "loss": 1.4302,
      "theoretical_loss": 3.41209099654999,
      "tokens_seen": 2364407808
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863791374122367,
      "loss": 1.3953,
      "theoretical_loss": 3.4120834318743984,
      "tokens_seen": 2364473344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863590772316951,
      "loss": 1.3037,
      "theoretical_loss": 3.4120758674671796,
      "tokens_seen": 2364538880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028633901705115344,
      "loss": 1.2538,
      "theoretical_loss": 3.412068303328316,
      "tokens_seen": 2364604416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863189568706118,
      "loss": 1.3618,
      "theoretical_loss": 3.412060739457791,
      "tokens_seen": 2364669952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862988966900702,
      "loss": 1.2846,
      "theoretical_loss": 3.4120531758555877,
      "tokens_seen": 2364735488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862788365095286,
      "loss": 1.3461,
      "theoretical_loss": 3.412045612521689,
      "tokens_seen": 2364801024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028625877632898694,
      "loss": 1.3408,
      "theoretical_loss": 3.4120380494560782,
      "tokens_seen": 2364866560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862387161484453,
      "loss": 1.3207,
      "theoretical_loss": 3.412030486658738,
      "tokens_seen": 2364932096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028621865596790367,
      "loss": 1.3847,
      "theoretical_loss": 3.412022924129652,
      "tokens_seen": 2364997632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861985957873621,
      "loss": 1.2924,
      "theoretical_loss": 3.4120153618688027,
      "tokens_seen": 2365063168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028617853560682045,
      "loss": 1.3711,
      "theoretical_loss": 3.4120077998761733,
      "tokens_seen": 2365128704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861584754262788,
      "loss": 1.2808,
      "theoretical_loss": 3.412000238151747,
      "tokens_seen": 2365194240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028613841524573717,
      "loss": 1.3641,
      "theoretical_loss": 3.411992676695507,
      "tokens_seen": 2365259776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861183550651956,
      "loss": 1.2874,
      "theoretical_loss": 3.411985115507436,
      "tokens_seen": 2365325312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028609829488465395,
      "loss": 1.3244,
      "theoretical_loss": 3.411977554587517,
      "tokens_seen": 2365390848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860782347041123,
      "loss": 1.3673,
      "theoretical_loss": 3.4119699939357337,
      "tokens_seen": 2365456384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860581745235707,
      "loss": 1.3912,
      "theoretical_loss": 3.4119624335520684,
      "tokens_seen": 2365521920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860381143430291,
      "loss": 1.3915,
      "theoretical_loss": 3.4119548734365046,
      "tokens_seen": 2365587456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860180541624875,
      "loss": 1.351,
      "theoretical_loss": 3.411947313589025,
      "tokens_seen": 2365652992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028599799398194587,
      "loss": 1.1967,
      "theoretical_loss": 3.4119397540096132,
      "tokens_seen": 2365718528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028597793380140423,
      "loss": 1.2212,
      "theoretical_loss": 3.411932194698252,
      "tokens_seen": 2365784064
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002859578736208626,
      "loss": 1.3651,
      "theoretical_loss": 3.411924635654924,
      "tokens_seen": 2365849600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285937813440321,
      "loss": 1.2923,
      "theoretical_loss": 3.411917076879613,
      "tokens_seen": 2365915136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002859177532597794,
      "loss": 1.2929,
      "theoretical_loss": 3.411909518372302,
      "tokens_seen": 2365980672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028589769307923774,
      "loss": 1.2801,
      "theoretical_loss": 3.411901960132974,
      "tokens_seen": 2366046208
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858776328986961,
      "loss": 1.3,
      "theoretical_loss": 3.4118944021616118,
      "tokens_seen": 2366111744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858575727181545,
      "loss": 1.3226,
      "theoretical_loss": 3.4118868444581985,
      "tokens_seen": 2366177280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858375125376129,
      "loss": 1.4061,
      "theoretical_loss": 3.4118792870227175,
      "tokens_seen": 2366242816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028581745235707124,
      "loss": 1.3107,
      "theoretical_loss": 3.411871729855151,
      "tokens_seen": 2366308352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857973921765296,
      "loss": 1.2826,
      "theoretical_loss": 3.4118641729554837,
      "tokens_seen": 2366373888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028577733199598796,
      "loss": 1.4263,
      "theoretical_loss": 3.411856616323697,
      "tokens_seen": 2366439424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857572718154464,
      "loss": 1.3163,
      "theoretical_loss": 3.4118490599597755,
      "tokens_seen": 2366504960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028573721163490474,
      "loss": 1.3527,
      "theoretical_loss": 3.411841503863701,
      "tokens_seen": 2366570496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857171514543631,
      "loss": 1.3231,
      "theoretical_loss": 3.4118339480354574,
      "tokens_seen": 2366636032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028569709127382147,
      "loss": 1.2775,
      "theoretical_loss": 3.411826392475027,
      "tokens_seen": 2366701568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856770310932799,
      "loss": 1.2985,
      "theoretical_loss": 3.4118188371823943,
      "tokens_seen": 2366767104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028565697091273825,
      "loss": 1.3201,
      "theoretical_loss": 3.411811282157541,
      "tokens_seen": 2366832640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856369107321966,
      "loss": 1.2818,
      "theoretical_loss": 3.4118037274004505,
      "tokens_seen": 2366898176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028561685055165497,
      "loss": 1.3636,
      "theoretical_loss": 3.411796172911106,
      "tokens_seen": 2366963712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028559679037111333,
      "loss": 1.2351,
      "theoretical_loss": 3.4117886186894912,
      "tokens_seen": 2367029248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028557673019057175,
      "loss": 1.3674,
      "theoretical_loss": 3.4117810647355884,
      "tokens_seen": 2367094784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855566700100301,
      "loss": 1.2995,
      "theoretical_loss": 3.411773511049381,
      "tokens_seen": 2367160320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855366098294885,
      "loss": 1.3726,
      "theoretical_loss": 3.4117659576308523,
      "tokens_seen": 2367225856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028551654964894684,
      "loss": 1.3567,
      "theoretical_loss": 3.411758404479985,
      "tokens_seen": 2367291392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028549648946840525,
      "loss": 1.3031,
      "theoretical_loss": 3.4117508515967625,
      "tokens_seen": 2367356928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854764292878636,
      "loss": 1.2887,
      "theoretical_loss": 3.411743298981168,
      "tokens_seen": 2367422464
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285456369107322,
      "loss": 1.1462,
      "theoretical_loss": 3.411735746633184,
      "tokens_seen": 2367488000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028543630892678034,
      "loss": 1.3338,
      "theoretical_loss": 3.411728194552794,
      "tokens_seen": 2367553536
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 861837,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3536452054977417,
      "objective/train/theoretical_loss": 3.411720642739981,
      "objective/train/tokens_used": 738078176,
      "theoretical_loss": 3.411720642739981,
      "tokens_seen": 2367619072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854162487462387,
      "loss": 1.3665,
      "theoretical_loss": 3.411720642739981,
      "tokens_seen": 2367619072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002853961885656971,
      "loss": 1.2447,
      "theoretical_loss": 3.411713091194729,
      "tokens_seen": 2367684608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853761283851555,
      "loss": 1.3943,
      "theoretical_loss": 3.41170553991702,
      "tokens_seen": 2367750144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028535606820461384,
      "loss": 1.2777,
      "theoretical_loss": 3.411697988906837,
      "tokens_seen": 2367815680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853360080240722,
      "loss": 1.2979,
      "theoretical_loss": 3.411690438164164,
      "tokens_seen": 2367881216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853159478435306,
      "loss": 1.4002,
      "theoretical_loss": 3.411682887688984,
      "tokens_seen": 2367946752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000285295887662989,
      "loss": 1.2942,
      "theoretical_loss": 3.411675337481279,
      "tokens_seen": 2368012288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028527582748244735,
      "loss": 1.1875,
      "theoretical_loss": 3.4116677875410337,
      "tokens_seen": 2368077824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852557673019057,
      "loss": 1.2492,
      "theoretical_loss": 3.4116602378682304,
      "tokens_seen": 2368143360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028523570712136407,
      "loss": 1.4297,
      "theoretical_loss": 3.411652688462852,
      "tokens_seen": 2368208896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852156469408225,
      "loss": 1.4129,
      "theoretical_loss": 3.4116451393248823,
      "tokens_seen": 2368274432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028519558676028085,
      "loss": 1.3353,
      "theoretical_loss": 3.411637590454304,
      "tokens_seen": 2368339968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851755265797392,
      "loss": 1.2396,
      "theoretical_loss": 3.4116300418511,
      "tokens_seen": 2368405504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028515546639919757,
      "loss": 1.4334,
      "theoretical_loss": 3.4116224935152544,
      "tokens_seen": 2368471040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000285135406218656,
      "loss": 1.2243,
      "theoretical_loss": 3.411614945446749,
      "tokens_seen": 2368536576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028511534603811435,
      "loss": 1.3635,
      "theoretical_loss": 3.411607397645568,
      "tokens_seen": 2368602112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850952858575727,
      "loss": 1.288,
      "theoretical_loss": 3.4115998501116938,
      "tokens_seen": 2368667648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850752256770311,
      "loss": 1.4284,
      "theoretical_loss": 3.41159230284511,
      "tokens_seen": 2368733184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850551654964895,
      "loss": 1.3042,
      "theoretical_loss": 3.4115847558457997,
      "tokens_seen": 2368798720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028503510531594785,
      "loss": 1.3574,
      "theoretical_loss": 3.4115772091137457,
      "tokens_seen": 2368864256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850150451354062,
      "loss": 1.4121,
      "theoretical_loss": 3.411569662648932,
      "tokens_seen": 2368929792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849949849548646,
      "loss": 1.3017,
      "theoretical_loss": 3.4115621164513406,
      "tokens_seen": 2368995328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028497492477432294,
      "loss": 1.3189,
      "theoretical_loss": 3.411554570520955,
      "tokens_seen": 2369060864
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028495486459378136,
      "loss": 1.2421,
      "theoretical_loss": 3.411547024857759,
      "tokens_seen": 2369126400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849348044132397,
      "loss": 1.1777,
      "theoretical_loss": 3.411539479461735,
      "tokens_seen": 2369191936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849147442326981,
      "loss": 1.3888,
      "theoretical_loss": 3.4115319343328667,
      "tokens_seen": 2369257472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028489468405215644,
      "loss": 1.3279,
      "theoretical_loss": 3.4115243894711367,
      "tokens_seen": 2369323008
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028487462387161486,
      "loss": 1.2475,
      "theoretical_loss": 3.4115168448765285,
      "tokens_seen": 2369388544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848545636910732,
      "loss": 1.3397,
      "theoretical_loss": 3.4115093005490253,
      "tokens_seen": 2369454080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848345035105316,
      "loss": 1.3847,
      "theoretical_loss": 3.41150175648861,
      "tokens_seen": 2369519616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028481444332998995,
      "loss": 1.303,
      "theoretical_loss": 3.411494212695266,
      "tokens_seen": 2369585152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847943831494483,
      "loss": 1.2856,
      "theoretical_loss": 3.4114866691689762,
      "tokens_seen": 2369650688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847743229689067,
      "loss": 1.2327,
      "theoretical_loss": 3.4114791259097244,
      "tokens_seen": 2369716224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847542627883651,
      "loss": 1.3176,
      "theoretical_loss": 3.4114715829174926,
      "tokens_seen": 2369781760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028473420260782345,
      "loss": 1.3408,
      "theoretical_loss": 3.411464040192265,
      "tokens_seen": 2369847296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847141424272818,
      "loss": 1.3041,
      "theoretical_loss": 3.411456497734025,
      "tokens_seen": 2369912832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028469408224674023,
      "loss": 1.3031,
      "theoretical_loss": 3.4114489555427543,
      "tokens_seen": 2369978368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846740220661986,
      "loss": 1.3527,
      "theoretical_loss": 3.4114414136184372,
      "tokens_seen": 2370043904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028465396188565695,
      "loss": 1.3059,
      "theoretical_loss": 3.411433871961057,
      "tokens_seen": 2370109440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846339017051153,
      "loss": 1.3545,
      "theoretical_loss": 3.411426330570596,
      "tokens_seen": 2370174976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846138415245737,
      "loss": 1.3494,
      "theoretical_loss": 3.411418789447038,
      "tokens_seen": 2370240512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845937813440321,
      "loss": 1.3317,
      "theoretical_loss": 3.4114112485903663,
      "tokens_seen": 2370306048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028457372116349046,
      "loss": 1.3315,
      "theoretical_loss": 3.4114037080005635,
      "tokens_seen": 2370371584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845536609829488,
      "loss": 1.2688,
      "theoretical_loss": 3.411396167677613,
      "tokens_seen": 2370437120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028453360080240724,
      "loss": 1.216,
      "theoretical_loss": 3.4113886276214984,
      "tokens_seen": 2370502656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028451354062186565,
      "loss": 1.2913,
      "theoretical_loss": 3.4113810878322024,
      "tokens_seen": 2370568192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284493480441324,
      "loss": 1.3825,
      "theoretical_loss": 3.411373548309708,
      "tokens_seen": 2370633728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844734202607824,
      "loss": 1.3224,
      "theoretical_loss": 3.411366009053999,
      "tokens_seen": 2370699264
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028445336008024074,
      "loss": 1.2686,
      "theoretical_loss": 3.4113584700650588,
      "tokens_seen": 2370764800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844332998996991,
      "loss": 1.3098,
      "theoretical_loss": 3.411350931342869,
      "tokens_seen": 2370830336
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 863154,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3327744007110596,
      "objective/train/theoretical_loss": 3.4113433928874146,
      "objective/train/tokens_used": 741354976,
      "theoretical_loss": 3.4113433928874146,
      "tokens_seen": 2370895872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844132397191575,
      "loss": 1.2354,
      "theoretical_loss": 3.4113433928874146,
      "tokens_seen": 2370895872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843931795386159,
      "loss": 1.3593,
      "theoretical_loss": 3.4113358546986783,
      "tokens_seen": 2370961408
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028437311935807424,
      "loss": 1.3502,
      "theoretical_loss": 3.411328316776643,
      "tokens_seen": 2371026944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843530591775326,
      "loss": 1.4473,
      "theoretical_loss": 3.4113207791212914,
      "tokens_seen": 2371092480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284332998996991,
      "loss": 1.3367,
      "theoretical_loss": 3.4113132417326075,
      "tokens_seen": 2371158016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843129388164494,
      "loss": 1.3558,
      "theoretical_loss": 3.4113057046105744,
      "tokens_seen": 2371223552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028429287863590775,
      "loss": 1.3908,
      "theoretical_loss": 3.411298167755175,
      "tokens_seen": 2371289088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842728184553661,
      "loss": 1.3502,
      "theoretical_loss": 3.4112906311663926,
      "tokens_seen": 2371354624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028425275827482447,
      "loss": 1.3454,
      "theoretical_loss": 3.4112830948442108,
      "tokens_seen": 2371420160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842326980942829,
      "loss": 1.3213,
      "theoretical_loss": 3.4112755587886117,
      "tokens_seen": 2371485696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028421263791374125,
      "loss": 1.4537,
      "theoretical_loss": 3.4112680229995798,
      "tokens_seen": 2371551232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841925777331996,
      "loss": 1.3642,
      "theoretical_loss": 3.4112604874770978,
      "tokens_seen": 2371616768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284172517552658,
      "loss": 1.3696,
      "theoretical_loss": 3.4112529522211483,
      "tokens_seen": 2371682304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841524573721164,
      "loss": 1.3731,
      "theoretical_loss": 3.4112454172317155,
      "tokens_seen": 2371747840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028413239719157475,
      "loss": 1.2867,
      "theoretical_loss": 3.4112378825087823,
      "tokens_seen": 2371813376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841123370110331,
      "loss": 1.3119,
      "theoretical_loss": 3.4112303480523316,
      "tokens_seen": 2371878912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840922768304915,
      "loss": 1.3483,
      "theoretical_loss": 3.411222813862347,
      "tokens_seen": 2371944448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840722166499499,
      "loss": 1.3446,
      "theoretical_loss": 3.411215279938811,
      "tokens_seen": 2372009984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028405215646940825,
      "loss": 1.2628,
      "theoretical_loss": 3.4112077462817076,
      "tokens_seen": 2372075520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840320962888666,
      "loss": 1.2579,
      "theoretical_loss": 3.41120021289102,
      "tokens_seen": 2372141056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284012036108325,
      "loss": 1.3252,
      "theoretical_loss": 3.4111926797667307,
      "tokens_seen": 2372206592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028399197592778334,
      "loss": 1.2872,
      "theoretical_loss": 3.4111851469088235,
      "tokens_seen": 2372272128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028397191574724176,
      "loss": 1.2153,
      "theoretical_loss": 3.411177614317282,
      "tokens_seen": 2372337664
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002839518555667001,
      "loss": 1.349,
      "theoretical_loss": 3.4111700819920885,
      "tokens_seen": 2372403200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002839317953861585,
      "loss": 1.1373,
      "theoretical_loss": 3.4111625499332265,
      "tokens_seen": 2372468736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028391173520561684,
      "loss": 1.3837,
      "theoretical_loss": 3.4111550181406796,
      "tokens_seen": 2372534272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028389167502507526,
      "loss": 1.3397,
      "theoretical_loss": 3.4111474866144307,
      "tokens_seen": 2372599808
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838716148445336,
      "loss": 1.3024,
      "theoretical_loss": 3.4111399553544635,
      "tokens_seen": 2372665344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283851554663992,
      "loss": 1.3625,
      "theoretical_loss": 3.4111324243607606,
      "tokens_seen": 2372730880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028383149448345035,
      "loss": 1.367,
      "theoretical_loss": 3.411124893633305,
      "tokens_seen": 2372796416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838114343029087,
      "loss": 1.3969,
      "theoretical_loss": 3.4111173631720813,
      "tokens_seen": 2372861952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837913741223671,
      "loss": 1.196,
      "theoretical_loss": 3.4111098329770715,
      "tokens_seen": 2372927488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837713139418255,
      "loss": 1.3239,
      "theoretical_loss": 3.411102303048259,
      "tokens_seen": 2372993024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028375125376128385,
      "loss": 1.3059,
      "theoretical_loss": 3.411094773385628,
      "tokens_seen": 2373058560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837311935807422,
      "loss": 1.5196,
      "theoretical_loss": 3.4110872439891597,
      "tokens_seen": 2373124096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028371113340020063,
      "loss": 1.383,
      "theoretical_loss": 3.4110797148588397,
      "tokens_seen": 2373189632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283691073219659,
      "loss": 1.466,
      "theoretical_loss": 3.41107218599465,
      "tokens_seen": 2373255168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028367101303911735,
      "loss": 1.2908,
      "theoretical_loss": 3.4110646573965737,
      "tokens_seen": 2373320704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836509528585757,
      "loss": 1.3131,
      "theoretical_loss": 3.411057129064594,
      "tokens_seen": 2373386240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836308926780341,
      "loss": 1.3103,
      "theoretical_loss": 3.4110496009986955,
      "tokens_seen": 2373451776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836108324974925,
      "loss": 1.3384,
      "theoretical_loss": 3.4110420731988595,
      "tokens_seen": 2373517312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028359077231695086,
      "loss": 1.3036,
      "theoretical_loss": 3.411034545665071,
      "tokens_seen": 2373582848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835707121364092,
      "loss": 1.3782,
      "theoretical_loss": 3.411027018397312,
      "tokens_seen": 2373648384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835506519558676,
      "loss": 1.2992,
      "theoretical_loss": 3.4110194913955665,
      "tokens_seen": 2373713920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283530591775326,
      "loss": 1.2488,
      "theoretical_loss": 3.411011964659817,
      "tokens_seen": 2373779456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028351053159478436,
      "loss": 1.3812,
      "theoretical_loss": 3.4110044381900475,
      "tokens_seen": 2373844992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834904714142427,
      "loss": 1.212,
      "theoretical_loss": 3.410996911986241,
      "tokens_seen": 2373910528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834704112337011,
      "loss": 1.304,
      "theoretical_loss": 3.4109893860483806,
      "tokens_seen": 2373976064
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028345035105315945,
      "loss": 1.3324,
      "theoretical_loss": 3.41098186037645,
      "tokens_seen": 2374041600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028343029087261786,
      "loss": 1.1888,
      "theoretical_loss": 3.4109743349704322,
      "tokens_seen": 2374107136
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 864337,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4003491401672363,
      "objective/train/theoretical_loss": 3.41096680983031,
      "objective/train/tokens_used": 744631776,
      "theoretical_loss": 3.41096680983031,
      "tokens_seen": 2374172672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834102306920762,
      "loss": 1.3379,
      "theoretical_loss": 3.41096680983031,
      "tokens_seen": 2374172672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833901705115346,
      "loss": 1.3047,
      "theoretical_loss": 3.4109592849560673,
      "tokens_seen": 2374238208
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028337011033099295,
      "loss": 1.2141,
      "theoretical_loss": 3.4109517603476873,
      "tokens_seen": 2374303744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028335005015045137,
      "loss": 1.3176,
      "theoretical_loss": 3.410944236005153,
      "tokens_seen": 2374369280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028332998996990973,
      "loss": 1.3421,
      "theoretical_loss": 3.4109367119284477,
      "tokens_seen": 2374434816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833099297893681,
      "loss": 1.3578,
      "theoretical_loss": 3.4109291881175547,
      "tokens_seen": 2374500352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028328986960882645,
      "loss": 1.3424,
      "theoretical_loss": 3.4109216645724576,
      "tokens_seen": 2374565888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832698094282848,
      "loss": 1.2745,
      "theoretical_loss": 3.4109141412931394,
      "tokens_seen": 2374631424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028324974924774323,
      "loss": 1.3234,
      "theoretical_loss": 3.4109066182795833,
      "tokens_seen": 2374696960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832296890672016,
      "loss": 1.2722,
      "theoretical_loss": 3.410899095531773,
      "tokens_seen": 2374762496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028320962888665996,
      "loss": 1.3156,
      "theoretical_loss": 3.410891573049691,
      "tokens_seen": 2374828032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831895687061183,
      "loss": 1.3126,
      "theoretical_loss": 3.410884050833321,
      "tokens_seen": 2374893568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028316950852557674,
      "loss": 1.3172,
      "theoretical_loss": 3.4108765288826466,
      "tokens_seen": 2374959104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831494483450351,
      "loss": 1.387,
      "theoretical_loss": 3.410869007197651,
      "tokens_seen": 2375024640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028312938816449346,
      "loss": 1.4503,
      "theoretical_loss": 3.410861485778317,
      "tokens_seen": 2375090176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831093279839518,
      "loss": 1.3157,
      "theoretical_loss": 3.410853964624628,
      "tokens_seen": 2375155712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830892678034102,
      "loss": 1.3085,
      "theoretical_loss": 3.4108464437365678,
      "tokens_seen": 2375221248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830692076228686,
      "loss": 1.3225,
      "theoretical_loss": 3.410838923114119,
      "tokens_seen": 2375286784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028304914744232696,
      "loss": 1.2802,
      "theoretical_loss": 3.4108314027572657,
      "tokens_seen": 2375352320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830290872617854,
      "loss": 1.2924,
      "theoretical_loss": 3.4108238826659902,
      "tokens_seen": 2375417856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028300902708124374,
      "loss": 1.3533,
      "theoretical_loss": 3.4108163628402766,
      "tokens_seen": 2375483392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028298896690070216,
      "loss": 1.338,
      "theoretical_loss": 3.4108088432801082,
      "tokens_seen": 2375548928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829689067201605,
      "loss": 1.2492,
      "theoretical_loss": 3.4108013239854675,
      "tokens_seen": 2375614464
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829488465396189,
      "loss": 1.315,
      "theoretical_loss": 3.410793804956339,
      "tokens_seen": 2375680000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028292878635907725,
      "loss": 1.4085,
      "theoretical_loss": 3.4107862861927045,
      "tokens_seen": 2375745536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028290872617853566,
      "loss": 1.3728,
      "theoretical_loss": 3.4107787676945485,
      "tokens_seen": 2375811072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282888665997994,
      "loss": 1.2542,
      "theoretical_loss": 3.4107712494618543,
      "tokens_seen": 2375876608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828686058174524,
      "loss": 1.3712,
      "theoretical_loss": 3.410763731494604,
      "tokens_seen": 2375942144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028284854563691075,
      "loss": 1.3305,
      "theoretical_loss": 3.4107562137927827,
      "tokens_seen": 2376007680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828284854563691,
      "loss": 1.316,
      "theoretical_loss": 3.410748696356372,
      "tokens_seen": 2376073216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828084252758275,
      "loss": 1.3245,
      "theoretical_loss": 3.410741179185356,
      "tokens_seen": 2376138752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827883650952859,
      "loss": 1.247,
      "theoretical_loss": 3.4107336622797186,
      "tokens_seen": 2376204288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028276830491474425,
      "loss": 1.3875,
      "theoretical_loss": 3.410726145639442,
      "tokens_seen": 2376269824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827482447342026,
      "loss": 1.2788,
      "theoretical_loss": 3.41071862926451,
      "tokens_seen": 2376335360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028272818455366103,
      "loss": 1.2089,
      "theoretical_loss": 3.410711113154906,
      "tokens_seen": 2376400896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827081243731194,
      "loss": 1.3753,
      "theoretical_loss": 3.4107035973106132,
      "tokens_seen": 2376466432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028268806419257775,
      "loss": 1.3201,
      "theoretical_loss": 3.410696081731615,
      "tokens_seen": 2376531968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826680040120361,
      "loss": 1.3418,
      "theoretical_loss": 3.4106885664178948,
      "tokens_seen": 2376597504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826479438314945,
      "loss": 1.3258,
      "theoretical_loss": 3.4106810513694357,
      "tokens_seen": 2376663040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826278836509529,
      "loss": 1.2971,
      "theoretical_loss": 3.410673536586221,
      "tokens_seen": 2376728576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028260782347041126,
      "loss": 1.2274,
      "theoretical_loss": 3.4106660220682343,
      "tokens_seen": 2376794112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825877632898696,
      "loss": 1.2881,
      "theoretical_loss": 3.4106585078154583,
      "tokens_seen": 2376859648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282567703109328,
      "loss": 1.3114,
      "theoretical_loss": 3.4106509938278773,
      "tokens_seen": 2376925184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825476429287864,
      "loss": 1.2747,
      "theoretical_loss": 3.410643480105474,
      "tokens_seen": 2376990720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028252758274824476,
      "loss": 1.3043,
      "theoretical_loss": 3.4106359666482318,
      "tokens_seen": 2377056256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825075225677031,
      "loss": 1.3016,
      "theoretical_loss": 3.410628453456134,
      "tokens_seen": 2377121792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824874623871615,
      "loss": 1.3818,
      "theoretical_loss": 3.4106209405291645,
      "tokens_seen": 2377187328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028246740220661985,
      "loss": 1.3147,
      "theoretical_loss": 3.4106134278673057,
      "tokens_seen": 2377252864
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028244734202607826,
      "loss": 1.3639,
      "theoretical_loss": 3.4106059154705415,
      "tokens_seen": 2377318400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824272818455366,
      "loss": 1.3316,
      "theoretical_loss": 3.410598403338855,
      "tokens_seen": 2377383936
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 865580,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3770675659179688,
      "objective/train/theoretical_loss": 3.41059089147223,
      "objective/train/tokens_used": 747908576,
      "theoretical_loss": 3.41059089147223,
      "tokens_seen": 2377449472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282407221664995,
      "loss": 1.3573,
      "theoretical_loss": 3.41059089147223,
      "tokens_seen": 2377449472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028238716148445335,
      "loss": 1.3974,
      "theoretical_loss": 3.410583379870649,
      "tokens_seen": 2377515008
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028236710130391177,
      "loss": 1.2249,
      "theoretical_loss": 3.410575868534096,
      "tokens_seen": 2377580544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028234704112337013,
      "loss": 1.2719,
      "theoretical_loss": 3.4105683574625543,
      "tokens_seen": 2377646080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002823269809428285,
      "loss": 1.2301,
      "theoretical_loss": 3.4105608466560073,
      "tokens_seen": 2377711616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028230692076228685,
      "loss": 1.3313,
      "theoretical_loss": 3.410553336114438,
      "tokens_seen": 2377777152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002822868605817452,
      "loss": 1.2619,
      "theoretical_loss": 3.41054582583783,
      "tokens_seen": 2377842688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028226680040120363,
      "loss": 1.3882,
      "theoretical_loss": 3.4105383158261664,
      "tokens_seen": 2377908224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282246740220662,
      "loss": 1.2948,
      "theoretical_loss": 3.4105308060794313,
      "tokens_seen": 2377973760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028222668004012036,
      "loss": 1.3541,
      "theoretical_loss": 3.410523296597607,
      "tokens_seen": 2378039296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002822066198595787,
      "loss": 1.3207,
      "theoretical_loss": 3.4105157873806773,
      "tokens_seen": 2378104832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028218655967903714,
      "loss": 1.3738,
      "theoretical_loss": 3.410508278428626,
      "tokens_seen": 2378170368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821664994984955,
      "loss": 1.2812,
      "theoretical_loss": 3.410500769741436,
      "tokens_seen": 2378235904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028214643931795386,
      "loss": 1.3414,
      "theoretical_loss": 3.41049326131909,
      "tokens_seen": 2378301440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821263791374122,
      "loss": 1.3073,
      "theoretical_loss": 3.4104857531615727,
      "tokens_seen": 2378366976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821063189568706,
      "loss": 1.3326,
      "theoretical_loss": 3.410478245268867,
      "tokens_seen": 2378432512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282086258776329,
      "loss": 1.3411,
      "theoretical_loss": 3.410470737640956,
      "tokens_seen": 2378498048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028206619859578736,
      "loss": 1.316,
      "theoretical_loss": 3.410463230277823,
      "tokens_seen": 2378563584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820461384152457,
      "loss": 1.3343,
      "theoretical_loss": 3.4104557231794512,
      "tokens_seen": 2378629120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820260782347041,
      "loss": 1.2912,
      "theoretical_loss": 3.410448216345825,
      "tokens_seen": 2378694656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820060180541625,
      "loss": 1.2536,
      "theoretical_loss": 3.4104407097769265,
      "tokens_seen": 2378760192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028198595787362087,
      "loss": 1.2707,
      "theoretical_loss": 3.4104332034727403,
      "tokens_seen": 2378825728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028196589769307923,
      "loss": 1.3409,
      "theoretical_loss": 3.4104256974332485,
      "tokens_seen": 2378891264
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002819458375125376,
      "loss": 1.3658,
      "theoretical_loss": 3.4104181916584357,
      "tokens_seen": 2378956800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281925777331996,
      "loss": 1.3991,
      "theoretical_loss": 3.410410686148284,
      "tokens_seen": 2379022336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028190571715145437,
      "loss": 1.4677,
      "theoretical_loss": 3.4104031809027777,
      "tokens_seen": 2379087872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028188565697091273,
      "loss": 1.3324,
      "theoretical_loss": 3.4103956759219,
      "tokens_seen": 2379153408
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002818655967903711,
      "loss": 1.3083,
      "theoretical_loss": 3.4103881712056343,
      "tokens_seen": 2379218944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028184553660982946,
      "loss": 1.2914,
      "theoretical_loss": 3.410380666753964,
      "tokens_seen": 2379284480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002818254764292879,
      "loss": 1.4032,
      "theoretical_loss": 3.4103731625668723,
      "tokens_seen": 2379350016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028180541624874624,
      "loss": 1.2813,
      "theoretical_loss": 3.4103656586443423,
      "tokens_seen": 2379415552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817853560682046,
      "loss": 1.2874,
      "theoretical_loss": 3.410358154986358,
      "tokens_seen": 2379481088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028176529588766296,
      "loss": 1.356,
      "theoretical_loss": 3.4103506515929025,
      "tokens_seen": 2379546624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817452357071214,
      "loss": 1.2697,
      "theoretical_loss": 3.4103431484639595,
      "tokens_seen": 2379612160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028172517552657974,
      "loss": 1.3486,
      "theoretical_loss": 3.4103356455995115,
      "tokens_seen": 2379677696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817051153460381,
      "loss": 1.3395,
      "theoretical_loss": 3.410328142999543,
      "tokens_seen": 2379743232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028168505516549646,
      "loss": 1.3587,
      "theoretical_loss": 3.410320640664037,
      "tokens_seen": 2379808768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816649949849548,
      "loss": 1.2731,
      "theoretical_loss": 3.4103131385929766,
      "tokens_seen": 2379874304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028164493480441324,
      "loss": 1.392,
      "theoretical_loss": 3.410305636786345,
      "tokens_seen": 2379939840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816248746238716,
      "loss": 1.3405,
      "theoretical_loss": 3.4102981352441266,
      "tokens_seen": 2380005376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028160481444332997,
      "loss": 1.2791,
      "theoretical_loss": 3.410290633966304,
      "tokens_seen": 2380070912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028158475426278833,
      "loss": 1.3354,
      "theoretical_loss": 3.410283132952861,
      "tokens_seen": 2380136448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028156469408224674,
      "loss": 1.1994,
      "theoretical_loss": 3.4102756322037804,
      "tokens_seen": 2380201984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815446339017051,
      "loss": 1.3435,
      "theoretical_loss": 3.4102681317190457,
      "tokens_seen": 2380267520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815245737211635,
      "loss": 1.2966,
      "theoretical_loss": 3.4102606314986414,
      "tokens_seen": 2380333056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815045135406219,
      "loss": 1.4798,
      "theoretical_loss": 3.4102531315425497,
      "tokens_seen": 2380398592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028148445336008025,
      "loss": 1.2646,
      "theoretical_loss": 3.4102456318507546,
      "tokens_seen": 2380464128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028146439317953866,
      "loss": 1.3102,
      "theoretical_loss": 3.410238132423239,
      "tokens_seen": 2380529664
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281444332998997,
      "loss": 1.3332,
      "theoretical_loss": 3.410230633259987,
      "tokens_seen": 2380595200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002814242728184554,
      "loss": 1.3754,
      "theoretical_loss": 3.4102231343609812,
      "tokens_seen": 2380660736
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 866853,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4650294780731201,
      "objective/train/theoretical_loss": 3.410215635726206,
      "objective/train/tokens_used": 751185376,
      "theoretical_loss": 3.410215635726206,
      "tokens_seen": 2380726272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028140421263791375,
      "loss": 1.3166,
      "theoretical_loss": 3.410215635726206,
      "tokens_seen": 2380726272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028138415245737217,
      "loss": 1.3479,
      "theoretical_loss": 3.410208137355644,
      "tokens_seen": 2380791808
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028136409227683053,
      "loss": 1.356,
      "theoretical_loss": 3.410200639249279,
      "tokens_seen": 2380857344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813440320962889,
      "loss": 1.3532,
      "theoretical_loss": 3.4101931414070945,
      "tokens_seen": 2380922880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028132397191574725,
      "loss": 1.2532,
      "theoretical_loss": 3.4101856438290734,
      "tokens_seen": 2380988416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813039117352056,
      "loss": 1.2538,
      "theoretical_loss": 3.4101781465151997,
      "tokens_seen": 2381053952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028128385155466403,
      "loss": 1.3647,
      "theoretical_loss": 3.4101706494654564,
      "tokens_seen": 2381119488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812637913741224,
      "loss": 1.3917,
      "theoretical_loss": 3.4101631526798273,
      "tokens_seen": 2381185024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028124373119358076,
      "loss": 1.3517,
      "theoretical_loss": 3.4101556561582957,
      "tokens_seen": 2381250560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812236710130391,
      "loss": 1.327,
      "theoretical_loss": 3.410148159900845,
      "tokens_seen": 2381316096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028120361083249754,
      "loss": 1.4366,
      "theoretical_loss": 3.410140663907458,
      "tokens_seen": 2381381632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811835506519559,
      "loss": 1.2831,
      "theoretical_loss": 3.4101331681781195,
      "tokens_seen": 2381447168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028116349047141426,
      "loss": 1.3691,
      "theoretical_loss": 3.4101256727128115,
      "tokens_seen": 2381512704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811434302908726,
      "loss": 1.2812,
      "theoretical_loss": 3.410118177511519,
      "tokens_seen": 2381578240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281123370110331,
      "loss": 1.2811,
      "theoretical_loss": 3.410110682574224,
      "tokens_seen": 2381643776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811033099297894,
      "loss": 1.4121,
      "theoretical_loss": 3.4101031879009103,
      "tokens_seen": 2381709312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028108324974924776,
      "loss": 1.3692,
      "theoretical_loss": 3.4100956934915616,
      "tokens_seen": 2381774848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810631895687061,
      "loss": 1.3754,
      "theoretical_loss": 3.4100881993461614,
      "tokens_seen": 2381840384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810431293881645,
      "loss": 1.2967,
      "theoretical_loss": 3.4100807054646927,
      "tokens_seen": 2381905920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810230692076229,
      "loss": 1.3883,
      "theoretical_loss": 3.4100732118471395,
      "tokens_seen": 2381971456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028100300902708127,
      "loss": 1.2094,
      "theoretical_loss": 3.410065718493485,
      "tokens_seen": 2382036992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028098294884653963,
      "loss": 1.3485,
      "theoretical_loss": 3.410058225403713,
      "tokens_seen": 2382102528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280962888665998,
      "loss": 1.3817,
      "theoretical_loss": 3.410050732577806,
      "tokens_seen": 2382168064
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028094282848545635,
      "loss": 1.4285,
      "theoretical_loss": 3.410043240015748,
      "tokens_seen": 2382233600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028092276830491477,
      "loss": 1.3323,
      "theoretical_loss": 3.4100357477175227,
      "tokens_seen": 2382299136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028090270812437313,
      "loss": 1.2598,
      "theoretical_loss": 3.4100282556831134,
      "tokens_seen": 2382364672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808826479438315,
      "loss": 1.3228,
      "theoretical_loss": 3.4100207639125033,
      "tokens_seen": 2382430208
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028086258776328986,
      "loss": 1.4418,
      "theoretical_loss": 3.4100132724056764,
      "tokens_seen": 2382495744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808425275827483,
      "loss": 1.308,
      "theoretical_loss": 3.4100057811626154,
      "tokens_seen": 2382561280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028082246740220664,
      "loss": 1.3924,
      "theoretical_loss": 3.4099982901833044,
      "tokens_seen": 2382626816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280802407221665,
      "loss": 1.3129,
      "theoretical_loss": 3.4099907994677263,
      "tokens_seen": 2382692352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028078234704112336,
      "loss": 1.3365,
      "theoretical_loss": 3.409983309015865,
      "tokens_seen": 2382757888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807622868605818,
      "loss": 1.3387,
      "theoretical_loss": 3.409975818827704,
      "tokens_seen": 2382823424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028074222668004014,
      "loss": 1.3871,
      "theoretical_loss": 3.4099683289032265,
      "tokens_seen": 2382888960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807221664994985,
      "loss": 1.3555,
      "theoretical_loss": 3.409960839242416,
      "tokens_seen": 2382954496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028070210631895686,
      "loss": 1.2642,
      "theoretical_loss": 3.409953349845256,
      "tokens_seen": 2383020032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002806820461384152,
      "loss": 1.2793,
      "theoretical_loss": 3.4099458607117303,
      "tokens_seen": 2383085568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028066198595787364,
      "loss": 1.3662,
      "theoretical_loss": 3.409938371841822,
      "tokens_seen": 2383151104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280641925777332,
      "loss": 1.3333,
      "theoretical_loss": 3.4099308832355146,
      "tokens_seen": 2383216640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028062186559679037,
      "loss": 1.4395,
      "theoretical_loss": 3.4099233948927914,
      "tokens_seen": 2383282176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028060180541624873,
      "loss": 1.3788,
      "theoretical_loss": 3.4099159068136364,
      "tokens_seen": 2383347712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028058174523570715,
      "loss": 1.2366,
      "theoretical_loss": 3.4099084189980324,
      "tokens_seen": 2383413248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805616850551655,
      "loss": 1.2712,
      "theoretical_loss": 3.4099009314459634,
      "tokens_seen": 2383478784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028054162487462387,
      "loss": 1.2927,
      "theoretical_loss": 3.409893444157413,
      "tokens_seen": 2383544320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028052156469408223,
      "loss": 1.2963,
      "theoretical_loss": 3.409885957132364,
      "tokens_seen": 2383609856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805015045135406,
      "loss": 1.4209,
      "theoretical_loss": 3.409878470370801,
      "tokens_seen": 2383675392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280481444332999,
      "loss": 1.39,
      "theoretical_loss": 3.4098709838727057,
      "tokens_seen": 2383740928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028046138415245737,
      "loss": 1.3058,
      "theoretical_loss": 3.4098634976380633,
      "tokens_seen": 2383806464
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028044132397191574,
      "loss": 1.2907,
      "theoretical_loss": 3.409856011666857,
      "tokens_seen": 2383872000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804212637913741,
      "loss": 1.285,
      "theoretical_loss": 3.4098485259590694,
      "tokens_seen": 2383937536
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 868029,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2294628620147705,
      "objective/train/theoretical_loss": 3.409841040514684,
      "objective/train/tokens_used": 754462176,
      "theoretical_loss": 3.409841040514684,
      "tokens_seen": 2384003072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804012036108325,
      "loss": 1.2541,
      "theoretical_loss": 3.409841040514684,
      "tokens_seen": 2384003072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002803811434302909,
      "loss": 1.2648,
      "theoretical_loss": 3.4098335553336856,
      "tokens_seen": 2384068608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028036108324974924,
      "loss": 1.3008,
      "theoretical_loss": 3.409826070416057,
      "tokens_seen": 2384134144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002803410230692076,
      "loss": 1.2395,
      "theoretical_loss": 3.4098185857617813,
      "tokens_seen": 2384199680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028032096288866596,
      "loss": 1.3242,
      "theoretical_loss": 3.4098111013708423,
      "tokens_seen": 2384265216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002803009027081244,
      "loss": 1.3508,
      "theoretical_loss": 3.409803617243224,
      "tokens_seen": 2384330752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028028084252758274,
      "loss": 1.3576,
      "theoretical_loss": 3.4097961333789084,
      "tokens_seen": 2384396288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802607823470411,
      "loss": 1.2337,
      "theoretical_loss": 3.4097886497778807,
      "tokens_seen": 2384461824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028024072216649947,
      "loss": 1.3077,
      "theoretical_loss": 3.4097811664401236,
      "tokens_seen": 2384527360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802206619859579,
      "loss": 1.3668,
      "theoretical_loss": 3.4097736833656205,
      "tokens_seen": 2384592896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028020060180541624,
      "loss": 1.3489,
      "theoretical_loss": 3.409766200554355,
      "tokens_seen": 2384658432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801805416248746,
      "loss": 1.3607,
      "theoretical_loss": 3.409758718006311,
      "tokens_seen": 2384723968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028016048144433297,
      "loss": 1.3002,
      "theoretical_loss": 3.4097512357214717,
      "tokens_seen": 2384789504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028014042126379133,
      "loss": 1.3373,
      "theoretical_loss": 3.4097437536998205,
      "tokens_seen": 2384855040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028012036108324975,
      "loss": 1.2911,
      "theoretical_loss": 3.409736271941341,
      "tokens_seen": 2384920576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801003009027081,
      "loss": 1.2898,
      "theoretical_loss": 3.4097287904460165,
      "tokens_seen": 2384986112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028008024072216647,
      "loss": 1.3524,
      "theoretical_loss": 3.409721309213831,
      "tokens_seen": 2385051648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028006018054162483,
      "loss": 1.3748,
      "theoretical_loss": 3.409713828244768,
      "tokens_seen": 2385117184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028004012036108325,
      "loss": 1.2955,
      "theoretical_loss": 3.4097063475388105,
      "tokens_seen": 2385182720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028002006018054167,
      "loss": 1.269,
      "theoretical_loss": 3.4096988670959427,
      "tokens_seen": 2385248256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028000000000000003,
      "loss": 1.2192,
      "theoretical_loss": 3.409691386916147,
      "tokens_seen": 2385313792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002799799398194584,
      "loss": 1.292,
      "theoretical_loss": 3.4096839069994083,
      "tokens_seen": 2385379328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027995987963891675,
      "loss": 1.3055,
      "theoretical_loss": 3.409676427345709,
      "tokens_seen": 2385444864
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027993981945837517,
      "loss": 1.2948,
      "theoretical_loss": 3.4096689479550335,
      "tokens_seen": 2385510400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027991975927783353,
      "loss": 1.2718,
      "theoretical_loss": 3.4096614688273648,
      "tokens_seen": 2385575936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798996990972919,
      "loss": 1.3102,
      "theoretical_loss": 3.4096539899626865,
      "tokens_seen": 2385641472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027987963891675026,
      "loss": 1.3307,
      "theoretical_loss": 3.4096465113609824,
      "tokens_seen": 2385707008
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798595787362087,
      "loss": 1.3586,
      "theoretical_loss": 3.4096390330222355,
      "tokens_seen": 2385772544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027983951855566704,
      "loss": 1.3346,
      "theoretical_loss": 3.4096315549464298,
      "tokens_seen": 2385838080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798194583751254,
      "loss": 1.2735,
      "theoretical_loss": 3.4096240771335484,
      "tokens_seen": 2385903616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027979939819458376,
      "loss": 1.2501,
      "theoretical_loss": 3.4096165995835754,
      "tokens_seen": 2385969152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797793380140422,
      "loss": 1.3442,
      "theoretical_loss": 3.409609122296494,
      "tokens_seen": 2386034688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027975927783350054,
      "loss": 1.3476,
      "theoretical_loss": 3.4096016452722875,
      "tokens_seen": 2386100224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797392176529589,
      "loss": 1.3751,
      "theoretical_loss": 3.4095941685109397,
      "tokens_seen": 2386165760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027971915747241726,
      "loss": 1.3278,
      "theoretical_loss": 3.4095866920124345,
      "tokens_seen": 2386231296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796990972918756,
      "loss": 1.3246,
      "theoretical_loss": 3.4095792157767546,
      "tokens_seen": 2386296832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027967903711133404,
      "loss": 1.2611,
      "theoretical_loss": 3.4095717398038845,
      "tokens_seen": 2386362368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796589769307924,
      "loss": 1.372,
      "theoretical_loss": 3.4095642640938073,
      "tokens_seen": 2386427904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027963891675025077,
      "loss": 1.3916,
      "theoretical_loss": 3.409556788646506,
      "tokens_seen": 2386493440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027961885656970913,
      "loss": 1.3651,
      "theoretical_loss": 3.4095493134619654,
      "tokens_seen": 2386558976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027959879638916755,
      "loss": 1.4094,
      "theoretical_loss": 3.409541838540168,
      "tokens_seen": 2386624512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795787362086259,
      "loss": 1.3346,
      "theoretical_loss": 3.4095343638810975,
      "tokens_seen": 2386690048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027955867602808427,
      "loss": 1.2631,
      "theoretical_loss": 3.4095268894847375,
      "tokens_seen": 2386755584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027953861584754263,
      "loss": 1.4029,
      "theoretical_loss": 3.409519415351072,
      "tokens_seen": 2386821120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279518555667001,
      "loss": 1.3067,
      "theoretical_loss": 3.4095119414800843,
      "tokens_seen": 2386886656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794984954864594,
      "loss": 1.223,
      "theoretical_loss": 3.4095044678717574,
      "tokens_seen": 2386952192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794784353059178,
      "loss": 1.2985,
      "theoretical_loss": 3.409496994526076,
      "tokens_seen": 2387017728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027945837512537614,
      "loss": 1.2422,
      "theoretical_loss": 3.4094895214430228,
      "tokens_seen": 2387083264
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794383149448345,
      "loss": 1.4276,
      "theoretical_loss": 3.409482048622581,
      "tokens_seen": 2387148800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794182547642929,
      "loss": 1.3104,
      "theoretical_loss": 3.4094745760647354,
      "tokens_seen": 2387214336
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 869259,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3626919984817505,
      "objective/train/theoretical_loss": 3.409467103769469,
      "objective/train/tokens_used": 757738976,
      "theoretical_loss": 3.409467103769469,
      "tokens_seen": 2387279872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793981945837513,
      "loss": 1.2977,
      "theoretical_loss": 3.409467103769469,
      "tokens_seen": 2387279872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027937813440320964,
      "loss": 1.4006,
      "theoretical_loss": 3.4094596317367647,
      "tokens_seen": 2387345408
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279358074222668,
      "loss": 1.185,
      "theoretical_loss": 3.409452159966607,
      "tokens_seen": 2387410944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027933801404212636,
      "loss": 1.4117,
      "theoretical_loss": 3.409444688458979,
      "tokens_seen": 2387476480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793179538615848,
      "loss": 1.2789,
      "theoretical_loss": 3.4094372172138643,
      "tokens_seen": 2387542016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027929789368104314,
      "loss": 1.4358,
      "theoretical_loss": 3.4094297462312464,
      "tokens_seen": 2387607552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792778335005015,
      "loss": 1.4048,
      "theoretical_loss": 3.409422275511109,
      "tokens_seen": 2387673088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027925777331995987,
      "loss": 1.2656,
      "theoretical_loss": 3.409414805053436,
      "tokens_seen": 2387738624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792377131394183,
      "loss": 1.2627,
      "theoretical_loss": 3.4094073348582103,
      "tokens_seen": 2387804160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027921765295887664,
      "loss": 1.3163,
      "theoretical_loss": 3.409399864925416,
      "tokens_seen": 2387869696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279197592778335,
      "loss": 1.3376,
      "theoretical_loss": 3.4093923952550367,
      "tokens_seen": 2387935232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027917753259779337,
      "loss": 1.3544,
      "theoretical_loss": 3.409384925847055,
      "tokens_seen": 2388000768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027915747241725173,
      "loss": 1.3194,
      "theoretical_loss": 3.4093774567014563,
      "tokens_seen": 2388066304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027913741223671015,
      "loss": 1.2566,
      "theoretical_loss": 3.4093699878182226,
      "tokens_seen": 2388131840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002791173520561685,
      "loss": 1.3513,
      "theoretical_loss": 3.4093625191973382,
      "tokens_seen": 2388197376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027909729187562687,
      "loss": 1.4182,
      "theoretical_loss": 3.4093550508387866,
      "tokens_seen": 2388262912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027907723169508523,
      "loss": 1.3653,
      "theoretical_loss": 3.4093475827425515,
      "tokens_seen": 2388328448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027905717151454365,
      "loss": 1.2908,
      "theoretical_loss": 3.4093401149086158,
      "tokens_seen": 2388393984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279037111334002,
      "loss": 1.4036,
      "theoretical_loss": 3.4093326473369636,
      "tokens_seen": 2388459520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002790170511534604,
      "loss": 1.3354,
      "theoretical_loss": 3.409325180027579,
      "tokens_seen": 2388525056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027899699097291874,
      "loss": 1.3056,
      "theoretical_loss": 3.4093177129804446,
      "tokens_seen": 2388590592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789769307923771,
      "loss": 1.2023,
      "theoretical_loss": 3.4093102461955445,
      "tokens_seen": 2388656128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789568706118355,
      "loss": 1.2744,
      "theoretical_loss": 3.4093027796728625,
      "tokens_seen": 2388721664
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789368104312939,
      "loss": 1.361,
      "theoretical_loss": 3.409295313412382,
      "tokens_seen": 2388787200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027891675025075224,
      "loss": 1.3486,
      "theoretical_loss": 3.4092878474140864,
      "tokens_seen": 2388852736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788966900702106,
      "loss": 1.248,
      "theoretical_loss": 3.4092803816779593,
      "tokens_seen": 2388918272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278876629889669,
      "loss": 1.2359,
      "theoretical_loss": 3.4092729162039848,
      "tokens_seen": 2388983808
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788565697091274,
      "loss": 1.3306,
      "theoretical_loss": 3.4092654509921463,
      "tokens_seen": 2389049344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027883650952858574,
      "loss": 1.3045,
      "theoretical_loss": 3.4092579860424266,
      "tokens_seen": 2389114880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788164493480441,
      "loss": 1.4198,
      "theoretical_loss": 3.4092505213548105,
      "tokens_seen": 2389180416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027879638916750247,
      "loss": 1.2892,
      "theoretical_loss": 3.4092430569292813,
      "tokens_seen": 2389245952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787763289869609,
      "loss": 1.3572,
      "theoretical_loss": 3.409235592765822,
      "tokens_seen": 2389311488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027875626880641925,
      "loss": 1.2506,
      "theoretical_loss": 3.4092281288644166,
      "tokens_seen": 2389377024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787362086258776,
      "loss": 1.3228,
      "theoretical_loss": 3.4092206652250487,
      "tokens_seen": 2389442560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027871614844533597,
      "loss": 1.2606,
      "theoretical_loss": 3.4092132018477024,
      "tokens_seen": 2389508096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786960882647944,
      "loss": 1.2479,
      "theoretical_loss": 3.40920573873236,
      "tokens_seen": 2389573632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027867602808425275,
      "loss": 1.4392,
      "theoretical_loss": 3.4091982758790067,
      "tokens_seen": 2389639168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786559679037111,
      "loss": 1.3376,
      "theoretical_loss": 3.4091908132876254,
      "tokens_seen": 2389704704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786359077231695,
      "loss": 1.3837,
      "theoretical_loss": 3.4091833509581995,
      "tokens_seen": 2389770240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786158475426279,
      "loss": 1.379,
      "theoretical_loss": 3.409175888890713,
      "tokens_seen": 2389835776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027859578736208625,
      "loss": 1.3274,
      "theoretical_loss": 3.409168427085149,
      "tokens_seen": 2389901312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002785757271815446,
      "loss": 1.3027,
      "theoretical_loss": 3.4091609655414916,
      "tokens_seen": 2389966848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278555667001003,
      "loss": 1.2772,
      "theoretical_loss": 3.4091535042597245,
      "tokens_seen": 2390032384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027853560682046134,
      "loss": 1.3184,
      "theoretical_loss": 3.409146043239831,
      "tokens_seen": 2390097920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002785155466399198,
      "loss": 1.328,
      "theoretical_loss": 3.409138582481795,
      "tokens_seen": 2390163456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784954864593782,
      "loss": 1.2777,
      "theoretical_loss": 3.4091311219855998,
      "tokens_seen": 2390228992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027847542627883654,
      "loss": 1.2834,
      "theoretical_loss": 3.409123661751229,
      "tokens_seen": 2390294528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784553660982949,
      "loss": 1.3037,
      "theoretical_loss": 3.4091162017786667,
      "tokens_seen": 2390360064
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784353059177533,
      "loss": 1.3447,
      "theoretical_loss": 3.4091087420678963,
      "tokens_seen": 2390425600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784152457372117,
      "loss": 1.3253,
      "theoretical_loss": 3.4091012826189013,
      "tokens_seen": 2390491136
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 870459,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.38811194896698,
      "objective/train/theoretical_loss": 3.4090938234316654,
      "objective/train/tokens_used": 761015776,
      "theoretical_loss": 3.4090938234316654,
      "tokens_seen": 2390556672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027839518555667004,
      "loss": 1.2951,
      "theoretical_loss": 3.4090938234316654,
      "tokens_seen": 2390556672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783751253761284,
      "loss": 1.3781,
      "theoretical_loss": 3.4090863645061726,
      "tokens_seen": 2390622208
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027835506519558676,
      "loss": 1.1978,
      "theoretical_loss": 3.409078905842406,
      "tokens_seen": 2390687744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783350050150452,
      "loss": 1.3197,
      "theoretical_loss": 3.4090714474403496,
      "tokens_seen": 2390753280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027831494483450354,
      "loss": 1.3465,
      "theoretical_loss": 3.409063989299987,
      "tokens_seen": 2390818816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782948846539619,
      "loss": 1.3921,
      "theoretical_loss": 3.409056531421302,
      "tokens_seen": 2390884352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027827482447342027,
      "loss": 1.3296,
      "theoretical_loss": 3.409049073804278,
      "tokens_seen": 2390949888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782547642928787,
      "loss": 1.4247,
      "theoretical_loss": 3.409041616448898,
      "tokens_seen": 2391015424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027823470411233705,
      "loss": 1.2734,
      "theoretical_loss": 3.4090341593551465,
      "tokens_seen": 2391080960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782146439317954,
      "loss": 1.2949,
      "theoretical_loss": 3.409026702523007,
      "tokens_seen": 2391146496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027819458375125377,
      "loss": 1.4054,
      "theoretical_loss": 3.4090192459524635,
      "tokens_seen": 2391212032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027817452357071213,
      "loss": 1.3173,
      "theoretical_loss": 3.4090117896434995,
      "tokens_seen": 2391277568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027815446339017055,
      "loss": 1.2453,
      "theoretical_loss": 3.409004333596098,
      "tokens_seen": 2391343104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781344032096289,
      "loss": 1.3776,
      "theoretical_loss": 3.4089968778102424,
      "tokens_seen": 2391408640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781143430290873,
      "loss": 1.2901,
      "theoretical_loss": 3.4089894222859183,
      "tokens_seen": 2391474176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027809428284854564,
      "loss": 1.4012,
      "theoretical_loss": 3.408981967023107,
      "tokens_seen": 2391539712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027807422266800405,
      "loss": 1.2961,
      "theoretical_loss": 3.408974512021794,
      "tokens_seen": 2391605248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002780541624874624,
      "loss": 1.2493,
      "theoretical_loss": 3.408967057281962,
      "tokens_seen": 2391670784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002780341023069208,
      "loss": 1.3881,
      "theoretical_loss": 3.408959602803595,
      "tokens_seen": 2391736320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027801404212637914,
      "loss": 1.5254,
      "theoretical_loss": 3.4089521485866765,
      "tokens_seen": 2391801856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779939819458375,
      "loss": 1.2621,
      "theoretical_loss": 3.40894469463119,
      "tokens_seen": 2391867392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779739217652959,
      "loss": 1.3712,
      "theoretical_loss": 3.4089372409371195,
      "tokens_seen": 2391932928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779538615847543,
      "loss": 1.3463,
      "theoretical_loss": 3.408929787504449,
      "tokens_seen": 2391998464
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027793380140421264,
      "loss": 1.2377,
      "theoretical_loss": 3.408922334333161,
      "tokens_seen": 2392064000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277913741223671,
      "loss": 1.3519,
      "theoretical_loss": 3.4089148814232404,
      "tokens_seen": 2392129536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778936810431294,
      "loss": 1.3265,
      "theoretical_loss": 3.40890742877467,
      "tokens_seen": 2392195072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778736208625878,
      "loss": 1.4015,
      "theoretical_loss": 3.408899976387434,
      "tokens_seen": 2392260608
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027785356068204614,
      "loss": 1.3939,
      "theoretical_loss": 3.4088925242615162,
      "tokens_seen": 2392326144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778335005015045,
      "loss": 1.321,
      "theoretical_loss": 3.4088850723968998,
      "tokens_seen": 2392391680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027781344032096287,
      "loss": 1.3623,
      "theoretical_loss": 3.408877620793569,
      "tokens_seen": 2392457216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777933801404213,
      "loss": 1.3018,
      "theoretical_loss": 3.4088701694515064,
      "tokens_seen": 2392522752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027777331995987965,
      "loss": 1.3178,
      "theoretical_loss": 3.408862718370697,
      "tokens_seen": 2392588288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277753259779338,
      "loss": 1.2818,
      "theoretical_loss": 3.408855267551124,
      "tokens_seen": 2392653824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027773319959879637,
      "loss": 1.4012,
      "theoretical_loss": 3.408847816992771,
      "tokens_seen": 2392719360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777131394182548,
      "loss": 1.392,
      "theoretical_loss": 3.4088403666956215,
      "tokens_seen": 2392784896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027769307923771315,
      "loss": 1.3222,
      "theoretical_loss": 3.4088329166596596,
      "tokens_seen": 2392850432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776730190571715,
      "loss": 1.2379,
      "theoretical_loss": 3.4088254668848683,
      "tokens_seen": 2392915968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776529588766299,
      "loss": 1.2514,
      "theoretical_loss": 3.4088180173712326,
      "tokens_seen": 2392981504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776328986960883,
      "loss": 1.3376,
      "theoretical_loss": 3.4088105681187346,
      "tokens_seen": 2393047040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027761283851554665,
      "loss": 1.3743,
      "theoretical_loss": 3.4088031191273593,
      "tokens_seen": 2393112576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277592778335005,
      "loss": 1.2103,
      "theoretical_loss": 3.40879567039709,
      "tokens_seen": 2393178112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775727181544634,
      "loss": 1.3838,
      "theoretical_loss": 3.4087882219279098,
      "tokens_seen": 2393243648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027755265797392174,
      "loss": 1.4406,
      "theoretical_loss": 3.408780773719803,
      "tokens_seen": 2393309184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027753259779338016,
      "loss": 1.4041,
      "theoretical_loss": 3.408773325772753,
      "tokens_seen": 2393374720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775125376128385,
      "loss": 1.1663,
      "theoretical_loss": 3.408765878086744,
      "tokens_seen": 2393440256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774924774322969,
      "loss": 1.1972,
      "theoretical_loss": 3.4087584306617593,
      "tokens_seen": 2393505792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027747241725175524,
      "loss": 1.3388,
      "theoretical_loss": 3.4087509834977823,
      "tokens_seen": 2393571328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027745235707121366,
      "loss": 1.2026,
      "theoretical_loss": 3.4087435365947973,
      "tokens_seen": 2393636864
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277432296890672,
      "loss": 1.3068,
      "theoretical_loss": 3.408736089952788,
      "tokens_seen": 2393702400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774122367101304,
      "loss": 1.2507,
      "theoretical_loss": 3.4087286435717377,
      "tokens_seen": 2393767936
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 871588,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3815319538116455,
      "objective/train/theoretical_loss": 3.4087211974516305,
      "objective/train/tokens_used": 764292576,
      "theoretical_loss": 3.4087211974516305,
      "tokens_seen": 2393833472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027739217652958875,
      "loss": 1.2661,
      "theoretical_loss": 3.4087211974516305,
      "tokens_seen": 2393833472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773721163490471,
      "loss": 1.2545,
      "theoretical_loss": 3.4087137515924497,
      "tokens_seen": 2393899008
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773520561685055,
      "loss": 1.2263,
      "theoretical_loss": 3.4087063059941793,
      "tokens_seen": 2393964544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773319959879639,
      "loss": 1.3182,
      "theoretical_loss": 3.408698860656803,
      "tokens_seen": 2394030080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027731193580742225,
      "loss": 1.3328,
      "theoretical_loss": 3.4086914155803045,
      "tokens_seen": 2394095616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772918756268806,
      "loss": 1.421,
      "theoretical_loss": 3.4086839707646677,
      "tokens_seen": 2394161152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027727181544633903,
      "loss": 1.2766,
      "theoretical_loss": 3.408676526209876,
      "tokens_seen": 2394226688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772517552657974,
      "loss": 1.4351,
      "theoretical_loss": 3.408669081915913,
      "tokens_seen": 2394292224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027723169508525575,
      "loss": 1.2423,
      "theoretical_loss": 3.4086616378827626,
      "tokens_seen": 2394357760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772116349047141,
      "loss": 1.323,
      "theoretical_loss": 3.4086541941104085,
      "tokens_seen": 2394423296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771915747241725,
      "loss": 1.3549,
      "theoretical_loss": 3.4086467505988347,
      "tokens_seen": 2394488832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771715145436309,
      "loss": 1.2289,
      "theoretical_loss": 3.408639307348025,
      "tokens_seen": 2394554368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027715145436308926,
      "loss": 1.2752,
      "theoretical_loss": 3.4086318643579627,
      "tokens_seen": 2394619904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771313941825476,
      "loss": 1.2042,
      "theoretical_loss": 3.408624421628631,
      "tokens_seen": 2394685440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277111334002006,
      "loss": 1.2713,
      "theoretical_loss": 3.408616979160015,
      "tokens_seen": 2394750976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770912738214644,
      "loss": 1.206,
      "theoretical_loss": 3.408609536952098,
      "tokens_seen": 2394816512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027707121364092276,
      "loss": 1.3655,
      "theoretical_loss": 3.408602095004863,
      "tokens_seen": 2394882048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770511534603811,
      "loss": 1.3771,
      "theoretical_loss": 3.4085946533182945,
      "tokens_seen": 2394947584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770310932798395,
      "loss": 1.3121,
      "theoretical_loss": 3.4085872118923755,
      "tokens_seen": 2395013120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770110330992979,
      "loss": 1.3309,
      "theoretical_loss": 3.4085797707270906,
      "tokens_seen": 2395078656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769909729187563,
      "loss": 1.3682,
      "theoretical_loss": 3.408572329822423,
      "tokens_seen": 2395144192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769709127382147,
      "loss": 1.3539,
      "theoretical_loss": 3.408564889178357,
      "tokens_seen": 2395209728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027695085255767304,
      "loss": 1.3183,
      "theoretical_loss": 3.408557448794875,
      "tokens_seen": 2395275264
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769307923771314,
      "loss": 1.4011,
      "theoretical_loss": 3.4085500086719622,
      "tokens_seen": 2395340800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769107321965898,
      "loss": 1.2749,
      "theoretical_loss": 3.408542568809602,
      "tokens_seen": 2395406336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768906720160482,
      "loss": 1.3038,
      "theoretical_loss": 3.408535129207778,
      "tokens_seen": 2395471872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027687061183550655,
      "loss": 1.4534,
      "theoretical_loss": 3.4085276898664736,
      "tokens_seen": 2395537408
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768505516549649,
      "loss": 1.3297,
      "theoretical_loss": 3.4085202507856724,
      "tokens_seen": 2395602944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027683049147442327,
      "loss": 1.1877,
      "theoretical_loss": 3.4085128119653594,
      "tokens_seen": 2395668480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768104312938817,
      "loss": 1.258,
      "theoretical_loss": 3.408505373405517,
      "tokens_seen": 2395734016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027679037111334005,
      "loss": 1.2916,
      "theoretical_loss": 3.40849793510613,
      "tokens_seen": 2395799552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767703109327984,
      "loss": 1.3794,
      "theoretical_loss": 3.4084904970671817,
      "tokens_seen": 2395865088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027675025075225677,
      "loss": 1.3948,
      "theoretical_loss": 3.4084830592886552,
      "tokens_seen": 2395930624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767301905717152,
      "loss": 1.3623,
      "theoretical_loss": 3.4084756217705356,
      "tokens_seen": 2395996160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027671013039117355,
      "loss": 1.2729,
      "theoretical_loss": 3.4084681845128055,
      "tokens_seen": 2396061696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766900702106319,
      "loss": 1.2556,
      "theoretical_loss": 3.4084607475154494,
      "tokens_seen": 2396127232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766700100300903,
      "loss": 1.323,
      "theoretical_loss": 3.4084533107784507,
      "tokens_seen": 2396192768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766499498495487,
      "loss": 1.3776,
      "theoretical_loss": 3.408445874301793,
      "tokens_seen": 2396258304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027662988966900705,
      "loss": 1.3659,
      "theoretical_loss": 3.4084384380854607,
      "tokens_seen": 2396323840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766098294884654,
      "loss": 1.3252,
      "theoretical_loss": 3.408431002129437,
      "tokens_seen": 2396389376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765897693079238,
      "loss": 1.3118,
      "theoretical_loss": 3.4084235664337057,
      "tokens_seen": 2396454912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027656970912738214,
      "loss": 1.3606,
      "theoretical_loss": 3.4084161309982504,
      "tokens_seen": 2396520448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027654964894684056,
      "loss": 1.3327,
      "theoretical_loss": 3.4084086958230557,
      "tokens_seen": 2396585984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765295887662989,
      "loss": 1.2386,
      "theoretical_loss": 3.408401260908105,
      "tokens_seen": 2396651520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765095285857573,
      "loss": 1.3603,
      "theoretical_loss": 3.4083938262533815,
      "tokens_seen": 2396717056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027648946840521564,
      "loss": 1.3089,
      "theoretical_loss": 3.40838639185887,
      "tokens_seen": 2396782592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027646940822467406,
      "loss": 1.2454,
      "theoretical_loss": 3.408378957724553,
      "tokens_seen": 2396848128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764493480441324,
      "loss": 1.3112,
      "theoretical_loss": 3.4083715238504153,
      "tokens_seen": 2396913664
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764292878635908,
      "loss": 1.3484,
      "theoretical_loss": 3.4083640902364403,
      "tokens_seen": 2396979200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027640922768304915,
      "loss": 1.367,
      "theoretical_loss": 3.408356656882612,
      "tokens_seen": 2397044736
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 872754,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.360323190689087,
      "objective/train/theoretical_loss": 3.4083492237889135,
      "objective/train/tokens_used": 767569376,
      "theoretical_loss": 3.4083492237889135,
      "tokens_seen": 2397110272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763891675025075,
      "loss": 1.3887,
      "theoretical_loss": 3.4083492237889135,
      "tokens_seen": 2397110272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763691073219659,
      "loss": 1.3327,
      "theoretical_loss": 3.4083417909553297,
      "tokens_seen": 2397175808
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763490471414243,
      "loss": 1.2866,
      "theoretical_loss": 3.4083343583818433,
      "tokens_seen": 2397241344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027632898696088265,
      "loss": 1.4575,
      "theoretical_loss": 3.408326926068439,
      "tokens_seen": 2397306880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276308926780341,
      "loss": 1.2744,
      "theoretical_loss": 3.4083194940150996,
      "tokens_seen": 2397372416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027628886659979943,
      "loss": 1.3538,
      "theoretical_loss": 3.40831206222181,
      "tokens_seen": 2397437952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762688064192578,
      "loss": 1.3931,
      "theoretical_loss": 3.408304630688553,
      "tokens_seen": 2397503488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027624874623871615,
      "loss": 1.2701,
      "theoretical_loss": 3.4082971994153133,
      "tokens_seen": 2397569024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762286860581745,
      "loss": 1.2899,
      "theoretical_loss": 3.4082897684020743,
      "tokens_seen": 2397634560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762086258776329,
      "loss": 1.3303,
      "theoretical_loss": 3.4082823376488194,
      "tokens_seen": 2397700096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761885656970913,
      "loss": 1.331,
      "theoretical_loss": 3.4082749071555325,
      "tokens_seen": 2397765632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027616850551654966,
      "loss": 1.2312,
      "theoretical_loss": 3.408267476922198,
      "tokens_seen": 2397831168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276148445336008,
      "loss": 1.337,
      "theoretical_loss": 3.4082600469488,
      "tokens_seen": 2397896704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761283851554664,
      "loss": 1.3525,
      "theoretical_loss": 3.4082526172353207,
      "tokens_seen": 2397962240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761083249749248,
      "loss": 1.3213,
      "theoretical_loss": 3.408245187781745,
      "tokens_seen": 2398027776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027608826479438316,
      "loss": 1.3443,
      "theoretical_loss": 3.4082377585880566,
      "tokens_seen": 2398093312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002760682046138415,
      "loss": 1.3525,
      "theoretical_loss": 3.408230329654239,
      "tokens_seen": 2398158848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002760481444332999,
      "loss": 1.3702,
      "theoretical_loss": 3.4082229009802765,
      "tokens_seen": 2398224384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027602808425275825,
      "loss": 1.2997,
      "theoretical_loss": 3.4082154725661526,
      "tokens_seen": 2398289920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027600802407221666,
      "loss": 1.3106,
      "theoretical_loss": 3.4082080444118517,
      "tokens_seen": 2398355456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000275987963891675,
      "loss": 1.2634,
      "theoretical_loss": 3.4082006165173566,
      "tokens_seen": 2398420992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002759679037111334,
      "loss": 1.313,
      "theoretical_loss": 3.4081931888826515,
      "tokens_seen": 2398486528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027594784353059175,
      "loss": 1.438,
      "theoretical_loss": 3.4081857615077205,
      "tokens_seen": 2398552064
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027592778335005017,
      "loss": 1.3885,
      "theoretical_loss": 3.4081783343925474,
      "tokens_seen": 2398617600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027590772316950853,
      "loss": 1.3855,
      "theoretical_loss": 3.408170907537116,
      "tokens_seen": 2398683136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758876629889669,
      "loss": 1.3965,
      "theoretical_loss": 3.4081634809414094,
      "tokens_seen": 2398748672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027586760280842525,
      "loss": 1.3221,
      "theoretical_loss": 3.4081560546054126,
      "tokens_seen": 2398814208
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758475426278836,
      "loss": 1.3831,
      "theoretical_loss": 3.4081486285291085,
      "tokens_seen": 2398879744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027582748244734203,
      "loss": 1.2525,
      "theoretical_loss": 3.408141202712481,
      "tokens_seen": 2398945280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758074222668004,
      "loss": 1.2654,
      "theoretical_loss": 3.4081337771555145,
      "tokens_seen": 2399010816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027578736208625876,
      "loss": 1.3737,
      "theoretical_loss": 3.4081263518581926,
      "tokens_seen": 2399076352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757673019057171,
      "loss": 1.3749,
      "theoretical_loss": 3.408118926820499,
      "tokens_seen": 2399141888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027574724172517554,
      "loss": 1.2796,
      "theoretical_loss": 3.4081115020424173,
      "tokens_seen": 2399207424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757271815446339,
      "loss": 1.2403,
      "theoretical_loss": 3.408104077523932,
      "tokens_seen": 2399272960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027570712136409226,
      "loss": 1.402,
      "theoretical_loss": 3.408096653265026,
      "tokens_seen": 2399338496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756870611835506,
      "loss": 1.3959,
      "theoretical_loss": 3.408089229265684,
      "tokens_seen": 2399404032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000275667001003009,
      "loss": 1.2776,
      "theoretical_loss": 3.4080818055258897,
      "tokens_seen": 2399469568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756469408224674,
      "loss": 1.3649,
      "theoretical_loss": 3.4080743820456263,
      "tokens_seen": 2399535104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027562688064192576,
      "loss": 1.194,
      "theoretical_loss": 3.4080669588248784,
      "tokens_seen": 2399600640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756068204613841,
      "loss": 1.3388,
      "theoretical_loss": 3.4080595358636296,
      "tokens_seen": 2399666176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755867602808425,
      "loss": 1.1689,
      "theoretical_loss": 3.4080521131618635,
      "tokens_seen": 2399731712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755667001003009,
      "loss": 1.2937,
      "theoretical_loss": 3.4080446907195636,
      "tokens_seen": 2399797248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027554663991975927,
      "loss": 1.3176,
      "theoretical_loss": 3.408037268536715,
      "tokens_seen": 2399862784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027552657973921763,
      "loss": 1.2586,
      "theoretical_loss": 3.4080298466133003,
      "tokens_seen": 2399928320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027550651955867604,
      "loss": 1.3707,
      "theoretical_loss": 3.408022424949304,
      "tokens_seen": 2399993856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027548645937813446,
      "loss": 1.3295,
      "theoretical_loss": 3.40801500354471,
      "tokens_seen": 2400059392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754663991975928,
      "loss": 1.2894,
      "theoretical_loss": 3.408007582399502,
      "tokens_seen": 2400124928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754463390170512,
      "loss": 1.4401,
      "theoretical_loss": 3.4080001615136633,
      "tokens_seen": 2400190464
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027542627883650955,
      "loss": 1.3138,
      "theoretical_loss": 3.4079927408871784,
      "tokens_seen": 2400256000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754062186559679,
      "loss": 1.2888,
      "theoretical_loss": 3.407985320520031,
      "tokens_seen": 2400321536
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 873951,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4510786533355713,
      "objective/train/theoretical_loss": 3.407977900412205,
      "objective/train/tokens_used": 770846176,
      "theoretical_loss": 3.407977900412205,
      "tokens_seen": 2400387072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753861584754263,
      "loss": 1.4135,
      "theoretical_loss": 3.407977900412205,
      "tokens_seen": 2400387072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753660982948847,
      "loss": 1.2497,
      "theoretical_loss": 3.4079704805636846,
      "tokens_seen": 2400452608
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027534603811434305,
      "loss": 1.4125,
      "theoretical_loss": 3.4079630609744527,
      "tokens_seen": 2400518144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753259779338014,
      "loss": 1.46,
      "theoretical_loss": 3.4079556416444943,
      "tokens_seen": 2400583680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027530591775325983,
      "loss": 1.3364,
      "theoretical_loss": 3.407948222573792,
      "tokens_seen": 2400649216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002752858575727182,
      "loss": 1.297,
      "theoretical_loss": 3.407940803762331,
      "tokens_seen": 2400714752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027526579739217655,
      "loss": 1.2674,
      "theoretical_loss": 3.4079333852100944,
      "tokens_seen": 2400780288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752457372116349,
      "loss": 1.2989,
      "theoretical_loss": 3.407925966917066,
      "tokens_seen": 2400845824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752256770310933,
      "loss": 1.2649,
      "theoretical_loss": 3.40791854888323,
      "tokens_seen": 2400911360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752056168505517,
      "loss": 1.3714,
      "theoretical_loss": 3.4079111311085697,
      "tokens_seen": 2400976896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752056168505517,
      "loss": 1.392,
      "theoretical_loss": 3.4079037135930697,
      "tokens_seen": 2401042432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027518555667001006,
      "loss": 1.4094,
      "theoretical_loss": 3.4078962963367134,
      "tokens_seen": 2401107968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751654964894684,
      "loss": 1.1846,
      "theoretical_loss": 3.4078888793394855,
      "tokens_seen": 2401173504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751454363089268,
      "loss": 1.3237,
      "theoretical_loss": 3.4078814626013685,
      "tokens_seen": 2401239040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751253761283852,
      "loss": 1.3889,
      "theoretical_loss": 3.4078740461223473,
      "tokens_seen": 2401304576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027510531594784356,
      "loss": 1.4075,
      "theoretical_loss": 3.407866629902405,
      "tokens_seen": 2401370112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750852557673019,
      "loss": 1.4103,
      "theoretical_loss": 3.407859213941527,
      "tokens_seen": 2401435648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750651955867603,
      "loss": 1.2524,
      "theoretical_loss": 3.4078517982396956,
      "tokens_seen": 2401501184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027504513540621865,
      "loss": 1.2593,
      "theoretical_loss": 3.407844382796895,
      "tokens_seen": 2401566720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027502507522567706,
      "loss": 1.4134,
      "theoretical_loss": 3.4078369676131093,
      "tokens_seen": 2401632256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750050150451354,
      "loss": 1.3256,
      "theoretical_loss": 3.4078295526883227,
      "tokens_seen": 2401697792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002749849548645938,
      "loss": 1.3262,
      "theoretical_loss": 3.407822138022518,
      "tokens_seen": 2401763328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027496489468405215,
      "loss": 1.3199,
      "theoretical_loss": 3.407814723615681,
      "tokens_seen": 2401828864
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027494483450351057,
      "loss": 1.1812,
      "theoretical_loss": 3.4078073094677936,
      "tokens_seen": 2401894400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027492477432296893,
      "loss": 1.3434,
      "theoretical_loss": 3.407799895578841,
      "tokens_seen": 2401959936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002749047141424273,
      "loss": 1.2822,
      "theoretical_loss": 3.4077924819488064,
      "tokens_seen": 2402025472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027488465396188565,
      "loss": 1.298,
      "theoretical_loss": 3.4077850685776743,
      "tokens_seen": 2402091008
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274864593781344,
      "loss": 1.2224,
      "theoretical_loss": 3.407777655465428,
      "tokens_seen": 2402156544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027484453360080243,
      "loss": 1.4537,
      "theoretical_loss": 3.4077702426120515,
      "tokens_seen": 2402222080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002748244734202608,
      "loss": 1.4607,
      "theoretical_loss": 3.4077628300175284,
      "tokens_seen": 2402287616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027480441323971916,
      "loss": 1.3727,
      "theoretical_loss": 3.4077554176818436,
      "tokens_seen": 2402353152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747843530591775,
      "loss": 1.2401,
      "theoretical_loss": 3.4077480056049803,
      "tokens_seen": 2402418688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027476429287863594,
      "loss": 1.2116,
      "theoretical_loss": 3.4077405937869227,
      "tokens_seen": 2402484224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747442326980943,
      "loss": 1.3508,
      "theoretical_loss": 3.407733182227654,
      "tokens_seen": 2402549760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027472417251755266,
      "loss": 1.4278,
      "theoretical_loss": 3.407725770927159,
      "tokens_seen": 2402615296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274704112337011,
      "loss": 1.2559,
      "theoretical_loss": 3.4077183598854215,
      "tokens_seen": 2402680832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746840521564694,
      "loss": 1.389,
      "theoretical_loss": 3.407710949102425,
      "tokens_seen": 2402746368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746639919759278,
      "loss": 1.3656,
      "theoretical_loss": 3.4077035385781533,
      "tokens_seen": 2402811904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027464393179538616,
      "loss": 1.3845,
      "theoretical_loss": 3.4076961283125904,
      "tokens_seen": 2402877440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746238716148445,
      "loss": 1.2302,
      "theoretical_loss": 3.407688718305721,
      "tokens_seen": 2402942976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746038114343029,
      "loss": 1.4383,
      "theoretical_loss": 3.4076813085575277,
      "tokens_seen": 2403008512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745837512537613,
      "loss": 1.3165,
      "theoretical_loss": 3.4076738990679956,
      "tokens_seen": 2403074048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027456369107321967,
      "loss": 1.2525,
      "theoretical_loss": 3.4076664898371076,
      "tokens_seen": 2403139584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027454363089267803,
      "loss": 1.2551,
      "theoretical_loss": 3.407659080864849,
      "tokens_seen": 2403205120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745235707121364,
      "loss": 1.3652,
      "theoretical_loss": 3.4076516721512022,
      "tokens_seen": 2403270656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027450351053159475,
      "loss": 1.2601,
      "theoretical_loss": 3.407644263696152,
      "tokens_seen": 2403336192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027448345035105317,
      "loss": 1.2774,
      "theoretical_loss": 3.407636855499682,
      "tokens_seen": 2403401728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027446339017051153,
      "loss": 1.3591,
      "theoretical_loss": 3.407629447561776,
      "tokens_seen": 2403467264
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744433299899699,
      "loss": 1.3509,
      "theoretical_loss": 3.4076220398824186,
      "tokens_seen": 2403532800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027442326980942826,
      "loss": 1.3731,
      "theoretical_loss": 3.4076146324615935,
      "tokens_seen": 2403598336
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 875100,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3885475397109985,
      "objective/train/theoretical_loss": 3.4076072252992837,
      "objective/train/tokens_used": 774122976,
      "theoretical_loss": 3.4076072252992837,
      "tokens_seen": 2403663872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027440320962888667,
      "loss": 1.3443,
      "theoretical_loss": 3.4076072252992837,
      "tokens_seen": 2403663872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027438314944834503,
      "loss": 1.4259,
      "theoretical_loss": 3.4075998183954743,
      "tokens_seen": 2403729408
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743630892678034,
      "loss": 1.338,
      "theoretical_loss": 3.407592411750149,
      "tokens_seen": 2403794944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027434302908726176,
      "loss": 1.3654,
      "theoretical_loss": 3.407585005363291,
      "tokens_seen": 2403860480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743229689067202,
      "loss": 1.3199,
      "theoretical_loss": 3.4075775992348847,
      "tokens_seen": 2403926016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027430290872617854,
      "loss": 1.3459,
      "theoretical_loss": 3.407570193364914,
      "tokens_seen": 2403991552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742828485456369,
      "loss": 1.4096,
      "theoretical_loss": 3.4075627877533634,
      "tokens_seen": 2404057088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027426278836509526,
      "loss": 1.253,
      "theoretical_loss": 3.4075553824002163,
      "tokens_seen": 2404122624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742427281845536,
      "loss": 1.4278,
      "theoretical_loss": 3.407547977305456,
      "tokens_seen": 2404188160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027422266800401204,
      "loss": 1.4905,
      "theoretical_loss": 3.407540572469068,
      "tokens_seen": 2404253696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742026078234704,
      "loss": 1.3587,
      "theoretical_loss": 3.407533167891035,
      "tokens_seen": 2404319232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027418254764292877,
      "loss": 1.3744,
      "theoretical_loss": 3.4075257635713414,
      "tokens_seen": 2404384768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027416248746238713,
      "loss": 1.5,
      "theoretical_loss": 3.407518359509971,
      "tokens_seen": 2404450304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027414242728184554,
      "loss": 1.2847,
      "theoretical_loss": 3.4075109557069077,
      "tokens_seen": 2404515840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741223671013039,
      "loss": 1.3629,
      "theoretical_loss": 3.4075035521621357,
      "tokens_seen": 2404581376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027410230692076227,
      "loss": 1.1964,
      "theoretical_loss": 3.407496148875639,
      "tokens_seen": 2404646912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027408224674022063,
      "loss": 1.2677,
      "theoretical_loss": 3.407488745847401,
      "tokens_seen": 2404712448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274062186559679,
      "loss": 1.3067,
      "theoretical_loss": 3.4074813430774062,
      "tokens_seen": 2404777984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002740421263791374,
      "loss": 1.3459,
      "theoretical_loss": 3.4074739405656382,
      "tokens_seen": 2404843520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002740220661985958,
      "loss": 1.399,
      "theoretical_loss": 3.4074665383120815,
      "tokens_seen": 2404909056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002740020060180542,
      "loss": 1.3337,
      "theoretical_loss": 3.407459136316719,
      "tokens_seen": 2404974592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027398194583751255,
      "loss": 1.306,
      "theoretical_loss": 3.407451734579536,
      "tokens_seen": 2405040128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027396188565697097,
      "loss": 1.3503,
      "theoretical_loss": 3.4074443331005155,
      "tokens_seen": 2405105664
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027394182547642933,
      "loss": 1.5323,
      "theoretical_loss": 3.407436931879642,
      "tokens_seen": 2405171200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002739217652958877,
      "loss": 1.3805,
      "theoretical_loss": 3.407429530916899,
      "tokens_seen": 2405236736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027390170511534605,
      "loss": 1.3784,
      "theoretical_loss": 3.4074221302122707,
      "tokens_seen": 2405302272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738816449348044,
      "loss": 1.5019,
      "theoretical_loss": 3.407414729765741,
      "tokens_seen": 2405367808
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027386158475426283,
      "loss": 1.2801,
      "theoretical_loss": 3.407407329577294,
      "tokens_seen": 2405433344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738415245737212,
      "loss": 1.3146,
      "theoretical_loss": 3.4073999296469135,
      "tokens_seen": 2405498880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027382146439317956,
      "loss": 1.3977,
      "theoretical_loss": 3.407392529974584,
      "tokens_seen": 2405564416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738014042126379,
      "loss": 1.2694,
      "theoretical_loss": 3.4073851305602885,
      "tokens_seen": 2405629952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027378134403209634,
      "loss": 1.386,
      "theoretical_loss": 3.4073777314040115,
      "tokens_seen": 2405695488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737612838515547,
      "loss": 1.306,
      "theoretical_loss": 3.407370332505737,
      "tokens_seen": 2405761024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027374122367101306,
      "loss": 1.2434,
      "theoretical_loss": 3.4073629338654494,
      "tokens_seen": 2405826560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737211634904714,
      "loss": 1.3107,
      "theoretical_loss": 3.407355535483132,
      "tokens_seen": 2405892096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737011033099298,
      "loss": 1.4423,
      "theoretical_loss": 3.4073481373587686,
      "tokens_seen": 2405957632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736810431293882,
      "loss": 1.4299,
      "theoretical_loss": 3.4073407394923443,
      "tokens_seen": 2406023168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027366098294884656,
      "loss": 1.3723,
      "theoretical_loss": 3.4073333418838416,
      "tokens_seen": 2406088704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736409227683049,
      "loss": 1.3525,
      "theoretical_loss": 3.407325944533246,
      "tokens_seen": 2406154240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736208625877633,
      "loss": 1.2994,
      "theoretical_loss": 3.40731854744054,
      "tokens_seen": 2406219776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736008024072217,
      "loss": 1.3442,
      "theoretical_loss": 3.407311150605709,
      "tokens_seen": 2406285312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027358074222668007,
      "loss": 1.3116,
      "theoretical_loss": 3.407303754028736,
      "tokens_seen": 2406350848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027356068204613843,
      "loss": 1.4537,
      "theoretical_loss": 3.4072963577096056,
      "tokens_seen": 2406416384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002735406218655968,
      "loss": 1.4134,
      "theoretical_loss": 3.407288961648301,
      "tokens_seen": 2406481920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027352056168505515,
      "loss": 1.398,
      "theoretical_loss": 3.407281565844807,
      "tokens_seen": 2406547456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027350050150451357,
      "loss": 1.3026,
      "theoretical_loss": 3.407274170299107,
      "tokens_seen": 2406612992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027348044132397193,
      "loss": 1.3736,
      "theoretical_loss": 3.4072667750111854,
      "tokens_seen": 2406678528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734603811434303,
      "loss": 1.4851,
      "theoretical_loss": 3.407259379981026,
      "tokens_seen": 2406744064
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027344032096288866,
      "loss": 1.2229,
      "theoretical_loss": 3.407251985208613,
      "tokens_seen": 2406809600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734202607823471,
      "loss": 1.3228,
      "theoretical_loss": 3.4072445906939306,
      "tokens_seen": 2406875136
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 876268,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4658091068267822,
      "objective/train/theoretical_loss": 3.407237196436962,
      "objective/train/tokens_used": 777399776,
      "theoretical_loss": 3.407237196436962,
      "tokens_seen": 2406940672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027340020060180544,
      "loss": 1.4443,
      "theoretical_loss": 3.407237196436962,
      "tokens_seen": 2406940672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002733801404212638,
      "loss": 1.2755,
      "theoretical_loss": 3.4072298024376915,
      "tokens_seen": 2407006208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027336008024072216,
      "loss": 1.2804,
      "theoretical_loss": 3.4072224086961036,
      "tokens_seen": 2407071744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002733400200601806,
      "loss": 1.3005,
      "theoretical_loss": 3.4072150152121816,
      "tokens_seen": 2407137280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027331995987963894,
      "loss": 1.3541,
      "theoretical_loss": 3.4072076219859104,
      "tokens_seen": 2407202816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732998996990973,
      "loss": 1.5378,
      "theoretical_loss": 3.407200229017273,
      "tokens_seen": 2407268352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027327983951855566,
      "loss": 1.366,
      "theoretical_loss": 3.4071928363062542,
      "tokens_seen": 2407333888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000273259779338014,
      "loss": 1.3577,
      "theoretical_loss": 3.4071854438528377,
      "tokens_seen": 2407399424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027323971915747244,
      "loss": 1.3391,
      "theoretical_loss": 3.407178051657007,
      "tokens_seen": 2407464960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732196589769308,
      "loss": 1.3225,
      "theoretical_loss": 3.4071706597187474,
      "tokens_seen": 2407530496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027319959879638917,
      "loss": 1.2143,
      "theoretical_loss": 3.4071632680380417,
      "tokens_seen": 2407596032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027317953861584753,
      "loss": 1.3921,
      "theoretical_loss": 3.4071558766148744,
      "tokens_seen": 2407661568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027315947843530594,
      "loss": 1.272,
      "theoretical_loss": 3.4071484854492295,
      "tokens_seen": 2407727104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731394182547643,
      "loss": 1.3395,
      "theoretical_loss": 3.407141094541091,
      "tokens_seen": 2407792640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027311935807422267,
      "loss": 1.3014,
      "theoretical_loss": 3.407133703890443,
      "tokens_seen": 2407858176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027309929789368103,
      "loss": 1.3823,
      "theoretical_loss": 3.4071263134972694,
      "tokens_seen": 2407923712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002730792377131394,
      "loss": 1.3122,
      "theoretical_loss": 3.407118923361554,
      "tokens_seen": 2407989248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002730591775325978,
      "loss": 1.3595,
      "theoretical_loss": 3.407111533483281,
      "tokens_seen": 2408054784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027303911735205617,
      "loss": 1.2864,
      "theoretical_loss": 3.407104143862435,
      "tokens_seen": 2408120320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027301905717151453,
      "loss": 1.2997,
      "theoretical_loss": 3.4070967544989994,
      "tokens_seen": 2408185856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729989969909729,
      "loss": 1.4726,
      "theoretical_loss": 3.407089365392958,
      "tokens_seen": 2408251392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729789368104313,
      "loss": 1.3011,
      "theoretical_loss": 3.4070819765442955,
      "tokens_seen": 2408316928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729588766298897,
      "loss": 1.3438,
      "theoretical_loss": 3.407074587952996,
      "tokens_seen": 2408382464
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027293881644934804,
      "loss": 1.2962,
      "theoretical_loss": 3.4070671996190423,
      "tokens_seen": 2408448000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729187562688064,
      "loss": 1.4236,
      "theoretical_loss": 3.4070598115424198,
      "tokens_seen": 2408513536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027289869608826476,
      "loss": 1.3333,
      "theoretical_loss": 3.407052423723112,
      "tokens_seen": 2408579072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728786359077232,
      "loss": 1.2338,
      "theoretical_loss": 3.4070450361611027,
      "tokens_seen": 2408644608
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027285857572718154,
      "loss": 1.2131,
      "theoretical_loss": 3.4070376488563765,
      "tokens_seen": 2408710144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728385155466399,
      "loss": 1.4353,
      "theoretical_loss": 3.407030261808917,
      "tokens_seen": 2408775680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027281845536609827,
      "loss": 1.2525,
      "theoretical_loss": 3.407022875018708,
      "tokens_seen": 2408841216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002727983951855567,
      "loss": 1.2966,
      "theoretical_loss": 3.407015488485735,
      "tokens_seen": 2408906752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027277833500501504,
      "loss": 1.2344,
      "theoretical_loss": 3.40700810220998,
      "tokens_seen": 2408972288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002727582748244734,
      "loss": 1.3516,
      "theoretical_loss": 3.407000716191428,
      "tokens_seen": 2409037824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027273821464393177,
      "loss": 1.2706,
      "theoretical_loss": 3.4069933304300637,
      "tokens_seen": 2409103360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027271815446339013,
      "loss": 1.276,
      "theoretical_loss": 3.40698594492587,
      "tokens_seen": 2409168896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027269809428284855,
      "loss": 1.3287,
      "theoretical_loss": 3.406978559678832,
      "tokens_seen": 2409234432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002726780341023069,
      "loss": 1.3735,
      "theoretical_loss": 3.4069711746889326,
      "tokens_seen": 2409299968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027265797392176527,
      "loss": 1.3452,
      "theoretical_loss": 3.4069637899561567,
      "tokens_seen": 2409365504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027263791374122363,
      "loss": 1.4255,
      "theoretical_loss": 3.406956405480488,
      "tokens_seen": 2409431040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027261785356068205,
      "loss": 1.2694,
      "theoretical_loss": 3.406949021261911,
      "tokens_seen": 2409496576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725977933801404,
      "loss": 1.4514,
      "theoretical_loss": 3.406941637300409,
      "tokens_seen": 2409562112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725777331995988,
      "loss": 1.4954,
      "theoretical_loss": 3.406934253595967,
      "tokens_seen": 2409627648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027255767301905714,
      "loss": 1.4562,
      "theoretical_loss": 3.406926870148568,
      "tokens_seen": 2409693184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725376128385155,
      "loss": 1.479,
      "theoretical_loss": 3.4069194869581967,
      "tokens_seen": 2409758720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027251755265797397,
      "loss": 1.2468,
      "theoretical_loss": 3.4069121040248374,
      "tokens_seen": 2409824256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027249749247743233,
      "loss": 1.27,
      "theoretical_loss": 3.4069047213484738,
      "tokens_seen": 2409889792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724774322968907,
      "loss": 1.3293,
      "theoretical_loss": 3.40689733892909,
      "tokens_seen": 2409955328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027245737211634906,
      "loss": 1.3974,
      "theoretical_loss": 3.40688995676667,
      "tokens_seen": 2410020864
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724373119358075,
      "loss": 1.4021,
      "theoretical_loss": 3.4068825748611977,
      "tokens_seen": 2410086400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027241725175526584,
      "loss": 1.2945,
      "theoretical_loss": 3.4068751932126577,
      "tokens_seen": 2410151936
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 877499,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2382419109344482,
      "objective/train/theoretical_loss": 3.4068678118210336,
      "objective/train/tokens_used": 780676576,
      "theoretical_loss": 3.4068678118210336,
      "tokens_seen": 2410217472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723971915747242,
      "loss": 1.3304,
      "theoretical_loss": 3.4068678118210336,
      "tokens_seen": 2410217472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027237713139418256,
      "loss": 1.3771,
      "theoretical_loss": 3.4068604306863097,
      "tokens_seen": 2410283008
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000272357071213641,
      "loss": 1.3283,
      "theoretical_loss": 3.40685304980847,
      "tokens_seen": 2410348544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027233701103309934,
      "loss": 1.4518,
      "theoretical_loss": 3.406845669187499,
      "tokens_seen": 2410414080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723169508525577,
      "loss": 1.2622,
      "theoretical_loss": 3.40683828882338,
      "tokens_seen": 2410479616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027229689067201606,
      "loss": 1.376,
      "theoretical_loss": 3.4068309087160973,
      "tokens_seen": 2410545152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722768304914744,
      "loss": 1.2939,
      "theoretical_loss": 3.406823528865635,
      "tokens_seen": 2410610688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027225677031093284,
      "loss": 1.3564,
      "theoretical_loss": 3.406816149271978,
      "tokens_seen": 2410676224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722367101303912,
      "loss": 1.3474,
      "theoretical_loss": 3.4068087699351093,
      "tokens_seen": 2410741760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027221664994984957,
      "loss": 1.3195,
      "theoretical_loss": 3.406801390855013,
      "tokens_seen": 2410807296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027219658976930793,
      "loss": 1.3918,
      "theoretical_loss": 3.406794012031674,
      "tokens_seen": 2410872832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027217652958876635,
      "loss": 1.347,
      "theoretical_loss": 3.4067866334650763,
      "tokens_seen": 2410938368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721564694082247,
      "loss": 1.3931,
      "theoretical_loss": 3.406779255155203,
      "tokens_seen": 2411003904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027213640922768307,
      "loss": 1.3192,
      "theoretical_loss": 3.406771877102039,
      "tokens_seen": 2411069440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027211634904714143,
      "loss": 1.3234,
      "theoretical_loss": 3.406764499305569,
      "tokens_seen": 2411134976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720962888665998,
      "loss": 1.2957,
      "theoretical_loss": 3.4067571217657755,
      "tokens_seen": 2411200512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720762286860582,
      "loss": 1.3583,
      "theoretical_loss": 3.4067497444826436,
      "tokens_seen": 2411266048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027205616850551657,
      "loss": 1.3308,
      "theoretical_loss": 3.406742367456157,
      "tokens_seen": 2411331584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027203610832497493,
      "loss": 1.3276,
      "theoretical_loss": 3.4067349906863003,
      "tokens_seen": 2411397120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720160481444333,
      "loss": 1.3166,
      "theoretical_loss": 3.406727614173057,
      "tokens_seen": 2411462656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719959879638917,
      "loss": 1.3315,
      "theoretical_loss": 3.406720237916412,
      "tokens_seen": 2411528192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719759277833501,
      "loss": 1.5321,
      "theoretical_loss": 3.4067128619163487,
      "tokens_seen": 2411593728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027195586760280844,
      "loss": 1.3522,
      "theoretical_loss": 3.406705486172851,
      "tokens_seen": 2411659264
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719358074222668,
      "loss": 1.3515,
      "theoretical_loss": 3.4066981106859036,
      "tokens_seen": 2411724800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027191574724172516,
      "loss": 1.1903,
      "theoretical_loss": 3.4066907354554905,
      "tokens_seen": 2411790336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718956870611836,
      "loss": 1.3984,
      "theoretical_loss": 3.406683360481596,
      "tokens_seen": 2411855872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027187562688064194,
      "loss": 1.2606,
      "theoretical_loss": 3.4066759857642035,
      "tokens_seen": 2411921408
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718555667001003,
      "loss": 1.3034,
      "theoretical_loss": 3.4066686113032976,
      "tokens_seen": 2411986944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027183550651955867,
      "loss": 1.4012,
      "theoretical_loss": 3.4066612370988625,
      "tokens_seen": 2412052480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718154463390171,
      "loss": 1.3774,
      "theoretical_loss": 3.4066538631508823,
      "tokens_seen": 2412118016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027179538615847544,
      "loss": 1.3222,
      "theoretical_loss": 3.406646489459341,
      "tokens_seen": 2412183552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002717753259779338,
      "loss": 1.3702,
      "theoretical_loss": 3.4066391160242224,
      "tokens_seen": 2412249088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027175526579739217,
      "loss": 1.2503,
      "theoretical_loss": 3.406631742845511,
      "tokens_seen": 2412314624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027173520561685053,
      "loss": 1.3344,
      "theoretical_loss": 3.4066243699231906,
      "tokens_seen": 2412380160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027171514543630895,
      "loss": 1.3219,
      "theoretical_loss": 3.4066169972572458,
      "tokens_seen": 2412445696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716950852557673,
      "loss": 1.3941,
      "theoretical_loss": 3.4066096248476607,
      "tokens_seen": 2412511232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027167502507522567,
      "loss": 1.3243,
      "theoretical_loss": 3.4066022526944186,
      "tokens_seen": 2412576768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027165496489468403,
      "loss": 1.5133,
      "theoretical_loss": 3.406594880797505,
      "tokens_seen": 2412642304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027163490471414245,
      "loss": 1.3319,
      "theoretical_loss": 3.4065875091569024,
      "tokens_seen": 2412707840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716148445336008,
      "loss": 1.306,
      "theoretical_loss": 3.4065801377725964,
      "tokens_seen": 2412773376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715947843530592,
      "loss": 1.4057,
      "theoretical_loss": 3.40657276664457,
      "tokens_seen": 2412838912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027157472417251754,
      "loss": 1.286,
      "theoretical_loss": 3.4065653957728084,
      "tokens_seen": 2412904448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715546639919759,
      "loss": 1.3028,
      "theoretical_loss": 3.4065580251572944,
      "tokens_seen": 2412969984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715346038114343,
      "loss": 1.4019,
      "theoretical_loss": 3.4065506547980138,
      "tokens_seen": 2413035520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715145436308927,
      "loss": 1.3433,
      "theoretical_loss": 3.406543284694949,
      "tokens_seen": 2413101056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027149448345035104,
      "loss": 1.3145,
      "theoretical_loss": 3.406535914848085,
      "tokens_seen": 2413166592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002714744232698094,
      "loss": 1.2907,
      "theoretical_loss": 3.4065285452574066,
      "tokens_seen": 2413232128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002714543630892678,
      "loss": 1.4535,
      "theoretical_loss": 3.406521175922897,
      "tokens_seen": 2413297664
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002714343029087262,
      "loss": 1.2803,
      "theoretical_loss": 3.40651380684454,
      "tokens_seen": 2413363200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027141424272818454,
      "loss": 1.4042,
      "theoretical_loss": 3.406506438022321,
      "tokens_seen": 2413428736
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 878709,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2260515689849854,
      "objective/train/theoretical_loss": 3.406499069456223,
      "objective/train/tokens_used": 783953376,
      "theoretical_loss": 3.406499069456223,
      "tokens_seen": 2413494272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713941825476429,
      "loss": 1.2944,
      "theoretical_loss": 3.406499069456223,
      "tokens_seen": 2413494272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027137412236710127,
      "loss": 1.2526,
      "theoretical_loss": 3.406491701146231,
      "tokens_seen": 2413559808
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713540621865597,
      "loss": 1.2666,
      "theoretical_loss": 3.4064843330923282,
      "tokens_seen": 2413625344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027133400200601805,
      "loss": 1.3362,
      "theoretical_loss": 3.4064769652944995,
      "tokens_seen": 2413690880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713139418254764,
      "loss": 1.3257,
      "theoretical_loss": 3.4064695977527286,
      "tokens_seen": 2413756416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027129388164493477,
      "loss": 1.3463,
      "theoretical_loss": 3.4064622304670005,
      "tokens_seen": 2413821952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712738214643932,
      "loss": 1.313,
      "theoretical_loss": 3.406454863437298,
      "tokens_seen": 2413887488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027125376128385155,
      "loss": 1.3596,
      "theoretical_loss": 3.4064474966636067,
      "tokens_seen": 2413953024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712337011033099,
      "loss": 1.3633,
      "theoretical_loss": 3.4064401301459095,
      "tokens_seen": 2414018560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712136409227683,
      "loss": 1.3177,
      "theoretical_loss": 3.4064327638841916,
      "tokens_seen": 2414084096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711935807422267,
      "loss": 1.4225,
      "theoretical_loss": 3.4064253978784365,
      "tokens_seen": 2414149632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027117352056168505,
      "loss": 1.4033,
      "theoretical_loss": 3.406418032128628,
      "tokens_seen": 2414215168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711534603811434,
      "loss": 1.3188,
      "theoretical_loss": 3.406410666634751,
      "tokens_seen": 2414280704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711334002006018,
      "loss": 1.3211,
      "theoretical_loss": 3.40640330139679,
      "tokens_seen": 2414346240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027111334002006014,
      "loss": 1.3446,
      "theoretical_loss": 3.406395936414728,
      "tokens_seen": 2414411776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027109327983951856,
      "loss": 1.3498,
      "theoretical_loss": 3.4063885716885496,
      "tokens_seen": 2414477312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710732196589769,
      "loss": 1.3331,
      "theoretical_loss": 3.4063812072182396,
      "tokens_seen": 2414542848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710531594784353,
      "loss": 1.3824,
      "theoretical_loss": 3.4063738430037813,
      "tokens_seen": 2414608384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027103309929789364,
      "loss": 1.3836,
      "theoretical_loss": 3.4063664790451598,
      "tokens_seen": 2414673920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710130391173521,
      "loss": 1.3344,
      "theoretical_loss": 3.406359115342358,
      "tokens_seen": 2414739456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709929789368105,
      "loss": 1.3293,
      "theoretical_loss": 3.4063517518953614,
      "tokens_seen": 2414804992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027097291875626884,
      "loss": 1.3345,
      "theoretical_loss": 3.406344388704153,
      "tokens_seen": 2414870528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709528585757272,
      "loss": 1.2308,
      "theoretical_loss": 3.406337025768718,
      "tokens_seen": 2414936064
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027093279839518556,
      "loss": 1.3315,
      "theoretical_loss": 3.40632966308904,
      "tokens_seen": 2415001600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000270912738214644,
      "loss": 1.3957,
      "theoretical_loss": 3.4063223006651033,
      "tokens_seen": 2415067136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027089267803410234,
      "loss": 1.3345,
      "theoretical_loss": 3.406314938496892,
      "tokens_seen": 2415132672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708726178535607,
      "loss": 1.2748,
      "theoretical_loss": 3.4063075765843904,
      "tokens_seen": 2415198208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027085255767301907,
      "loss": 1.3211,
      "theoretical_loss": 3.4063002149275823,
      "tokens_seen": 2415263744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708324974924775,
      "loss": 1.2757,
      "theoretical_loss": 3.4062928535264527,
      "tokens_seen": 2415329280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027081243731193584,
      "loss": 1.3569,
      "theoretical_loss": 3.406285492380985,
      "tokens_seen": 2415394816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707923771313942,
      "loss": 1.4019,
      "theoretical_loss": 3.4062781314911637,
      "tokens_seen": 2415460352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027077231695085257,
      "loss": 1.2499,
      "theoretical_loss": 3.406270770856973,
      "tokens_seen": 2415525888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027075225677031093,
      "loss": 1.3974,
      "theoretical_loss": 3.406263410478397,
      "tokens_seen": 2415591424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027073219658976935,
      "loss": 1.3,
      "theoretical_loss": 3.40625605035542,
      "tokens_seen": 2415656960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707121364092277,
      "loss": 1.4814,
      "theoretical_loss": 3.406248690488026,
      "tokens_seen": 2415722496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027069207622868607,
      "loss": 1.2349,
      "theoretical_loss": 3.4062413308761994,
      "tokens_seen": 2415788032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027067201604814443,
      "loss": 1.3823,
      "theoretical_loss": 3.4062339715199244,
      "tokens_seen": 2415853568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027065195586760285,
      "loss": 1.324,
      "theoretical_loss": 3.406226612419185,
      "tokens_seen": 2415919104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002706318956870612,
      "loss": 1.4081,
      "theoretical_loss": 3.4062192535739655,
      "tokens_seen": 2415984640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002706118355065196,
      "loss": 1.3212,
      "theoretical_loss": 3.4062118949842497,
      "tokens_seen": 2416050176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027059177532597794,
      "loss": 1.291,
      "theoretical_loss": 3.406204536650023,
      "tokens_seen": 2416115712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705717151454363,
      "loss": 1.3856,
      "theoretical_loss": 3.406197178571268,
      "tokens_seen": 2416181248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705516549648947,
      "loss": 1.4,
      "theoretical_loss": 3.40618982074797,
      "tokens_seen": 2416246784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705315947843531,
      "loss": 1.3138,
      "theoretical_loss": 3.406182463180113,
      "tokens_seen": 2416312320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027051153460381144,
      "loss": 1.2567,
      "theoretical_loss": 3.406175105867681,
      "tokens_seen": 2416377856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704914744232698,
      "loss": 1.3544,
      "theoretical_loss": 3.4061677488106583,
      "tokens_seen": 2416443392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704714142427282,
      "loss": 1.437,
      "theoretical_loss": 3.406160392009029,
      "tokens_seen": 2416508928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704513540621866,
      "loss": 1.3461,
      "theoretical_loss": 3.406153035462778,
      "tokens_seen": 2416574464
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027043129388164494,
      "loss": 1.3394,
      "theoretical_loss": 3.406145679171888,
      "tokens_seen": 2416640000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704112337011033,
      "loss": 1.3404,
      "theoretical_loss": 3.4061383231363447,
      "tokens_seen": 2416705536
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 880009,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3705439567565918,
      "objective/train/theoretical_loss": 3.4061309673561317,
      "objective/train/tokens_used": 787230176,
      "theoretical_loss": 3.4061309673561317,
      "tokens_seen": 2416771072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027039117352056167,
      "loss": 1.3516,
      "theoretical_loss": 3.4061309673561317,
      "tokens_seen": 2416771072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703711133400201,
      "loss": 1.4004,
      "theoretical_loss": 3.4061236118312332,
      "tokens_seen": 2416836608
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027035105315947845,
      "loss": 1.4456,
      "theoretical_loss": 3.4061162565616336,
      "tokens_seen": 2416902144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703309929789368,
      "loss": 1.3281,
      "theoretical_loss": 3.406108901547317,
      "tokens_seen": 2416967680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027031093279839517,
      "loss": 1.3332,
      "theoretical_loss": 3.4061015467882676,
      "tokens_seen": 2417033216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002702908726178536,
      "loss": 1.2566,
      "theoretical_loss": 3.4060941922844696,
      "tokens_seen": 2417098752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027027081243731195,
      "loss": 1.4096,
      "theoretical_loss": 3.406086838035907,
      "tokens_seen": 2417164288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002702507522567703,
      "loss": 1.4685,
      "theoretical_loss": 3.4060794840425648,
      "tokens_seen": 2417229824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002702306920762287,
      "loss": 1.5282,
      "theoretical_loss": 3.4060721303044263,
      "tokens_seen": 2417295360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002702106318956871,
      "loss": 1.3421,
      "theoretical_loss": 3.4060647768214762,
      "tokens_seen": 2417360896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027019057171514545,
      "loss": 1.3677,
      "theoretical_loss": 3.406057423593699,
      "tokens_seen": 2417426432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701705115346038,
      "loss": 1.3555,
      "theoretical_loss": 3.4060500706210783,
      "tokens_seen": 2417491968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701504513540622,
      "loss": 1.4035,
      "theoretical_loss": 3.4060427179035986,
      "tokens_seen": 2417557504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027013039117352054,
      "loss": 1.2473,
      "theoretical_loss": 3.406035365441244,
      "tokens_seen": 2417623040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027011033099297896,
      "loss": 1.3873,
      "theoretical_loss": 3.406028013233999,
      "tokens_seen": 2417688576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700902708124373,
      "loss": 1.3009,
      "theoretical_loss": 3.4060206612818478,
      "tokens_seen": 2417754112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700702106318957,
      "loss": 1.3812,
      "theoretical_loss": 3.4060133095847744,
      "tokens_seen": 2417819648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027005015045135404,
      "loss": 1.2852,
      "theoretical_loss": 3.4060059581427633,
      "tokens_seen": 2417885184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027003009027081246,
      "loss": 1.2703,
      "theoretical_loss": 3.4059986069557984,
      "tokens_seen": 2417950720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700100300902708,
      "loss": 1.3213,
      "theoretical_loss": 3.4059912560238645,
      "tokens_seen": 2418016256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699899699097292,
      "loss": 1.2318,
      "theoretical_loss": 3.4059839053469454,
      "tokens_seen": 2418081792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026996990972918755,
      "loss": 1.3667,
      "theoretical_loss": 3.4059765549250254,
      "tokens_seen": 2418147328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699498495486459,
      "loss": 1.2922,
      "theoretical_loss": 3.4059692047580885,
      "tokens_seen": 2418212864
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699297893681043,
      "loss": 1.354,
      "theoretical_loss": 3.4059618548461197,
      "tokens_seen": 2418278400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699097291875627,
      "loss": 1.3234,
      "theoretical_loss": 3.4059545051891025,
      "tokens_seen": 2418343936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026988966900702105,
      "loss": 1.2766,
      "theoretical_loss": 3.4059471557870213,
      "tokens_seen": 2418409472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698696088264794,
      "loss": 1.2844,
      "theoretical_loss": 3.4059398066398607,
      "tokens_seen": 2418475008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026984954864593783,
      "loss": 1.3467,
      "theoretical_loss": 3.4059324577476047,
      "tokens_seen": 2418540544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698294884653962,
      "loss": 1.2876,
      "theoretical_loss": 3.4059251091102376,
      "tokens_seen": 2418606080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026980942828485455,
      "loss": 1.3701,
      "theoretical_loss": 3.4059177607277435,
      "tokens_seen": 2418671616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697893681043129,
      "loss": 1.4601,
      "theoretical_loss": 3.405910412600107,
      "tokens_seen": 2418737152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697693079237713,
      "loss": 1.3751,
      "theoretical_loss": 3.405903064727312,
      "tokens_seen": 2418802688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697492477432297,
      "loss": 1.3724,
      "theoretical_loss": 3.4058957171093427,
      "tokens_seen": 2418868224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026972918756268806,
      "loss": 1.2093,
      "theoretical_loss": 3.405888369746184,
      "tokens_seen": 2418933760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697091273821464,
      "loss": 1.359,
      "theoretical_loss": 3.405881022637819,
      "tokens_seen": 2418999296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696890672016048,
      "loss": 1.3386,
      "theoretical_loss": 3.405873675784233,
      "tokens_seen": 2419064832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696690070210632,
      "loss": 1.3968,
      "theoretical_loss": 3.4058663291854105,
      "tokens_seen": 2419130368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026964894684052156,
      "loss": 1.3508,
      "theoretical_loss": 3.405858982841335,
      "tokens_seen": 2419195904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696288866599799,
      "loss": 1.3091,
      "theoretical_loss": 3.4058516367519904,
      "tokens_seen": 2419261440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696088264794383,
      "loss": 1.2846,
      "theoretical_loss": 3.405844290917362,
      "tokens_seen": 2419326976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026958876629889665,
      "loss": 1.3162,
      "theoretical_loss": 3.405836945337433,
      "tokens_seen": 2419392512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026956870611835506,
      "loss": 1.3486,
      "theoretical_loss": 3.405829600012189,
      "tokens_seen": 2419458048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695486459378134,
      "loss": 1.3165,
      "theoretical_loss": 3.4058222549416133,
      "tokens_seen": 2419523584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695285857572718,
      "loss": 1.3118,
      "theoretical_loss": 3.4058149101256903,
      "tokens_seen": 2419589120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695085255767302,
      "loss": 1.3184,
      "theoretical_loss": 3.4058075655644044,
      "tokens_seen": 2419654656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694884653961886,
      "loss": 1.3961,
      "theoretical_loss": 3.4058002212577403,
      "tokens_seen": 2419720192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269468405215647,
      "loss": 1.3534,
      "theoretical_loss": 3.4057928772056814,
      "tokens_seen": 2419785728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026944834503510534,
      "loss": 1.399,
      "theoretical_loss": 3.4057855334082126,
      "tokens_seen": 2419851264
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694282848545637,
      "loss": 1.2392,
      "theoretical_loss": 3.405778189865318,
      "tokens_seen": 2419916800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026940822467402207,
      "loss": 1.3299,
      "theoretical_loss": 3.4057708465769814,
      "tokens_seen": 2419982336
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 881081,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4548755884170532,
      "objective/train/theoretical_loss": 3.405763503543188,
      "objective/train/tokens_used": 790506976,
      "theoretical_loss": 3.405763503543188,
      "tokens_seen": 2420047872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002693881644934805,
      "loss": 1.3281,
      "theoretical_loss": 3.405763503543188,
      "tokens_seen": 2420047872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026936810431293885,
      "loss": 1.2841,
      "theoretical_loss": 3.4057561607639215,
      "tokens_seen": 2420113408
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002693480441323972,
      "loss": 1.3131,
      "theoretical_loss": 3.4057488182391666,
      "tokens_seen": 2420178944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026932798395185557,
      "loss": 1.5235,
      "theoretical_loss": 3.4057414759689073,
      "tokens_seen": 2420244480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269307923771314,
      "loss": 1.3,
      "theoretical_loss": 3.405734133953128,
      "tokens_seen": 2420310016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026928786359077235,
      "loss": 1.3561,
      "theoretical_loss": 3.4057267921918126,
      "tokens_seen": 2420375552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692678034102307,
      "loss": 1.3014,
      "theoretical_loss": 3.4057194506849457,
      "tokens_seen": 2420441088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692477432296891,
      "loss": 1.368,
      "theoretical_loss": 3.4057121094325113,
      "tokens_seen": 2420506624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692276830491475,
      "loss": 1.3419,
      "theoretical_loss": 3.4057047684344948,
      "tokens_seen": 2420572160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026920762286860585,
      "loss": 1.31,
      "theoretical_loss": 3.405697427690879,
      "tokens_seen": 2420637696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691875626880642,
      "loss": 1.2438,
      "theoretical_loss": 3.4056900872016493,
      "tokens_seen": 2420703232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691675025075226,
      "loss": 1.2784,
      "theoretical_loss": 3.4056827469667894,
      "tokens_seen": 2420768768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026914744232698094,
      "loss": 1.3955,
      "theoretical_loss": 3.4056754069862833,
      "tokens_seen": 2420834304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026912738214643936,
      "loss": 1.2967,
      "theoretical_loss": 3.405668067260116,
      "tokens_seen": 2420899840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691073219658977,
      "loss": 1.351,
      "theoretical_loss": 3.405660727788272,
      "tokens_seen": 2420965376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690872617853561,
      "loss": 1.3782,
      "theoretical_loss": 3.4056533885707347,
      "tokens_seen": 2421030912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026906720160481444,
      "loss": 1.328,
      "theoretical_loss": 3.4056460496074896,
      "tokens_seen": 2421096448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026904714142427286,
      "loss": 1.411,
      "theoretical_loss": 3.4056387108985193,
      "tokens_seen": 2421161984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690270812437312,
      "loss": 1.3244,
      "theoretical_loss": 3.4056313724438096,
      "tokens_seen": 2421227520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690070210631896,
      "loss": 1.3849,
      "theoretical_loss": 3.4056240342433446,
      "tokens_seen": 2421293056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026898696088264795,
      "loss": 1.2887,
      "theoretical_loss": 3.4056166962971077,
      "tokens_seen": 2421358592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689669007021063,
      "loss": 1.4169,
      "theoretical_loss": 3.4056093586050844,
      "tokens_seen": 2421424128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689468405215647,
      "loss": 1.4382,
      "theoretical_loss": 3.4056020211672577,
      "tokens_seen": 2421489664
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689267803410231,
      "loss": 1.2975,
      "theoretical_loss": 3.4055946839836135,
      "tokens_seen": 2421555200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026890672016048145,
      "loss": 1.3015,
      "theoretical_loss": 3.4055873470541345,
      "tokens_seen": 2421620736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688866599799398,
      "loss": 1.3456,
      "theoretical_loss": 3.4055800103788063,
      "tokens_seen": 2421686272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026886659979939823,
      "loss": 1.4122,
      "theoretical_loss": 3.4055726739576127,
      "tokens_seen": 2421751808
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688465396188566,
      "loss": 1.3985,
      "theoretical_loss": 3.4055653377905375,
      "tokens_seen": 2421817344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026882647943831495,
      "loss": 1.2206,
      "theoretical_loss": 3.4055580018775657,
      "tokens_seen": 2421882880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688064192577733,
      "loss": 1.4452,
      "theoretical_loss": 3.405550666218682,
      "tokens_seen": 2421948416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687863590772317,
      "loss": 1.2921,
      "theoretical_loss": 3.40554333081387,
      "tokens_seen": 2422013952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687662988966901,
      "loss": 1.3464,
      "theoretical_loss": 3.4055359956631137,
      "tokens_seen": 2422079488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026874623871614846,
      "loss": 1.4432,
      "theoretical_loss": 3.4055286607663984,
      "tokens_seen": 2422145024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687261785356068,
      "loss": 1.1195,
      "theoretical_loss": 3.405521326123708,
      "tokens_seen": 2422210560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687061183550652,
      "loss": 1.3454,
      "theoretical_loss": 3.4055139917350266,
      "tokens_seen": 2422276096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686860581745236,
      "loss": 1.3041,
      "theoretical_loss": 3.4055066576003385,
      "tokens_seen": 2422341632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026866599799398196,
      "loss": 1.368,
      "theoretical_loss": 3.4054993237196287,
      "tokens_seen": 2422407168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686459378134403,
      "loss": 1.4555,
      "theoretical_loss": 3.405491990092881,
      "tokens_seen": 2422472704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686258776328987,
      "loss": 1.3074,
      "theoretical_loss": 3.4054846567200796,
      "tokens_seen": 2422538240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026860581745235705,
      "loss": 1.4823,
      "theoretical_loss": 3.4054773236012097,
      "tokens_seen": 2422603776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026858575727181546,
      "loss": 1.2772,
      "theoretical_loss": 3.4054699907362544,
      "tokens_seen": 2422669312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685656970912738,
      "loss": 1.3256,
      "theoretical_loss": 3.4054626581251988,
      "tokens_seen": 2422734848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685456369107322,
      "loss": 1.2827,
      "theoretical_loss": 3.405455325768027,
      "tokens_seen": 2422800384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026852557673019055,
      "loss": 1.2093,
      "theoretical_loss": 3.4054479936647235,
      "tokens_seen": 2422865920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026850551654964897,
      "loss": 1.3149,
      "theoretical_loss": 3.405440661815273,
      "tokens_seen": 2422931456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026848545636910733,
      "loss": 1.3921,
      "theoretical_loss": 3.4054333302196587,
      "tokens_seen": 2422996992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684653961885657,
      "loss": 1.3477,
      "theoretical_loss": 3.4054259988778663,
      "tokens_seen": 2423062528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026844533600802405,
      "loss": 1.3062,
      "theoretical_loss": 3.405418667789879,
      "tokens_seen": 2423128064
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684252758274824,
      "loss": 1.2946,
      "theoretical_loss": 3.405411336955682,
      "tokens_seen": 2423193600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026840521564694083,
      "loss": 1.3284,
      "theoretical_loss": 3.4054040063752593,
      "tokens_seen": 2423259136
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 882232,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3485541343688965,
      "objective/train/theoretical_loss": 3.405396676048595,
      "objective/train/tokens_used": 793783776,
      "theoretical_loss": 3.405396676048595,
      "tokens_seen": 2423324672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683851554663992,
      "loss": 1.3357,
      "theoretical_loss": 3.405396676048595,
      "tokens_seen": 2423324672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026836509528585756,
      "loss": 1.3971,
      "theoretical_loss": 3.405389345975674,
      "tokens_seen": 2423390208
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683450351053159,
      "loss": 1.2751,
      "theoretical_loss": 3.40538201615648,
      "tokens_seen": 2423455744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026832497492477433,
      "loss": 1.4781,
      "theoretical_loss": 3.4053746865909984,
      "tokens_seen": 2423521280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683049147442327,
      "loss": 1.3827,
      "theoretical_loss": 3.4053673572792125,
      "tokens_seen": 2423586816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026828485456369106,
      "loss": 1.4542,
      "theoretical_loss": 3.405360028221107,
      "tokens_seen": 2423652352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682647943831494,
      "loss": 1.4376,
      "theoretical_loss": 3.4053526994166665,
      "tokens_seen": 2423717888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682447342026078,
      "loss": 1.3843,
      "theoretical_loss": 3.405345370865875,
      "tokens_seen": 2423783424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682246740220662,
      "loss": 1.5332,
      "theoretical_loss": 3.405338042568717,
      "tokens_seen": 2423848960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026820461384152456,
      "loss": 1.3222,
      "theoretical_loss": 3.405330714525177,
      "tokens_seen": 2423914496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681845536609829,
      "loss": 1.3028,
      "theoretical_loss": 3.4053233867352395,
      "tokens_seen": 2423980032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681644934804413,
      "loss": 1.4225,
      "theoretical_loss": 3.405316059198888,
      "tokens_seen": 2424045568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681444332998997,
      "loss": 1.3292,
      "theoretical_loss": 3.405308731916108,
      "tokens_seen": 2424111104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026812437311935807,
      "loss": 1.377,
      "theoretical_loss": 3.4053014048868833,
      "tokens_seen": 2424176640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026810431293881643,
      "loss": 1.3468,
      "theoretical_loss": 3.4052940781111984,
      "tokens_seen": 2424242176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002680842527582748,
      "loss": 1.3393,
      "theoretical_loss": 3.4052867515890375,
      "tokens_seen": 2424307712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026806419257773315,
      "loss": 1.2582,
      "theoretical_loss": 3.405279425320385,
      "tokens_seen": 2424373248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026804413239719157,
      "loss": 1.3182,
      "theoretical_loss": 3.4052720993052255,
      "tokens_seen": 2424438784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026802407221664993,
      "loss": 1.261,
      "theoretical_loss": 3.4052647735435433,
      "tokens_seen": 2424504320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026800401203610835,
      "loss": 1.3059,
      "theoretical_loss": 3.4052574480353224,
      "tokens_seen": 2424569856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002679839518555667,
      "loss": 1.3245,
      "theoretical_loss": 3.405250122780548,
      "tokens_seen": 2424635392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002679638916750251,
      "loss": 1.2669,
      "theoretical_loss": 3.405242797779204,
      "tokens_seen": 2424700928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002679438314944835,
      "loss": 1.2908,
      "theoretical_loss": 3.405235473031274,
      "tokens_seen": 2424766464
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026792377131394185,
      "loss": 1.2577,
      "theoretical_loss": 3.405228148536744,
      "tokens_seen": 2424832000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002679037111334002,
      "loss": 1.4439,
      "theoretical_loss": 3.405220824295597,
      "tokens_seen": 2424897536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026788365095285863,
      "loss": 1.3717,
      "theoretical_loss": 3.405213500307818,
      "tokens_seen": 2424963072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267863590772317,
      "loss": 1.3757,
      "theoretical_loss": 3.4052061765733916,
      "tokens_seen": 2425028608
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026784353059177535,
      "loss": 1.3104,
      "theoretical_loss": 3.4051988530923016,
      "tokens_seen": 2425094144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678234704112337,
      "loss": 1.2968,
      "theoretical_loss": 3.405191529864533,
      "tokens_seen": 2425159680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678034102306921,
      "loss": 1.3218,
      "theoretical_loss": 3.4051842068900697,
      "tokens_seen": 2425225216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677833500501505,
      "loss": 1.504,
      "theoretical_loss": 3.4051768841688963,
      "tokens_seen": 2425290752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026776328986960886,
      "loss": 1.2117,
      "theoretical_loss": 3.4051695617009967,
      "tokens_seen": 2425356288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677432296890672,
      "loss": 1.3506,
      "theoretical_loss": 3.4051622394863563,
      "tokens_seen": 2425421824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677231695085256,
      "loss": 1.4068,
      "theoretical_loss": 3.405154917524959,
      "tokens_seen": 2425487360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267703109327984,
      "loss": 1.2232,
      "theoretical_loss": 3.405147595816789,
      "tokens_seen": 2425552896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026768304914744236,
      "loss": 1.3203,
      "theoretical_loss": 3.405140274361831,
      "tokens_seen": 2425618432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676629889669007,
      "loss": 1.2788,
      "theoretical_loss": 3.405132953160069,
      "tokens_seen": 2425683968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676429287863591,
      "loss": 1.4393,
      "theoretical_loss": 3.405125632211488,
      "tokens_seen": 2425749504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026762286860581745,
      "loss": 1.3712,
      "theoretical_loss": 3.405118311516072,
      "tokens_seen": 2425815040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026760280842527586,
      "loss": 1.3375,
      "theoretical_loss": 3.4051109910738053,
      "tokens_seen": 2425880576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002675827482447342,
      "loss": 1.4066,
      "theoretical_loss": 3.4051036708846727,
      "tokens_seen": 2425946112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002675626880641926,
      "loss": 1.3846,
      "theoretical_loss": 3.405096350948658,
      "tokens_seen": 2426011648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026754262788365095,
      "loss": 1.3323,
      "theoretical_loss": 3.4050890312657467,
      "tokens_seen": 2426077184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026752256770310937,
      "loss": 1.2681,
      "theoretical_loss": 3.4050817118359222,
      "tokens_seen": 2426142720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026750250752256773,
      "loss": 1.3565,
      "theoretical_loss": 3.405074392659169,
      "tokens_seen": 2426208256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674824473420261,
      "loss": 1.5088,
      "theoretical_loss": 3.405067073735472,
      "tokens_seen": 2426273792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026746238716148445,
      "loss": 1.3013,
      "theoretical_loss": 3.405059755064815,
      "tokens_seen": 2426339328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674423269809428,
      "loss": 1.2945,
      "theoretical_loss": 3.4050524366471833,
      "tokens_seen": 2426404864
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026742226680040123,
      "loss": 1.3537,
      "theoretical_loss": 3.4050451184825605,
      "tokens_seen": 2426470400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674022066198596,
      "loss": 1.2138,
      "theoretical_loss": 3.4050378005709314,
      "tokens_seen": 2426535936
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 883303,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3030461072921753,
      "objective/train/theoretical_loss": 3.4050304829122804,
      "objective/train/tokens_used": 797060576,
      "theoretical_loss": 3.4050304829122804,
      "tokens_seen": 2426601472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026738214643931796,
      "loss": 1.258,
      "theoretical_loss": 3.4050304829122804,
      "tokens_seen": 2426601472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673620862587763,
      "loss": 1.4453,
      "theoretical_loss": 3.405023165506592,
      "tokens_seen": 2426667008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026734202607823474,
      "loss": 1.33,
      "theoretical_loss": 3.40501584835385,
      "tokens_seen": 2426732544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673219658976931,
      "loss": 1.4086,
      "theoretical_loss": 3.4050085314540395,
      "tokens_seen": 2426798080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026730190571715146,
      "loss": 1.2477,
      "theoretical_loss": 3.405001214807145,
      "tokens_seen": 2426863616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672818455366098,
      "loss": 1.2665,
      "theoretical_loss": 3.4049938984131507,
      "tokens_seen": 2426929152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672617853560682,
      "loss": 1.1864,
      "theoretical_loss": 3.4049865822720404,
      "tokens_seen": 2426994688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672417251755266,
      "loss": 1.2912,
      "theoretical_loss": 3.4049792663838,
      "tokens_seen": 2427060224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026722166499498496,
      "loss": 1.2859,
      "theoretical_loss": 3.404971950748412,
      "tokens_seen": 2427125760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672016048144433,
      "loss": 1.3571,
      "theoretical_loss": 3.4049646353658627,
      "tokens_seen": 2427191296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671815446339017,
      "loss": 1.2856,
      "theoretical_loss": 3.404957320236136,
      "tokens_seen": 2427256832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671614844533601,
      "loss": 1.25,
      "theoretical_loss": 3.404950005359215,
      "tokens_seen": 2427322368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026714142427281847,
      "loss": 1.258,
      "theoretical_loss": 3.404942690735086,
      "tokens_seen": 2427387904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026712136409227683,
      "loss": 1.3684,
      "theoretical_loss": 3.4049353763637327,
      "tokens_seen": 2427453440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671013039117352,
      "loss": 1.4446,
      "theoretical_loss": 3.404928062245139,
      "tokens_seen": 2427518976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026708124373119355,
      "loss": 1.2308,
      "theoretical_loss": 3.40492074837929,
      "tokens_seen": 2427584512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026706118355065197,
      "loss": 1.3159,
      "theoretical_loss": 3.4049134347661703,
      "tokens_seen": 2427650048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026704112337011033,
      "loss": 1.4214,
      "theoretical_loss": 3.4049061214057637,
      "tokens_seen": 2427715584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002670210631895687,
      "loss": 1.4214,
      "theoretical_loss": 3.404898808298055,
      "tokens_seen": 2427781120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026700100300902706,
      "loss": 1.391,
      "theoretical_loss": 3.4048914954430285,
      "tokens_seen": 2427846656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026698094282848547,
      "loss": 1.3988,
      "theoretical_loss": 3.404884182840669,
      "tokens_seen": 2427912192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026696088264794383,
      "loss": 1.2703,
      "theoretical_loss": 3.4048768704909604,
      "tokens_seen": 2427977728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002669408224674022,
      "loss": 1.3277,
      "theoretical_loss": 3.4048695583938877,
      "tokens_seen": 2428043264
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026692076228686056,
      "loss": 1.3554,
      "theoretical_loss": 3.404862246549435,
      "tokens_seen": 2428108800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000266900702106319,
      "loss": 1.5205,
      "theoretical_loss": 3.404854934957587,
      "tokens_seen": 2428174336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026688064192577734,
      "loss": 1.3562,
      "theoretical_loss": 3.404847623618328,
      "tokens_seen": 2428239872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668605817452357,
      "loss": 1.4315,
      "theoretical_loss": 3.4048403125316424,
      "tokens_seen": 2428305408
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026684052156469406,
      "loss": 1.2801,
      "theoretical_loss": 3.404833001697515,
      "tokens_seen": 2428370944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668204613841524,
      "loss": 1.3293,
      "theoretical_loss": 3.4048256911159296,
      "tokens_seen": 2428436480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026680040120361084,
      "loss": 1.3755,
      "theoretical_loss": 3.404818380786871,
      "tokens_seen": 2428502016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667803410230692,
      "loss": 1.2545,
      "theoretical_loss": 3.4048110707103243,
      "tokens_seen": 2428567552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026676028084252757,
      "loss": 1.3037,
      "theoretical_loss": 3.404803760886273,
      "tokens_seen": 2428633088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026674022066198593,
      "loss": 1.3218,
      "theoretical_loss": 3.4047964513147018,
      "tokens_seen": 2428698624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026672016048144434,
      "loss": 1.3399,
      "theoretical_loss": 3.404789141995596,
      "tokens_seen": 2428764160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667001003009027,
      "loss": 1.1972,
      "theoretical_loss": 3.4047818329289385,
      "tokens_seen": 2428829696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026668004012036107,
      "loss": 1.3512,
      "theoretical_loss": 3.404774524114715,
      "tokens_seen": 2428895232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026665997993981943,
      "loss": 1.2622,
      "theoretical_loss": 3.40476721555291,
      "tokens_seen": 2428960768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666399197592778,
      "loss": 1.3282,
      "theoretical_loss": 3.404759907243507,
      "tokens_seen": 2429026304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666198595787362,
      "loss": 1.2539,
      "theoretical_loss": 3.4047525991864913,
      "tokens_seen": 2429091840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026659979939819457,
      "loss": 1.4264,
      "theoretical_loss": 3.4047452913818472,
      "tokens_seen": 2429157376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026657973921765293,
      "loss": 1.3425,
      "theoretical_loss": 3.404737983829559,
      "tokens_seen": 2429222912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665596790371113,
      "loss": 1.2743,
      "theoretical_loss": 3.4047306765296117,
      "tokens_seen": 2429288448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665396188565697,
      "loss": 1.33,
      "theoretical_loss": 3.404723369481989,
      "tokens_seen": 2429353984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665195586760281,
      "loss": 1.2016,
      "theoretical_loss": 3.404716062686676,
      "tokens_seen": 2429419520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664994984954865,
      "loss": 1.2349,
      "theoretical_loss": 3.4047087561436564,
      "tokens_seen": 2429485056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026647943831494485,
      "loss": 1.3166,
      "theoretical_loss": 3.404701449852916,
      "tokens_seen": 2429550592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664593781344032,
      "loss": 1.3174,
      "theoretical_loss": 3.4046941438144382,
      "tokens_seen": 2429616128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026643931795386163,
      "loss": 1.3426,
      "theoretical_loss": 3.4046868380282076,
      "tokens_seen": 2429681664
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026641925777332,
      "loss": 1.3944,
      "theoretical_loss": 3.4046795324942094,
      "tokens_seen": 2429747200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026639919759277836,
      "loss": 1.2684,
      "theoretical_loss": 3.404672227212427,
      "tokens_seen": 2429812736
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 884501,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3901500701904297,
      "objective/train/theoretical_loss": 3.4046649221828456,
      "objective/train/tokens_used": 800337376,
      "theoretical_loss": 3.4046649221828456,
      "tokens_seen": 2429878272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663791374122367,
      "loss": 1.3589,
      "theoretical_loss": 3.4046649221828456,
      "tokens_seen": 2429878272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026635907723169514,
      "loss": 1.4137,
      "theoretical_loss": 3.4046576174054497,
      "tokens_seen": 2429943808
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663390170511535,
      "loss": 1.2823,
      "theoretical_loss": 3.4046503128802237,
      "tokens_seen": 2430009344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026631895687061186,
      "loss": 1.3052,
      "theoretical_loss": 3.404643008607152,
      "tokens_seen": 2430074880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662988966900702,
      "loss": 1.374,
      "theoretical_loss": 3.404635704586219,
      "tokens_seen": 2430140416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662788365095286,
      "loss": 1.3538,
      "theoretical_loss": 3.4046284008174093,
      "tokens_seen": 2430205952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000266258776328987,
      "loss": 1.4029,
      "theoretical_loss": 3.4046210973007076,
      "tokens_seen": 2430271488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026623871614844536,
      "loss": 1.4687,
      "theoretical_loss": 3.404613794036098,
      "tokens_seen": 2430337024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662186559679037,
      "loss": 1.4055,
      "theoretical_loss": 3.4046064910235656,
      "tokens_seen": 2430402560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661985957873621,
      "loss": 1.2183,
      "theoretical_loss": 3.404599188263094,
      "tokens_seen": 2430468096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661785356068205,
      "loss": 1.2853,
      "theoretical_loss": 3.4045918857546686,
      "tokens_seen": 2430533632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026615847542627887,
      "loss": 1.3001,
      "theoretical_loss": 3.404584583498274,
      "tokens_seen": 2430599168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026613841524573723,
      "loss": 1.4721,
      "theoretical_loss": 3.4045772814938937,
      "tokens_seen": 2430664704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661183550651956,
      "loss": 1.1923,
      "theoretical_loss": 3.4045699797415128,
      "tokens_seen": 2430730240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026609829488465395,
      "loss": 1.3096,
      "theoretical_loss": 3.404562678241116,
      "tokens_seen": 2430795776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026607823470411237,
      "loss": 1.2664,
      "theoretical_loss": 3.404555376992687,
      "tokens_seen": 2430861312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026605817452357073,
      "loss": 1.2991,
      "theoretical_loss": 3.4045480759962117,
      "tokens_seen": 2430926848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002660381143430291,
      "loss": 1.3548,
      "theoretical_loss": 3.4045407752516734,
      "tokens_seen": 2430992384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026601805416248746,
      "loss": 1.2728,
      "theoretical_loss": 3.4045334747590568,
      "tokens_seen": 2431057920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026599799398194587,
      "loss": 1.394,
      "theoretical_loss": 3.404526174518347,
      "tokens_seen": 2431123456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026597793380140423,
      "loss": 1.3311,
      "theoretical_loss": 3.4045188745295283,
      "tokens_seen": 2431188992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002659578736208626,
      "loss": 1.2233,
      "theoretical_loss": 3.4045115747925845,
      "tokens_seen": 2431254528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026593781344032096,
      "loss": 1.5064,
      "theoretical_loss": 3.404504275307501,
      "tokens_seen": 2431320064
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002659177532597794,
      "loss": 1.2455,
      "theoretical_loss": 3.4044969760742623,
      "tokens_seen": 2431385600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026589769307923774,
      "loss": 1.2831,
      "theoretical_loss": 3.404489677092852,
      "tokens_seen": 2431451136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002658776328986961,
      "loss": 1.3527,
      "theoretical_loss": 3.404482378363256,
      "tokens_seen": 2431516672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026585757271815446,
      "loss": 1.306,
      "theoretical_loss": 3.4044750798854575,
      "tokens_seen": 2431582208
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002658375125376128,
      "loss": 1.3541,
      "theoretical_loss": 3.4044677816594415,
      "tokens_seen": 2431647744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026581745235707124,
      "loss": 1.4123,
      "theoretical_loss": 3.4044604836851935,
      "tokens_seen": 2431713280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657973921765296,
      "loss": 1.2521,
      "theoretical_loss": 3.4044531859626965,
      "tokens_seen": 2431778816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026577733199598797,
      "loss": 1.2421,
      "theoretical_loss": 3.404445888491936,
      "tokens_seen": 2431844352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026575727181544633,
      "loss": 1.3793,
      "theoretical_loss": 3.404438591272896,
      "tokens_seen": 2431909888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026573721163490474,
      "loss": 1.3008,
      "theoretical_loss": 3.404431294305561,
      "tokens_seen": 2431975424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657171514543631,
      "loss": 1.2833,
      "theoretical_loss": 3.4044239975899164,
      "tokens_seen": 2432040960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026569709127382147,
      "loss": 1.4058,
      "theoretical_loss": 3.404416701125946,
      "tokens_seen": 2432106496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026567703109327983,
      "loss": 1.3082,
      "theoretical_loss": 3.4044094049136344,
      "tokens_seen": 2432172032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656569709127382,
      "loss": 1.3279,
      "theoretical_loss": 3.4044021089529664,
      "tokens_seen": 2432237568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656369107321966,
      "loss": 1.2609,
      "theoretical_loss": 3.404394813243926,
      "tokens_seen": 2432303104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026561685055165497,
      "loss": 1.3395,
      "theoretical_loss": 3.4043875177864984,
      "tokens_seen": 2432368640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026559679037111333,
      "loss": 1.2916,
      "theoretical_loss": 3.4043802225806674,
      "tokens_seen": 2432434176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655767301905717,
      "loss": 1.3477,
      "theoretical_loss": 3.4043729276264187,
      "tokens_seen": 2432499712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655566700100301,
      "loss": 1.3896,
      "theoretical_loss": 3.4043656329237355,
      "tokens_seen": 2432565248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655366098294885,
      "loss": 1.2429,
      "theoretical_loss": 3.4043583384726035,
      "tokens_seen": 2432630784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026551654964894684,
      "loss": 1.4478,
      "theoretical_loss": 3.4043510442730067,
      "tokens_seen": 2432696320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002654964894684052,
      "loss": 1.3163,
      "theoretical_loss": 3.404343750324929,
      "tokens_seen": 2432761856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026547642928786356,
      "loss": 1.2848,
      "theoretical_loss": 3.4043364566283563,
      "tokens_seen": 2432827392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000265456369107322,
      "loss": 1.3382,
      "theoretical_loss": 3.4043291631832724,
      "tokens_seen": 2432892928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026543630892678034,
      "loss": 1.2891,
      "theoretical_loss": 3.404321869989662,
      "tokens_seen": 2432958464
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002654162487462387,
      "loss": 1.2454,
      "theoretical_loss": 3.4043145770475096,
      "tokens_seen": 2433024000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026539618856569707,
      "loss": 1.3167,
      "theoretical_loss": 3.4043072843567996,
      "tokens_seen": 2433089536
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 885704,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2719991207122803,
      "objective/train/theoretical_loss": 3.4042999919175165,
      "objective/train/tokens_used": 803614176,
      "theoretical_loss": 3.4042999919175165,
      "tokens_seen": 2433155072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653761283851555,
      "loss": 1.2637,
      "theoretical_loss": 3.4042999919175165,
      "tokens_seen": 2433155072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026535606820461384,
      "loss": 1.352,
      "theoretical_loss": 3.4042926997296457,
      "tokens_seen": 2433220608
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653360080240722,
      "loss": 1.3194,
      "theoretical_loss": 3.4042854077931706,
      "tokens_seen": 2433286144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026531594784353057,
      "loss": 1.2921,
      "theoretical_loss": 3.4042781161080766,
      "tokens_seen": 2433351680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026529588766298893,
      "loss": 1.3071,
      "theoretical_loss": 3.404270824674348,
      "tokens_seen": 2433417216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026527582748244735,
      "loss": 1.23,
      "theoretical_loss": 3.404263533491969,
      "tokens_seen": 2433482752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002652557673019057,
      "loss": 1.2463,
      "theoretical_loss": 3.4042562425609244,
      "tokens_seen": 2433548288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026523570712136407,
      "loss": 1.405,
      "theoretical_loss": 3.404248951881199,
      "tokens_seen": 2433613824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026521564694082243,
      "loss": 1.3834,
      "theoretical_loss": 3.4042416614527777,
      "tokens_seen": 2433679360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026519558676028085,
      "loss": 1.4448,
      "theoretical_loss": 3.404234371275644,
      "tokens_seen": 2433744896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651755265797392,
      "loss": 1.3802,
      "theoretical_loss": 3.4042270813497835,
      "tokens_seen": 2433810432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651554663991976,
      "loss": 1.3237,
      "theoretical_loss": 3.4042197916751804,
      "tokens_seen": 2433875968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026513540621865594,
      "loss": 1.2549,
      "theoretical_loss": 3.404212502251819,
      "tokens_seen": 2433941504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651153460381143,
      "loss": 1.431,
      "theoretical_loss": 3.404205213079684,
      "tokens_seen": 2434007040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650952858575727,
      "loss": 1.3434,
      "theoretical_loss": 3.40419792415876,
      "tokens_seen": 2434072576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650752256770311,
      "loss": 1.4047,
      "theoretical_loss": 3.404190635489032,
      "tokens_seen": 2434138112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026505516549648944,
      "loss": 1.2805,
      "theoretical_loss": 3.404183347070484,
      "tokens_seen": 2434203648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650351053159478,
      "loss": 1.2965,
      "theoretical_loss": 3.404176058903101,
      "tokens_seen": 2434269184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650150451354062,
      "loss": 1.2121,
      "theoretical_loss": 3.4041687709868675,
      "tokens_seen": 2434334720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026499498495486464,
      "loss": 1.3883,
      "theoretical_loss": 3.4041614833217677,
      "tokens_seen": 2434400256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264974924774323,
      "loss": 1.3495,
      "theoretical_loss": 3.404154195907787,
      "tokens_seen": 2434465792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026495486459378136,
      "loss": 1.3549,
      "theoretical_loss": 3.404146908744909,
      "tokens_seen": 2434531328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002649348044132398,
      "loss": 1.3597,
      "theoretical_loss": 3.404139621833119,
      "tokens_seen": 2434596864
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026491474423269814,
      "loss": 1.2416,
      "theoretical_loss": 3.404132335172401,
      "tokens_seen": 2434662400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648946840521565,
      "loss": 1.304,
      "theoretical_loss": 3.40412504876274,
      "tokens_seen": 2434727936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026487462387161486,
      "loss": 1.307,
      "theoretical_loss": 3.4041177626041206,
      "tokens_seen": 2434793472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648545636910732,
      "loss": 1.3008,
      "theoretical_loss": 3.4041104766965273,
      "tokens_seen": 2434859008
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026483450351053164,
      "loss": 1.2966,
      "theoretical_loss": 3.404103191039945,
      "tokens_seen": 2434924544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026481444332999,
      "loss": 1.2517,
      "theoretical_loss": 3.4040959056343576,
      "tokens_seen": 2434990080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026479438314944837,
      "loss": 1.2835,
      "theoretical_loss": 3.4040886204797505,
      "tokens_seen": 2435055616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026477432296890673,
      "loss": 1.4008,
      "theoretical_loss": 3.4040813355761075,
      "tokens_seen": 2435121152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026475426278836514,
      "loss": 1.3413,
      "theoretical_loss": 3.404074050923414,
      "tokens_seen": 2435186688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647342026078235,
      "loss": 1.3696,
      "theoretical_loss": 3.404066766521654,
      "tokens_seen": 2435252224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026471414242728187,
      "loss": 1.3266,
      "theoretical_loss": 3.4040594823708124,
      "tokens_seen": 2435317760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026469408224674023,
      "loss": 1.2926,
      "theoretical_loss": 3.4040521984708736,
      "tokens_seen": 2435383296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646740220661986,
      "loss": 1.2218,
      "theoretical_loss": 3.404044914821822,
      "tokens_seen": 2435448832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264653961885657,
      "loss": 1.2744,
      "theoretical_loss": 3.404037631423643,
      "tokens_seen": 2435514368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026463390170511537,
      "loss": 1.3833,
      "theoretical_loss": 3.404030348276321,
      "tokens_seen": 2435579904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026461384152457373,
      "loss": 1.3537,
      "theoretical_loss": 3.40402306537984,
      "tokens_seen": 2435645440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645937813440321,
      "loss": 1.2694,
      "theoretical_loss": 3.404015782734185,
      "tokens_seen": 2435710976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645737211634905,
      "loss": 1.2989,
      "theoretical_loss": 3.4040085003393403,
      "tokens_seen": 2435776512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645536609829489,
      "loss": 1.3162,
      "theoretical_loss": 3.4040012181952912,
      "tokens_seen": 2435842048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026453360080240724,
      "loss": 1.3371,
      "theoretical_loss": 3.4039939363020215,
      "tokens_seen": 2435907584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645135406218656,
      "loss": 1.273,
      "theoretical_loss": 3.4039866546595166,
      "tokens_seen": 2435973120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026449348044132396,
      "loss": 1.236,
      "theoretical_loss": 3.4039793732677603,
      "tokens_seen": 2436038656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002644734202607824,
      "loss": 1.3863,
      "theoretical_loss": 3.4039720921267382,
      "tokens_seen": 2436104192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026445336008024074,
      "loss": 1.3751,
      "theoretical_loss": 3.4039648112364342,
      "tokens_seen": 2436169728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002644332998996991,
      "loss": 1.4152,
      "theoretical_loss": 3.4039575305968333,
      "tokens_seen": 2436235264
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026441323971915747,
      "loss": 1.3586,
      "theoretical_loss": 3.4039502502079193,
      "tokens_seen": 2436300800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002643931795386159,
      "loss": 1.2942,
      "theoretical_loss": 3.403942970069678,
      "tokens_seen": 2436366336
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 886917,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2300962209701538,
      "objective/train/theoretical_loss": 3.4039356901820934,
      "objective/train/tokens_used": 806890976,
      "theoretical_loss": 3.4039356901820934,
      "tokens_seen": 2436431872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026437311935807424,
      "loss": 1.3366,
      "theoretical_loss": 3.4039356901820934,
      "tokens_seen": 2436431872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002643530591775326,
      "loss": 1.302,
      "theoretical_loss": 3.4039284105451504,
      "tokens_seen": 2436497408
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026433299899699097,
      "loss": 1.335,
      "theoretical_loss": 3.4039211311588327,
      "tokens_seen": 2436562944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026431293881644933,
      "loss": 1.3802,
      "theoretical_loss": 3.4039138520231265,
      "tokens_seen": 2436628480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026429287863590775,
      "loss": 1.2518,
      "theoretical_loss": 3.4039065731380154,
      "tokens_seen": 2436694016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002642728184553661,
      "loss": 1.3233,
      "theoretical_loss": 3.403899294503484,
      "tokens_seen": 2436759552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026425275827482447,
      "loss": 1.3496,
      "theoretical_loss": 3.403892016119517,
      "tokens_seen": 2436825088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026423269809428283,
      "loss": 1.2334,
      "theoretical_loss": 3.4038847379860995,
      "tokens_seen": 2436890624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026421263791374125,
      "loss": 1.2668,
      "theoretical_loss": 3.403877460103216,
      "tokens_seen": 2436956160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641925777331996,
      "loss": 1.2685,
      "theoretical_loss": 3.403870182470851,
      "tokens_seen": 2437021696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264172517552658,
      "loss": 1.3218,
      "theoretical_loss": 3.403862905088989,
      "tokens_seen": 2437087232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026415245737211634,
      "loss": 1.283,
      "theoretical_loss": 3.4038556279576144,
      "tokens_seen": 2437152768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641323971915747,
      "loss": 1.3254,
      "theoretical_loss": 3.4038483510767126,
      "tokens_seen": 2437218304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641123370110331,
      "loss": 1.3425,
      "theoretical_loss": 3.4038410744462677,
      "tokens_seen": 2437283840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640922768304915,
      "loss": 1.2866,
      "theoretical_loss": 3.4038337980662643,
      "tokens_seen": 2437349376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026407221664994984,
      "loss": 1.3734,
      "theoretical_loss": 3.403826521936687,
      "tokens_seen": 2437414912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640521564694082,
      "loss": 1.3671,
      "theoretical_loss": 3.403819246057522,
      "tokens_seen": 2437480448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640320962888666,
      "loss": 1.3858,
      "theoretical_loss": 3.4038119704287517,
      "tokens_seen": 2437545984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264012036108325,
      "loss": 1.3578,
      "theoretical_loss": 3.4038046950503618,
      "tokens_seen": 2437611520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026399197592778334,
      "loss": 1.4024,
      "theoretical_loss": 3.403797419922337,
      "tokens_seen": 2437677056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002639719157472417,
      "loss": 1.2692,
      "theoretical_loss": 3.4037901450446615,
      "tokens_seen": 2437742592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026395185556670007,
      "loss": 1.3677,
      "theoretical_loss": 3.40378287041732,
      "tokens_seen": 2437808128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002639317953861585,
      "loss": 1.3466,
      "theoretical_loss": 3.403775596040298,
      "tokens_seen": 2437873664
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026391173520561685,
      "loss": 1.2702,
      "theoretical_loss": 3.4037683219135793,
      "tokens_seen": 2437939200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002638916750250752,
      "loss": 1.38,
      "theoretical_loss": 3.4037610480371487,
      "tokens_seen": 2438004736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026387161484453357,
      "loss": 1.4064,
      "theoretical_loss": 3.4037537744109914,
      "tokens_seen": 2438070272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000263851554663992,
      "loss": 1.4466,
      "theoretical_loss": 3.4037465010350916,
      "tokens_seen": 2438135808
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026383149448345035,
      "loss": 1.3869,
      "theoretical_loss": 3.403739227909434,
      "tokens_seen": 2438201344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002638114343029087,
      "loss": 1.1736,
      "theoretical_loss": 3.4037319550340026,
      "tokens_seen": 2438266880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637913741223671,
      "loss": 1.2752,
      "theoretical_loss": 3.4037246824087837,
      "tokens_seen": 2438332416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637713139418255,
      "loss": 1.2928,
      "theoretical_loss": 3.40371741003376,
      "tokens_seen": 2438397952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026375125376128385,
      "loss": 1.2996,
      "theoretical_loss": 3.4037101379089183,
      "tokens_seen": 2438463488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637311935807422,
      "loss": 1.3287,
      "theoretical_loss": 3.4037028660342417,
      "tokens_seen": 2438529024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637111334002006,
      "loss": 1.2418,
      "theoretical_loss": 3.403695594409715,
      "tokens_seen": 2438594560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026369107321965894,
      "loss": 1.4114,
      "theoretical_loss": 3.403688323035323,
      "tokens_seen": 2438660096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026367101303911736,
      "loss": 1.3061,
      "theoretical_loss": 3.403681051911051,
      "tokens_seen": 2438725632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636509528585757,
      "loss": 1.3155,
      "theoretical_loss": 3.4036737810368836,
      "tokens_seen": 2438791168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636308926780341,
      "loss": 1.31,
      "theoretical_loss": 3.4036665104128048,
      "tokens_seen": 2438856704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026361083249749244,
      "loss": 1.3362,
      "theoretical_loss": 3.4036592400387997,
      "tokens_seen": 2438922240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026359077231695086,
      "loss": 1.2689,
      "theoretical_loss": 3.4036519699148524,
      "tokens_seen": 2438987776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635707121364092,
      "loss": 1.3707,
      "theoretical_loss": 3.403644700040948,
      "tokens_seen": 2439053312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635506519558676,
      "loss": 1.3808,
      "theoretical_loss": 3.403637430417072,
      "tokens_seen": 2439118848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026353059177532595,
      "loss": 1.2902,
      "theoretical_loss": 3.403630161043208,
      "tokens_seen": 2439184384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026351053159478436,
      "loss": 1.2855,
      "theoretical_loss": 3.4036228919193405,
      "tokens_seen": 2439249920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634904714142428,
      "loss": 1.3485,
      "theoretical_loss": 3.4036156230454555,
      "tokens_seen": 2439315456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026347041123370114,
      "loss": 1.2817,
      "theoretical_loss": 3.4036083544215363,
      "tokens_seen": 2439380992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634503510531595,
      "loss": 1.3304,
      "theoretical_loss": 3.4036010860475683,
      "tokens_seen": 2439446528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026343029087261787,
      "loss": 1.3429,
      "theoretical_loss": 3.4035938179235363,
      "tokens_seen": 2439512064
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634102306920763,
      "loss": 1.2538,
      "theoretical_loss": 3.4035865500494245,
      "tokens_seen": 2439577600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026339017051153464,
      "loss": 1.3226,
      "theoretical_loss": 3.403579282425218,
      "tokens_seen": 2439643136
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 888050,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.376977562904358,
      "objective/train/theoretical_loss": 3.4035720150509015,
      "objective/train/tokens_used": 810167776,
      "theoretical_loss": 3.4035720150509015,
      "tokens_seen": 2439708672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000263370110330993,
      "loss": 1.432,
      "theoretical_loss": 3.4035720150509015,
      "tokens_seen": 2439708672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026335005015045137,
      "loss": 1.2653,
      "theoretical_loss": 3.403564747926459,
      "tokens_seen": 2439774208
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026332998996990973,
      "loss": 1.3982,
      "theoretical_loss": 3.403557481051876,
      "tokens_seen": 2439839744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026330992978936815,
      "loss": 1.2869,
      "theoretical_loss": 3.403550214427137,
      "tokens_seen": 2439905280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632898696088265,
      "loss": 1.3165,
      "theoretical_loss": 3.4035429480522263,
      "tokens_seen": 2439970816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026326980942828487,
      "loss": 1.337,
      "theoretical_loss": 3.4035356819271296,
      "tokens_seen": 2440036352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026324974924774323,
      "loss": 1.3174,
      "theoretical_loss": 3.4035284160518304,
      "tokens_seen": 2440101888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026322968906720165,
      "loss": 1.3245,
      "theoretical_loss": 3.403521150426314,
      "tokens_seen": 2440167424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026320962888666,
      "loss": 1.379,
      "theoretical_loss": 3.403513885050565,
      "tokens_seen": 2440232960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631895687061184,
      "loss": 1.2523,
      "theoretical_loss": 3.4035066199245683,
      "tokens_seen": 2440298496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026316950852557674,
      "loss": 1.2459,
      "theoretical_loss": 3.403499355048308,
      "tokens_seen": 2440364032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631494483450351,
      "loss": 1.2831,
      "theoretical_loss": 3.40349209042177,
      "tokens_seen": 2440429568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631293881644935,
      "loss": 1.3923,
      "theoretical_loss": 3.403484826044938,
      "tokens_seen": 2440495104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631093279839519,
      "loss": 1.3163,
      "theoretical_loss": 3.403477561917797,
      "tokens_seen": 2440560640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026308926780341024,
      "loss": 1.3579,
      "theoretical_loss": 3.4034702980403315,
      "tokens_seen": 2440626176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002630692076228686,
      "loss": 1.3234,
      "theoretical_loss": 3.4034630344125265,
      "tokens_seen": 2440691712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000263049147442327,
      "loss": 1.3498,
      "theoretical_loss": 3.4034557710343667,
      "tokens_seen": 2440757248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002630290872617854,
      "loss": 1.3074,
      "theoretical_loss": 3.4034485079058365,
      "tokens_seen": 2440822784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026300902708124374,
      "loss": 1.3297,
      "theoretical_loss": 3.4034412450269214,
      "tokens_seen": 2440888320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002629889669007021,
      "loss": 1.3327,
      "theoretical_loss": 3.4034339823976048,
      "tokens_seen": 2440953856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026296890672016047,
      "loss": 1.3149,
      "theoretical_loss": 3.403426720017873,
      "tokens_seen": 2441019392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002629488465396189,
      "loss": 1.3262,
      "theoretical_loss": 3.4034194578877095,
      "tokens_seen": 2441084928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026292878635907725,
      "loss": 1.395,
      "theoretical_loss": 3.4034121960070993,
      "tokens_seen": 2441150464
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002629087261785356,
      "loss": 1.3374,
      "theoretical_loss": 3.4034049343760278,
      "tokens_seen": 2441216000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026288866599799397,
      "loss": 1.266,
      "theoretical_loss": 3.403397672994479,
      "tokens_seen": 2441281536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628686058174524,
      "loss": 1.2661,
      "theoretical_loss": 3.4033904118624374,
      "tokens_seen": 2441347072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026284854563691075,
      "loss": 1.2804,
      "theoretical_loss": 3.4033831509798884,
      "tokens_seen": 2441412608
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628284854563691,
      "loss": 1.4598,
      "theoretical_loss": 3.4033758903468168,
      "tokens_seen": 2441478144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628084252758275,
      "loss": 1.3232,
      "theoretical_loss": 3.4033686299632064,
      "tokens_seen": 2441543680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002627883650952859,
      "loss": 1.414,
      "theoretical_loss": 3.4033613698290432,
      "tokens_seen": 2441609216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026276830491474425,
      "loss": 1.4005,
      "theoretical_loss": 3.4033541099443108,
      "tokens_seen": 2441674752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002627482447342026,
      "loss": 1.2182,
      "theoretical_loss": 3.4033468503089948,
      "tokens_seen": 2441740288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262728184553661,
      "loss": 1.3543,
      "theoretical_loss": 3.4033395909230792,
      "tokens_seen": 2441805824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026270812437311934,
      "loss": 1.2899,
      "theoretical_loss": 3.4033323317865496,
      "tokens_seen": 2441871360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026268806419257776,
      "loss": 1.4146,
      "theoretical_loss": 3.40332507289939,
      "tokens_seen": 2441936896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626680040120361,
      "loss": 1.4143,
      "theoretical_loss": 3.403317814261585,
      "tokens_seen": 2442002432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626479438314945,
      "loss": 1.369,
      "theoretical_loss": 3.4033105558731203,
      "tokens_seen": 2442067968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026262788365095284,
      "loss": 1.2887,
      "theoretical_loss": 3.40330329773398,
      "tokens_seen": 2442133504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026260782347041126,
      "loss": 1.2332,
      "theoretical_loss": 3.4032960398441485,
      "tokens_seen": 2442199040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625877632898696,
      "loss": 1.2901,
      "theoretical_loss": 3.403288782203611,
      "tokens_seen": 2442264576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262567703109328,
      "loss": 1.3541,
      "theoretical_loss": 3.4032815248123525,
      "tokens_seen": 2442330112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026254764292878635,
      "loss": 1.356,
      "theoretical_loss": 3.4032742676703576,
      "tokens_seen": 2442395648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625275827482447,
      "loss": 1.3593,
      "theoretical_loss": 3.4032670107776104,
      "tokens_seen": 2442461184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625075225677031,
      "loss": 1.3006,
      "theoretical_loss": 3.4032597541340968,
      "tokens_seen": 2442526720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624874623871615,
      "loss": 1.3696,
      "theoretical_loss": 3.4032524977398,
      "tokens_seen": 2442592256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026246740220661985,
      "loss": 1.2597,
      "theoretical_loss": 3.4032452415947065,
      "tokens_seen": 2442657792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624473420260782,
      "loss": 1.2731,
      "theoretical_loss": 3.4032379856988,
      "tokens_seen": 2442723328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026242728184553663,
      "loss": 1.3704,
      "theoretical_loss": 3.4032307300520657,
      "tokens_seen": 2442788864
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262407221664995,
      "loss": 1.3342,
      "theoretical_loss": 3.403223474654488,
      "tokens_seen": 2442854400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026238716148445335,
      "loss": 1.2058,
      "theoretical_loss": 3.403216219506051,
      "tokens_seen": 2442919936
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 889343,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2826709747314453,
      "objective/train/theoretical_loss": 3.4032089646067414,
      "objective/train/tokens_used": 813444576,
      "theoretical_loss": 3.4032089646067414,
      "tokens_seen": 2442985472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623671013039117,
      "loss": 1.3074,
      "theoretical_loss": 3.4032089646067414,
      "tokens_seen": 2442985472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623470411233701,
      "loss": 1.3213,
      "theoretical_loss": 3.403201709956542,
      "tokens_seen": 2443051008
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623269809428285,
      "loss": 1.4409,
      "theoretical_loss": 3.4031944555554388,
      "tokens_seen": 2443116544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026230692076228686,
      "loss": 1.3941,
      "theoretical_loss": 3.4031872014034166,
      "tokens_seen": 2443182080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622868605817452,
      "loss": 1.2495,
      "theoretical_loss": 3.403179947500459,
      "tokens_seen": 2443247616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622668004012036,
      "loss": 1.3324,
      "theoretical_loss": 3.4031726938465514,
      "tokens_seen": 2443313152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262246740220662,
      "loss": 1.241,
      "theoretical_loss": 3.403165440441679,
      "tokens_seen": 2443378688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026222668004012036,
      "loss": 1.3866,
      "theoretical_loss": 3.4031581872858263,
      "tokens_seen": 2443444224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622066198595787,
      "loss": 1.4343,
      "theoretical_loss": 3.403150934378978,
      "tokens_seen": 2443509760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621865596790371,
      "loss": 1.3575,
      "theoretical_loss": 3.4031436817211187,
      "tokens_seen": 2443575296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026216649949849545,
      "loss": 1.4157,
      "theoretical_loss": 3.4031364293122337,
      "tokens_seen": 2443640832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026214643931795386,
      "loss": 1.3242,
      "theoretical_loss": 3.403129177152307,
      "tokens_seen": 2443706368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621263791374122,
      "loss": 1.3735,
      "theoretical_loss": 3.403121925241324,
      "tokens_seen": 2443771904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621063189568706,
      "loss": 1.3958,
      "theoretical_loss": 3.4031146735792692,
      "tokens_seen": 2443837440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026208625877632895,
      "loss": 1.2829,
      "theoretical_loss": 3.4031074221661273,
      "tokens_seen": 2443902976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026206619859578737,
      "loss": 1.2901,
      "theoretical_loss": 3.4031001710018836,
      "tokens_seen": 2443968512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026204613841524573,
      "loss": 1.3508,
      "theoretical_loss": 3.4030929200865225,
      "tokens_seen": 2444034048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002620260782347041,
      "loss": 1.2354,
      "theoretical_loss": 3.4030856694200287,
      "tokens_seen": 2444099584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002620060180541625,
      "loss": 1.3049,
      "theoretical_loss": 3.4030784190023873,
      "tokens_seen": 2444165120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026198595787362087,
      "loss": 1.3019,
      "theoretical_loss": 3.4030711688335824,
      "tokens_seen": 2444230656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002619658976930793,
      "loss": 1.3713,
      "theoretical_loss": 3.4030639189135994,
      "tokens_seen": 2444296192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026194583751253765,
      "loss": 1.3814,
      "theoretical_loss": 3.4030566692424236,
      "tokens_seen": 2444361728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261925777331996,
      "loss": 1.2917,
      "theoretical_loss": 3.4030494198200385,
      "tokens_seen": 2444427264
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026190571715145437,
      "loss": 1.3146,
      "theoretical_loss": 3.4030421706464296,
      "tokens_seen": 2444492800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618856569709128,
      "loss": 1.2822,
      "theoretical_loss": 3.4030349217215816,
      "tokens_seen": 2444558336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026186559679037115,
      "loss": 1.3221,
      "theoretical_loss": 3.4030276730454796,
      "tokens_seen": 2444623872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618455366098295,
      "loss": 1.4244,
      "theoretical_loss": 3.403020424618108,
      "tokens_seen": 2444689408
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618254764292879,
      "loss": 1.2541,
      "theoretical_loss": 3.403013176439452,
      "tokens_seen": 2444754944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618054162487463,
      "loss": 1.3652,
      "theoretical_loss": 3.4030059285094953,
      "tokens_seen": 2444820480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026178535606820465,
      "loss": 1.3091,
      "theoretical_loss": 3.402998680828224,
      "tokens_seen": 2444886016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261765295887663,
      "loss": 1.2622,
      "theoretical_loss": 3.4029914333956226,
      "tokens_seen": 2444951552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002617452357071214,
      "loss": 1.327,
      "theoretical_loss": 3.402984186211676,
      "tokens_seen": 2445017088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026172517552657974,
      "loss": 1.305,
      "theoretical_loss": 3.4029769392763685,
      "tokens_seen": 2445082624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026170511534603816,
      "loss": 1.34,
      "theoretical_loss": 3.402969692589685,
      "tokens_seen": 2445148160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002616850551654965,
      "loss": 1.4476,
      "theoretical_loss": 3.4029624461516104,
      "tokens_seen": 2445213696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002616649949849549,
      "loss": 1.3878,
      "theoretical_loss": 3.4029551999621295,
      "tokens_seen": 2445279232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026164493480441324,
      "loss": 1.3156,
      "theoretical_loss": 3.4029479540212275,
      "tokens_seen": 2445344768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026162487462387166,
      "loss": 1.5031,
      "theoretical_loss": 3.4029407083288885,
      "tokens_seen": 2445410304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026160481444333,
      "loss": 1.2931,
      "theoretical_loss": 3.402933462885098,
      "tokens_seen": 2445475840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615847542627884,
      "loss": 1.3988,
      "theoretical_loss": 3.4029262176898403,
      "tokens_seen": 2445541376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026156469408224675,
      "loss": 1.2403,
      "theoretical_loss": 3.4029189727431004,
      "tokens_seen": 2445606912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615446339017051,
      "loss": 1.2563,
      "theoretical_loss": 3.402911728044863,
      "tokens_seen": 2445672448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615245737211635,
      "loss": 1.299,
      "theoretical_loss": 3.4029044835951137,
      "tokens_seen": 2445737984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615045135406219,
      "loss": 1.3516,
      "theoretical_loss": 3.402897239393836,
      "tokens_seen": 2445803520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026148445336008025,
      "loss": 1.3638,
      "theoretical_loss": 3.4028899954410154,
      "tokens_seen": 2445869056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614643931795386,
      "loss": 1.3364,
      "theoretical_loss": 3.4028827517366373,
      "tokens_seen": 2445934592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026144433299899703,
      "loss": 1.3023,
      "theoretical_loss": 3.4028755082806854,
      "tokens_seen": 2446000128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614242728184554,
      "loss": 1.3301,
      "theoretical_loss": 3.4028682650731454,
      "tokens_seen": 2446065664
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026140421263791375,
      "loss": 1.224,
      "theoretical_loss": 3.4028610221140014,
      "tokens_seen": 2446131200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613841524573721,
      "loss": 1.258,
      "theoretical_loss": 3.402853779403239,
      "tokens_seen": 2446196736
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 890537,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.372076153755188,
      "objective/train/theoretical_loss": 3.4028465369408423,
      "objective/train/tokens_used": 816721376,
      "theoretical_loss": 3.4028465369408423,
      "tokens_seen": 2446262272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613640922768305,
      "loss": 1.3616,
      "theoretical_loss": 3.4028465369408423,
      "tokens_seen": 2446262272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613440320962889,
      "loss": 1.3643,
      "theoretical_loss": 3.402839294726797,
      "tokens_seen": 2446327808
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026132397191574726,
      "loss": 1.3351,
      "theoretical_loss": 3.4028320527610867,
      "tokens_seen": 2446393344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613039117352056,
      "loss": 1.4195,
      "theoretical_loss": 3.4028248110436974,
      "tokens_seen": 2446458880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261283851554664,
      "loss": 1.286,
      "theoretical_loss": 3.402817569574613,
      "tokens_seen": 2446524416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612637913741224,
      "loss": 1.217,
      "theoretical_loss": 3.4028103283538194,
      "tokens_seen": 2446589952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026124373119358076,
      "loss": 1.2749,
      "theoretical_loss": 3.4028030873813004,
      "tokens_seen": 2446655488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612236710130391,
      "loss": 1.3649,
      "theoretical_loss": 3.4027958466570416,
      "tokens_seen": 2446721024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612036108324975,
      "loss": 1.3661,
      "theoretical_loss": 3.4027886061810273,
      "tokens_seen": 2446786560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026118355065195585,
      "loss": 1.3086,
      "theoretical_loss": 3.402781365953243,
      "tokens_seen": 2446852096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026116349047141426,
      "loss": 1.2858,
      "theoretical_loss": 3.4027741259736723,
      "tokens_seen": 2446917632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002611434302908726,
      "loss": 1.3788,
      "theoretical_loss": 3.4027668862423015,
      "tokens_seen": 2446983168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261123370110331,
      "loss": 1.266,
      "theoretical_loss": 3.4027596467591144,
      "tokens_seen": 2447048704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026110330992978935,
      "loss": 1.2482,
      "theoretical_loss": 3.4027524075240967,
      "tokens_seen": 2447114240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026108324974924777,
      "loss": 1.3155,
      "theoretical_loss": 3.402745168537232,
      "tokens_seen": 2447179776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026106318956870613,
      "loss": 1.2567,
      "theoretical_loss": 3.4027379297985068,
      "tokens_seen": 2447245312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002610431293881645,
      "loss": 1.3292,
      "theoretical_loss": 3.4027306913079043,
      "tokens_seen": 2447310848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026102306920762285,
      "loss": 1.2774,
      "theoretical_loss": 3.4027234530654105,
      "tokens_seen": 2447376384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002610030090270812,
      "loss": 1.2333,
      "theoretical_loss": 3.4027162150710097,
      "tokens_seen": 2447441920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026098294884653963,
      "loss": 1.2712,
      "theoretical_loss": 3.4027089773246875,
      "tokens_seen": 2447507456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260962888665998,
      "loss": 1.2864,
      "theoretical_loss": 3.4027017398264277,
      "tokens_seen": 2447572992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026094282848545636,
      "loss": 1.29,
      "theoretical_loss": 3.4026945025762156,
      "tokens_seen": 2447638528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002609227683049147,
      "loss": 1.4223,
      "theoretical_loss": 3.4026872655740363,
      "tokens_seen": 2447704064
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026090270812437313,
      "loss": 1.2668,
      "theoretical_loss": 3.402680028819874,
      "tokens_seen": 2447769600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608826479438315,
      "loss": 1.4167,
      "theoretical_loss": 3.402672792313715,
      "tokens_seen": 2447835136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026086258776328986,
      "loss": 1.3056,
      "theoretical_loss": 3.4026655560555423,
      "tokens_seen": 2447900672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608425275827482,
      "loss": 1.343,
      "theoretical_loss": 3.402658320045342,
      "tokens_seen": 2447966208
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608224674022066,
      "loss": 1.1956,
      "theoretical_loss": 3.4026510842830984,
      "tokens_seen": 2448031744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260802407221665,
      "loss": 1.3718,
      "theoretical_loss": 3.402643848768797,
      "tokens_seen": 2448097280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026078234704112336,
      "loss": 1.3934,
      "theoretical_loss": 3.4026366135024215,
      "tokens_seen": 2448162816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607622868605817,
      "loss": 1.2082,
      "theoretical_loss": 3.402629378483958,
      "tokens_seen": 2448228352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607422266800401,
      "loss": 1.2699,
      "theoretical_loss": 3.402622143713391,
      "tokens_seen": 2448293888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607221664994985,
      "loss": 1.4717,
      "theoretical_loss": 3.4026149091907048,
      "tokens_seen": 2448359424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026070210631895687,
      "loss": 1.3513,
      "theoretical_loss": 3.402607674915885,
      "tokens_seen": 2448424960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026068204613841523,
      "loss": 1.3045,
      "theoretical_loss": 3.402600440888916,
      "tokens_seen": 2448490496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002606619859578736,
      "loss": 1.4123,
      "theoretical_loss": 3.4025932071097835,
      "tokens_seen": 2448556032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026064192577733195,
      "loss": 1.3672,
      "theoretical_loss": 3.4025859735784714,
      "tokens_seen": 2448621568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026062186559679037,
      "loss": 1.2683,
      "theoretical_loss": 3.4025787402949645,
      "tokens_seen": 2448687104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026060180541624873,
      "loss": 1.3746,
      "theoretical_loss": 3.4025715072592484,
      "tokens_seen": 2448752640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002605817452357071,
      "loss": 1.2209,
      "theoretical_loss": 3.402564274471308,
      "tokens_seen": 2448818176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026056168505516546,
      "loss": 1.3007,
      "theoretical_loss": 3.4025570419311273,
      "tokens_seen": 2448883712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026054162487462387,
      "loss": 1.2537,
      "theoretical_loss": 3.402549809638692,
      "tokens_seen": 2448949248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026052156469408223,
      "loss": 1.4302,
      "theoretical_loss": 3.402542577593987,
      "tokens_seen": 2449014784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026050150451354065,
      "loss": 1.3112,
      "theoretical_loss": 3.4025353457969967,
      "tokens_seen": 2449080320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260481444332999,
      "loss": 1.3001,
      "theoretical_loss": 3.4025281142477057,
      "tokens_seen": 2449145856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026046138415245743,
      "loss": 1.2979,
      "theoretical_loss": 3.4025208829461002,
      "tokens_seen": 2449211392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002604413239719158,
      "loss": 1.4334,
      "theoretical_loss": 3.4025136518921637,
      "tokens_seen": 2449276928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026042126379137415,
      "loss": 1.3403,
      "theoretical_loss": 3.402506421085882,
      "tokens_seen": 2449342464
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002604012036108325,
      "loss": 1.17,
      "theoretical_loss": 3.40249919052724,
      "tokens_seen": 2449408000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603811434302909,
      "loss": 1.1721,
      "theoretical_loss": 3.4024919602162216,
      "tokens_seen": 2449473536
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 891543,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2751030921936035,
      "objective/train/theoretical_loss": 3.4024847301528123,
      "objective/train/tokens_used": 819998176,
      "theoretical_loss": 3.4024847301528123,
      "tokens_seen": 2449539072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603610832497493,
      "loss": 1.2485,
      "theoretical_loss": 3.4024847301528123,
      "tokens_seen": 2449539072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026034102306920766,
      "loss": 1.226,
      "theoretical_loss": 3.4024775003369974,
      "tokens_seen": 2449604608
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260320962888666,
      "loss": 1.2568,
      "theoretical_loss": 3.4024702707687613,
      "tokens_seen": 2449670144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603009027081244,
      "loss": 1.3388,
      "theoretical_loss": 3.4024630414480894,
      "tokens_seen": 2449735680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602808425275828,
      "loss": 1.2031,
      "theoretical_loss": 3.4024558123749657,
      "tokens_seen": 2449801216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026026078234704116,
      "loss": 1.2208,
      "theoretical_loss": 3.402448583549376,
      "tokens_seen": 2449866752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602407221664995,
      "loss": 1.2523,
      "theoretical_loss": 3.4024413549713044,
      "tokens_seen": 2449932288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602206619859579,
      "loss": 1.2126,
      "theoretical_loss": 3.402434126640737,
      "tokens_seen": 2449997824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026020060180541625,
      "loss": 1.2344,
      "theoretical_loss": 3.4024268985576573,
      "tokens_seen": 2450063360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026018054162487466,
      "loss": 1.274,
      "theoretical_loss": 3.4024196707220513,
      "tokens_seen": 2450128896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260160481444333,
      "loss": 1.2667,
      "theoretical_loss": 3.4024124431339033,
      "tokens_seen": 2450194432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002601404212637914,
      "loss": 1.1339,
      "theoretical_loss": 3.4024052157931983,
      "tokens_seen": 2450259968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026012036108324975,
      "loss": 1.2779,
      "theoretical_loss": 3.4023979886999216,
      "tokens_seen": 2450325504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026010030090270817,
      "loss": 1.2613,
      "theoretical_loss": 3.402390761854057,
      "tokens_seen": 2450391040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026008024072216653,
      "loss": 1.127,
      "theoretical_loss": 3.402383535255591,
      "tokens_seen": 2450456576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002600601805416249,
      "loss": 1.3083,
      "theoretical_loss": 3.4023763089045076,
      "tokens_seen": 2450522112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026004012036108325,
      "loss": 1.3698,
      "theoretical_loss": 3.4023690828007918,
      "tokens_seen": 2450587648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002600200601805416,
      "loss": 1.3242,
      "theoretical_loss": 3.4023618569444283,
      "tokens_seen": 2450653184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026000000000000003,
      "loss": 1.2993,
      "theoretical_loss": 3.4023546313354025,
      "tokens_seen": 2450718720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002599799398194584,
      "loss": 1.1712,
      "theoretical_loss": 3.4023474059736993,
      "tokens_seen": 2450784256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025995987963891676,
      "loss": 1.3599,
      "theoretical_loss": 3.4023401808593032,
      "tokens_seen": 2450849792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002599398194583751,
      "loss": 1.3202,
      "theoretical_loss": 3.4023329559921995,
      "tokens_seen": 2450915328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025991975927783353,
      "loss": 1.2785,
      "theoretical_loss": 3.4023257313723727,
      "tokens_seen": 2450980864
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598996990972919,
      "loss": 1.2373,
      "theoretical_loss": 3.402318506999808,
      "tokens_seen": 2451046400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025987963891675026,
      "loss": 1.2973,
      "theoretical_loss": 3.402311282874491,
      "tokens_seen": 2451111936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598595787362086,
      "loss": 1.2893,
      "theoretical_loss": 3.402304058996405,
      "tokens_seen": 2451177472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259839518555667,
      "loss": 1.2794,
      "theoretical_loss": 3.4022968353655365,
      "tokens_seen": 2451243008
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598194583751254,
      "loss": 1.2109,
      "theoretical_loss": 3.4022896119818697,
      "tokens_seen": 2451308544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025979939819458376,
      "loss": 1.2111,
      "theoretical_loss": 3.4022823888453892,
      "tokens_seen": 2451374080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597793380140421,
      "loss": 1.2783,
      "theoretical_loss": 3.402275165956081,
      "tokens_seen": 2451439616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597592778335005,
      "loss": 1.2987,
      "theoretical_loss": 3.4022679433139293,
      "tokens_seen": 2451505152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597392176529589,
      "loss": 1.2217,
      "theoretical_loss": 3.402260720918919,
      "tokens_seen": 2451570688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025971915747241727,
      "loss": 1.2945,
      "theoretical_loss": 3.402253498771035,
      "tokens_seen": 2451636224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025969909729187563,
      "loss": 1.3011,
      "theoretical_loss": 3.4022462768702626,
      "tokens_seen": 2451701760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259679037111334,
      "loss": 1.3744,
      "theoretical_loss": 3.4022390552165866,
      "tokens_seen": 2451767296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025965897693079235,
      "loss": 1.3144,
      "theoretical_loss": 3.4022318338099917,
      "tokens_seen": 2451832832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025963891675025077,
      "loss": 1.0988,
      "theoretical_loss": 3.402224612650463,
      "tokens_seen": 2451898368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025961885656970913,
      "loss": 1.2753,
      "theoretical_loss": 3.402217391737986,
      "tokens_seen": 2451963904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002595987963891675,
      "loss": 1.2413,
      "theoretical_loss": 3.4022101710725448,
      "tokens_seen": 2452029440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025957873620862586,
      "loss": 1.214,
      "theoretical_loss": 3.4022029506541247,
      "tokens_seen": 2452094976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025955867602808427,
      "loss": 1.25,
      "theoretical_loss": 3.4021957304827106,
      "tokens_seen": 2452160512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025953861584754263,
      "loss": 1.1689,
      "theoretical_loss": 3.4021885105582874,
      "tokens_seen": 2452226048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259518555667001,
      "loss": 1.3016,
      "theoretical_loss": 3.4021812908808404,
      "tokens_seen": 2452291584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025949849548645936,
      "loss": 1.2182,
      "theoretical_loss": 3.402174071450354,
      "tokens_seen": 2452357120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594784353059178,
      "loss": 1.35,
      "theoretical_loss": 3.4021668522668134,
      "tokens_seen": 2452422656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025945837512537614,
      "loss": 1.4057,
      "theoretical_loss": 3.4021596333302035,
      "tokens_seen": 2452488192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594383149448345,
      "loss": 1.2432,
      "theoretical_loss": 3.40215241464051,
      "tokens_seen": 2452553728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025941825476429286,
      "loss": 1.2764,
      "theoretical_loss": 3.4021451961977167,
      "tokens_seen": 2452619264
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593981945837512,
      "loss": 1.1841,
      "theoretical_loss": 3.402137978001809,
      "tokens_seen": 2452684800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025937813440320964,
      "loss": 1.2657,
      "theoretical_loss": 3.402130760052772,
      "tokens_seen": 2452750336
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 892743,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4029309749603271,
      "objective/train/theoretical_loss": 3.4021235423505907,
      "objective/train/tokens_used": 823274976,
      "theoretical_loss": 3.4021235423505907,
      "tokens_seen": 2452815872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259358074222668,
      "loss": 1.3487,
      "theoretical_loss": 3.4021235423505907,
      "tokens_seen": 2452815872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025933801404212636,
      "loss": 1.3066,
      "theoretical_loss": 3.4021163248952497,
      "tokens_seen": 2452881408
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593179538615847,
      "loss": 1.251,
      "theoretical_loss": 3.4021091076867345,
      "tokens_seen": 2452946944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025929789368104314,
      "loss": 1.3149,
      "theoretical_loss": 3.4021018907250298,
      "tokens_seen": 2453012480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592778335005015,
      "loss": 1.2697,
      "theoretical_loss": 3.40209467401012,
      "tokens_seen": 2453078016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025925777331995987,
      "loss": 1.3618,
      "theoretical_loss": 3.402087457541991,
      "tokens_seen": 2453143552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025923771313941823,
      "loss": 1.3523,
      "theoretical_loss": 3.4020802413206273,
      "tokens_seen": 2453209088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592176529588766,
      "loss": 1.312,
      "theoretical_loss": 3.402073025346014,
      "tokens_seen": 2453274624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259197592778335,
      "loss": 1.2164,
      "theoretical_loss": 3.402065809618136,
      "tokens_seen": 2453340160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025917753259779337,
      "loss": 1.2917,
      "theoretical_loss": 3.402058594136978,
      "tokens_seen": 2453405696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025915747241725173,
      "loss": 1.1973,
      "theoretical_loss": 3.4020513789025255,
      "tokens_seen": 2453471232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002591374122367101,
      "loss": 1.3843,
      "theoretical_loss": 3.402044163914763,
      "tokens_seen": 2453536768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002591173520561685,
      "loss": 1.0975,
      "theoretical_loss": 3.4020369491736764,
      "tokens_seen": 2453602304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590972918756269,
      "loss": 1.3724,
      "theoretical_loss": 3.4020297346792496,
      "tokens_seen": 2453667840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025907723169508524,
      "loss": 1.3255,
      "theoretical_loss": 3.4020225204314674,
      "tokens_seen": 2453733376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590571715145436,
      "loss": 1.2759,
      "theoretical_loss": 3.4020153064303162,
      "tokens_seen": 2453798912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025903711133400196,
      "loss": 1.1484,
      "theoretical_loss": 3.4020080926757794,
      "tokens_seen": 2453864448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590170511534604,
      "loss": 1.2668,
      "theoretical_loss": 3.402000879167843,
      "tokens_seen": 2453929984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002589969909729188,
      "loss": 1.2611,
      "theoretical_loss": 3.401993665906492,
      "tokens_seen": 2453995520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025897693079237716,
      "loss": 1.335,
      "theoretical_loss": 3.4019864528917108,
      "tokens_seen": 2454061056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002589568706118355,
      "loss": 1.2145,
      "theoretical_loss": 3.401979240123485,
      "tokens_seen": 2454126592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025893681043129393,
      "loss": 1.3004,
      "theoretical_loss": 3.401972027601799,
      "tokens_seen": 2454192128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002589167502507523,
      "loss": 1.4319,
      "theoretical_loss": 3.4019648153266377,
      "tokens_seen": 2454257664
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025889669007021066,
      "loss": 1.1735,
      "theoretical_loss": 3.401957603297987,
      "tokens_seen": 2454323200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258876629889669,
      "loss": 1.2995,
      "theoretical_loss": 3.401950391515831,
      "tokens_seen": 2454388736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002588565697091274,
      "loss": 1.3106,
      "theoretical_loss": 3.4019431799801554,
      "tokens_seen": 2454454272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002588365095285858,
      "loss": 1.2075,
      "theoretical_loss": 3.4019359686909443,
      "tokens_seen": 2454519808
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025881644934804416,
      "loss": 1.3379,
      "theoretical_loss": 3.4019287576481836,
      "tokens_seen": 2454585344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587963891675025,
      "loss": 1.3133,
      "theoretical_loss": 3.4019215468518578,
      "tokens_seen": 2454650880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587763289869609,
      "loss": 1.2299,
      "theoretical_loss": 3.4019143363019517,
      "tokens_seen": 2454716416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587562688064193,
      "loss": 1.4027,
      "theoretical_loss": 3.401907125998451,
      "tokens_seen": 2454781952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025873620862587767,
      "loss": 1.3717,
      "theoretical_loss": 3.4018999159413403,
      "tokens_seen": 2454847488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025871614844533603,
      "loss": 1.1951,
      "theoretical_loss": 3.4018927061306043,
      "tokens_seen": 2454913024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586960882647944,
      "loss": 1.316,
      "theoretical_loss": 3.4018854965662286,
      "tokens_seen": 2454978560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025867602808425275,
      "loss": 1.187,
      "theoretical_loss": 3.401878287248198,
      "tokens_seen": 2455044096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025865596790371117,
      "loss": 1.3822,
      "theoretical_loss": 3.401871078176497,
      "tokens_seen": 2455109632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025863590772316953,
      "loss": 1.3052,
      "theoretical_loss": 3.4018638693511116,
      "tokens_seen": 2455175168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586158475426279,
      "loss": 1.4023,
      "theoretical_loss": 3.401856660772026,
      "tokens_seen": 2455240704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025859578736208626,
      "loss": 1.2247,
      "theoretical_loss": 3.4018494524392255,
      "tokens_seen": 2455306240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025857572718154467,
      "loss": 1.2596,
      "theoretical_loss": 3.401842244352695,
      "tokens_seen": 2455371776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025855566700100303,
      "loss": 1.3451,
      "theoretical_loss": 3.40183503651242,
      "tokens_seen": 2455437312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002585356068204614,
      "loss": 1.3343,
      "theoretical_loss": 3.4018278289183845,
      "tokens_seen": 2455502848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025851554663991976,
      "loss": 1.2938,
      "theoretical_loss": 3.4018206215705744,
      "tokens_seen": 2455568384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584954864593782,
      "loss": 1.2978,
      "theoretical_loss": 3.4018134144689745,
      "tokens_seen": 2455633920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025847542627883654,
      "loss": 1.2582,
      "theoretical_loss": 3.4018062076135696,
      "tokens_seen": 2455699456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584553660982949,
      "loss": 1.1933,
      "theoretical_loss": 3.4017990010043446,
      "tokens_seen": 2455764992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025843530591775326,
      "loss": 1.3183,
      "theoretical_loss": 3.4017917946412854,
      "tokens_seen": 2455830528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584152457372116,
      "loss": 1.2937,
      "theoretical_loss": 3.401784588524376,
      "tokens_seen": 2455896064
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025839518555667004,
      "loss": 1.2139,
      "theoretical_loss": 3.401777382653602,
      "tokens_seen": 2455961600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002583751253761284,
      "loss": 1.2555,
      "theoretical_loss": 3.4017701770289483,
      "tokens_seen": 2456027136
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 893908,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1507830619812012,
      "objective/train/theoretical_loss": 3.4017629716504,
      "objective/train/tokens_used": 826551776,
      "theoretical_loss": 3.4017629716504,
      "tokens_seen": 2456092672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025835506519558677,
      "loss": 1.1521,
      "theoretical_loss": 3.4017629716504,
      "tokens_seen": 2456092672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025833500501504513,
      "loss": 1.3059,
      "theoretical_loss": 3.4017557665179416,
      "tokens_seen": 2456158208
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025831494483450354,
      "loss": 1.2003,
      "theoretical_loss": 3.401748561631559,
      "tokens_seen": 2456223744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002582948846539619,
      "loss": 1.3157,
      "theoretical_loss": 3.4017413569912365,
      "tokens_seen": 2456289280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025827482447342027,
      "loss": 1.1688,
      "theoretical_loss": 3.4017341525969598,
      "tokens_seen": 2456354816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025825476429287863,
      "loss": 1.331,
      "theoretical_loss": 3.401726948448713,
      "tokens_seen": 2456420352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258234704112337,
      "loss": 1.2889,
      "theoretical_loss": 3.401719744546482,
      "tokens_seen": 2456485888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002582146439317954,
      "loss": 1.3216,
      "theoretical_loss": 3.4017125408902515,
      "tokens_seen": 2456551424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025819458375125377,
      "loss": 1.3415,
      "theoretical_loss": 3.4017053374800064,
      "tokens_seen": 2456616960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025817452357071213,
      "loss": 1.3383,
      "theoretical_loss": 3.4016981343157324,
      "tokens_seen": 2456682496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581544633901705,
      "loss": 1.1546,
      "theoretical_loss": 3.4016909313974133,
      "tokens_seen": 2456748032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581344032096289,
      "loss": 1.3471,
      "theoretical_loss": 3.401683728725035,
      "tokens_seen": 2456813568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581143430290873,
      "loss": 1.1808,
      "theoretical_loss": 3.401676526298583,
      "tokens_seen": 2456879104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025809428284854564,
      "loss": 1.2624,
      "theoretical_loss": 3.4016693241180413,
      "tokens_seen": 2456944640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258074222668004,
      "loss": 1.3343,
      "theoretical_loss": 3.401662122183396,
      "tokens_seen": 2457010176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025805416248746236,
      "loss": 1.2565,
      "theoretical_loss": 3.4016549204946305,
      "tokens_seen": 2457075712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002580341023069208,
      "loss": 1.2243,
      "theoretical_loss": 3.4016477190517316,
      "tokens_seen": 2457141248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025801404212637914,
      "loss": 1.3084,
      "theoretical_loss": 3.4016405178546836,
      "tokens_seen": 2457206784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002579939819458375,
      "loss": 1.326,
      "theoretical_loss": 3.4016333169034714,
      "tokens_seen": 2457272320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025797392176529586,
      "loss": 1.3367,
      "theoretical_loss": 3.4016261161980803,
      "tokens_seen": 2457337856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002579538615847543,
      "loss": 1.2767,
      "theoretical_loss": 3.4016189157384953,
      "tokens_seen": 2457403392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025793380140421264,
      "loss": 1.394,
      "theoretical_loss": 3.4016117155247017,
      "tokens_seen": 2457468928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257913741223671,
      "loss": 1.3321,
      "theoretical_loss": 3.4016045155566843,
      "tokens_seen": 2457534464
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025789368104312937,
      "loss": 1.2814,
      "theoretical_loss": 3.4015973158344277,
      "tokens_seen": 2457600000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025787362086258773,
      "loss": 1.3176,
      "theoretical_loss": 3.4015901163579176,
      "tokens_seen": 2457665536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025785356068204615,
      "loss": 1.1297,
      "theoretical_loss": 3.401582917127139,
      "tokens_seen": 2457731072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002578335005015045,
      "loss": 1.3789,
      "theoretical_loss": 3.401575718142077,
      "tokens_seen": 2457796608
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025781344032096287,
      "loss": 1.21,
      "theoretical_loss": 3.4015685194027165,
      "tokens_seen": 2457862144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025779338014042123,
      "loss": 1.3223,
      "theoretical_loss": 3.4015613209090425,
      "tokens_seen": 2457927680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025777331995987965,
      "loss": 1.417,
      "theoretical_loss": 3.40155412266104,
      "tokens_seen": 2457993216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257753259779338,
      "loss": 1.082,
      "theoretical_loss": 3.4015469246586942,
      "tokens_seen": 2458058752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002577331995987964,
      "loss": 1.2411,
      "theoretical_loss": 3.4015397269019902,
      "tokens_seen": 2458124288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025771313941825474,
      "loss": 1.2932,
      "theoretical_loss": 3.401532529390913,
      "tokens_seen": 2458189824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576930792377131,
      "loss": 1.1867,
      "theoretical_loss": 3.401525332125448,
      "tokens_seen": 2458255360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576730190571715,
      "loss": 1.3752,
      "theoretical_loss": 3.4015181351055794,
      "tokens_seen": 2458320896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576529588766299,
      "loss": 1.2384,
      "theoretical_loss": 3.4015109383312936,
      "tokens_seen": 2458386432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025763289869608824,
      "loss": 1.2866,
      "theoretical_loss": 3.4015037418025744,
      "tokens_seen": 2458451968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576128385155466,
      "loss": 1.3302,
      "theoretical_loss": 3.4014965455194073,
      "tokens_seen": 2458517504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257592778335005,
      "loss": 1.2366,
      "theoretical_loss": 3.401489349481778,
      "tokens_seen": 2458583040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575727181544634,
      "loss": 1.2519,
      "theoretical_loss": 3.4014821536896704,
      "tokens_seen": 2458648576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025755265797392174,
      "loss": 1.2704,
      "theoretical_loss": 3.401474958143071,
      "tokens_seen": 2458714112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575325977933801,
      "loss": 1.2832,
      "theoretical_loss": 3.401467762841963,
      "tokens_seen": 2458779648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025751253761283847,
      "loss": 1.2439,
      "theoretical_loss": 3.4014605677863337,
      "tokens_seen": 2458845184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025749247743229694,
      "loss": 1.2907,
      "theoretical_loss": 3.4014533729761665,
      "tokens_seen": 2458910720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002574724172517553,
      "loss": 1.2892,
      "theoretical_loss": 3.4014461784114474,
      "tokens_seen": 2458976256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025745235707121366,
      "loss": 1.3071,
      "theoretical_loss": 3.401438984092161,
      "tokens_seen": 2459041792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257432296890672,
      "loss": 1.3161,
      "theoretical_loss": 3.401431790018292,
      "tokens_seen": 2459107328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025741223671013044,
      "loss": 1.3173,
      "theoretical_loss": 3.401424596189827,
      "tokens_seen": 2459172864
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002573921765295888,
      "loss": 1.3224,
      "theoretical_loss": 3.4014174026067496,
      "tokens_seen": 2459238400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025737211634904717,
      "loss": 1.2555,
      "theoretical_loss": 3.401410209269045,
      "tokens_seen": 2459303936
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 895121,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.0762553215026855,
      "objective/train/theoretical_loss": 3.4014030161766993,
      "objective/train/tokens_used": 829828576,
      "theoretical_loss": 3.4014030161766993,
      "tokens_seen": 2459369472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025735205616850553,
      "loss": 1.1005,
      "theoretical_loss": 3.4014030161766993,
      "tokens_seen": 2459369472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025733199598796394,
      "loss": 1.1974,
      "theoretical_loss": 3.401395823329697,
      "tokens_seen": 2459435008
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002573119358074223,
      "loss": 1.3607,
      "theoretical_loss": 3.401388630728023,
      "tokens_seen": 2459500544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025729187562688067,
      "loss": 1.2499,
      "theoretical_loss": 3.4013814383716623,
      "tokens_seen": 2459566080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025727181544633903,
      "loss": 1.2714,
      "theoretical_loss": 3.401374246260601,
      "tokens_seen": 2459631616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572517552657974,
      "loss": 1.2088,
      "theoretical_loss": 3.4013670543948225,
      "tokens_seen": 2459697152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572316950852558,
      "loss": 1.2267,
      "theoretical_loss": 3.4013598627743136,
      "tokens_seen": 2459762688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025721163490471417,
      "loss": 1.2654,
      "theoretical_loss": 3.4013526713990583,
      "tokens_seen": 2459828224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025719157472417253,
      "loss": 1.2365,
      "theoretical_loss": 3.4013454802690424,
      "tokens_seen": 2459893760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571715145436309,
      "loss": 1.2702,
      "theoretical_loss": 3.4013382893842503,
      "tokens_seen": 2459959296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571514543630893,
      "loss": 1.2643,
      "theoretical_loss": 3.401331098744668,
      "tokens_seen": 2460024832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571313941825477,
      "loss": 1.3124,
      "theoretical_loss": 3.4013239083502795,
      "tokens_seen": 2460090368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025711133400200604,
      "loss": 1.3635,
      "theoretical_loss": 3.4013167182010706,
      "tokens_seen": 2460155904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570912738214644,
      "loss": 1.1712,
      "theoretical_loss": 3.4013095282970265,
      "tokens_seen": 2460221440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025707121364092276,
      "loss": 1.225,
      "theoretical_loss": 3.401302338638132,
      "tokens_seen": 2460286976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570511534603812,
      "loss": 1.1791,
      "theoretical_loss": 3.4012951492243726,
      "tokens_seen": 2460352512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025703109327983954,
      "loss": 1.245,
      "theoretical_loss": 3.4012879600557326,
      "tokens_seen": 2460418048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570110330992979,
      "loss": 1.3002,
      "theoretical_loss": 3.4012807711321984,
      "tokens_seen": 2460483584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025699097291875626,
      "loss": 1.3345,
      "theoretical_loss": 3.4012735824537534,
      "tokens_seen": 2460549120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002569709127382147,
      "loss": 1.266,
      "theoretical_loss": 3.401266394020385,
      "tokens_seen": 2460614656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025695085255767304,
      "loss": 1.1927,
      "theoretical_loss": 3.401259205832076,
      "tokens_seen": 2460680192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002569307923771314,
      "loss": 1.2874,
      "theoretical_loss": 3.4012520178888126,
      "tokens_seen": 2460745728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025691073219658977,
      "loss": 1.2789,
      "theoretical_loss": 3.40124483019058,
      "tokens_seen": 2460811264
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025689067201604813,
      "loss": 1.3014,
      "theoretical_loss": 3.401237642737363,
      "tokens_seen": 2460876800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025687061183550655,
      "loss": 1.237,
      "theoretical_loss": 3.4012304555291473,
      "tokens_seen": 2460942336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002568505516549649,
      "loss": 1.3844,
      "theoretical_loss": 3.4012232685659174,
      "tokens_seen": 2461007872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025683049147442327,
      "loss": 1.2603,
      "theoretical_loss": 3.4012160818476587,
      "tokens_seen": 2461073408
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025681043129388163,
      "loss": 1.415,
      "theoretical_loss": 3.401208895374356,
      "tokens_seen": 2461138944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025679037111334005,
      "loss": 1.2863,
      "theoretical_loss": 3.4012017091459947,
      "tokens_seen": 2461204480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567703109327984,
      "loss": 1.2419,
      "theoretical_loss": 3.40119452316256,
      "tokens_seen": 2461270016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567502507522568,
      "loss": 1.2632,
      "theoretical_loss": 3.4011873374240373,
      "tokens_seen": 2461335552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025673019057171514,
      "loss": 1.2436,
      "theoretical_loss": 3.401180151930411,
      "tokens_seen": 2461401088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567101303911735,
      "loss": 1.2178,
      "theoretical_loss": 3.4011729666816666,
      "tokens_seen": 2461466624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566900702106319,
      "loss": 1.3521,
      "theoretical_loss": 3.4011657816777894,
      "tokens_seen": 2461532160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566700100300903,
      "loss": 1.2555,
      "theoretical_loss": 3.4011585969187648,
      "tokens_seen": 2461597696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025664994984954864,
      "loss": 1.1631,
      "theoretical_loss": 3.4011514124045767,
      "tokens_seen": 2461663232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256629889669007,
      "loss": 1.3963,
      "theoretical_loss": 3.4011442281352116,
      "tokens_seen": 2461728768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566098294884654,
      "loss": 1.2355,
      "theoretical_loss": 3.401137044110654,
      "tokens_seen": 2461794304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565897693079238,
      "loss": 1.2411,
      "theoretical_loss": 3.4011298603308893,
      "tokens_seen": 2461859840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025656970912738214,
      "loss": 1.2878,
      "theoretical_loss": 3.401122676795902,
      "tokens_seen": 2461925376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565496489468405,
      "loss": 1.2499,
      "theoretical_loss": 3.401115493505678,
      "tokens_seen": 2461990912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025652958876629887,
      "loss": 1.382,
      "theoretical_loss": 3.4011083104602022,
      "tokens_seen": 2462056448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565095285857573,
      "loss": 1.2319,
      "theoretical_loss": 3.40110112765946,
      "tokens_seen": 2462121984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025648946840521565,
      "loss": 1.2203,
      "theoretical_loss": 3.401093945103436,
      "tokens_seen": 2462187520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256469408224674,
      "loss": 1.2867,
      "theoretical_loss": 3.4010867627921155,
      "tokens_seen": 2462253056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025644934804413237,
      "loss": 1.2172,
      "theoretical_loss": 3.4010795807254843,
      "tokens_seen": 2462318592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002564292878635908,
      "loss": 1.3653,
      "theoretical_loss": 3.4010723989035263,
      "tokens_seen": 2462384128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025640922768304915,
      "loss": 1.3665,
      "theoretical_loss": 3.401065217326228,
      "tokens_seen": 2462449664
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002563891675025075,
      "loss": 1.2652,
      "theoretical_loss": 3.4010580359935734,
      "tokens_seen": 2462515200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002563691073219659,
      "loss": 1.1988,
      "theoretical_loss": 3.4010508549055487,
      "tokens_seen": 2462580736
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 896263,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3929803371429443,
      "objective/train/theoretical_loss": 3.401043674062138,
      "objective/train/tokens_used": 833105376,
      "theoretical_loss": 3.401043674062138,
      "tokens_seen": 2462646272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025634904714142424,
      "loss": 1.3371,
      "theoretical_loss": 3.401043674062138,
      "tokens_seen": 2462646272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025632898696088265,
      "loss": 1.1339,
      "theoretical_loss": 3.4010364934633275,
      "tokens_seen": 2462711808
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256308926780341,
      "loss": 1.267,
      "theoretical_loss": 3.4010293131091016,
      "tokens_seen": 2462777344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562888665997994,
      "loss": 1.3505,
      "theoretical_loss": 3.401022132999446,
      "tokens_seen": 2462842880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025626880641925774,
      "loss": 1.292,
      "theoretical_loss": 3.401014953134345,
      "tokens_seen": 2462908416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025624874623871616,
      "loss": 1.2029,
      "theoretical_loss": 3.401007773513785,
      "tokens_seen": 2462973952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562286860581745,
      "loss": 1.1188,
      "theoretical_loss": 3.4010005941377504,
      "tokens_seen": 2463039488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562086258776329,
      "loss": 1.3494,
      "theoretical_loss": 3.400993415006226,
      "tokens_seen": 2463105024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025618856569709124,
      "loss": 1.3765,
      "theoretical_loss": 3.4009862361191976,
      "tokens_seen": 2463170560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025616850551654966,
      "loss": 1.2538,
      "theoretical_loss": 3.4009790574766505,
      "tokens_seen": 2463236096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256148445336008,
      "loss": 1.3168,
      "theoretical_loss": 3.4009718790785692,
      "tokens_seen": 2463301632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002561283851554664,
      "loss": 1.2821,
      "theoretical_loss": 3.40096470092494,
      "tokens_seen": 2463367168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025610832497492475,
      "loss": 1.1803,
      "theoretical_loss": 3.4009575230157467,
      "tokens_seen": 2463432704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560882647943831,
      "loss": 1.2587,
      "theoretical_loss": 3.4009503453509753,
      "tokens_seen": 2463498240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560682046138415,
      "loss": 1.2135,
      "theoretical_loss": 3.4009431679306106,
      "tokens_seen": 2463563776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560481444332999,
      "loss": 1.2501,
      "theoretical_loss": 3.400935990754638,
      "tokens_seen": 2463629312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025602808425275825,
      "loss": 1.3872,
      "theoretical_loss": 3.4009288138230427,
      "tokens_seen": 2463694848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560080240722166,
      "loss": 1.288,
      "theoretical_loss": 3.4009216371358097,
      "tokens_seen": 2463760384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002559879638916751,
      "loss": 1.1709,
      "theoretical_loss": 3.4009144606929245,
      "tokens_seen": 2463825920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025596790371113344,
      "loss": 1.2113,
      "theoretical_loss": 3.400907284494372,
      "tokens_seen": 2463891456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002559478435305918,
      "loss": 1.2224,
      "theoretical_loss": 3.4009001085401374,
      "tokens_seen": 2463956992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025592778335005017,
      "loss": 1.2911,
      "theoretical_loss": 3.4008929328302058,
      "tokens_seen": 2464022528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025590772316950853,
      "loss": 1.3853,
      "theoretical_loss": 3.400885757364563,
      "tokens_seen": 2464088064
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025588766298896695,
      "loss": 1.2285,
      "theoretical_loss": 3.4008785821431933,
      "tokens_seen": 2464153600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002558676028084253,
      "loss": 1.2755,
      "theoretical_loss": 3.400871407166082,
      "tokens_seen": 2464219136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025584754262788367,
      "loss": 1.2036,
      "theoretical_loss": 3.4008642324332152,
      "tokens_seen": 2464284672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025582748244734203,
      "loss": 1.3086,
      "theoretical_loss": 3.4008570579445774,
      "tokens_seen": 2464350208
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025580742226680045,
      "loss": 1.2557,
      "theoretical_loss": 3.4008498837001535,
      "tokens_seen": 2464415744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557873620862588,
      "loss": 1.3022,
      "theoretical_loss": 3.4008427096999294,
      "tokens_seen": 2464481280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557673019057172,
      "loss": 1.2379,
      "theoretical_loss": 3.40083553594389,
      "tokens_seen": 2464546816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025574724172517554,
      "loss": 1.3034,
      "theoretical_loss": 3.40082836243202,
      "tokens_seen": 2464612352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557271815446339,
      "loss": 1.2848,
      "theoretical_loss": 3.400821189164305,
      "tokens_seen": 2464677888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557071213640923,
      "loss": 1.2275,
      "theoretical_loss": 3.400814016140731,
      "tokens_seen": 2464743424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556870611835507,
      "loss": 1.2393,
      "theoretical_loss": 3.4008068433612824,
      "tokens_seen": 2464808960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025566700100300904,
      "loss": 1.2202,
      "theoretical_loss": 3.400799670825944,
      "tokens_seen": 2464874496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556469408224674,
      "loss": 1.3058,
      "theoretical_loss": 3.4007924985347016,
      "tokens_seen": 2464940032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556268806419258,
      "loss": 1.3109,
      "theoretical_loss": 3.40078532648754,
      "tokens_seen": 2465005568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556068204613842,
      "loss": 1.2705,
      "theoretical_loss": 3.4007781546844447,
      "tokens_seen": 2465071104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025558676028084254,
      "loss": 1.2203,
      "theoretical_loss": 3.400770983125401,
      "tokens_seen": 2465136640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002555667001003009,
      "loss": 1.4566,
      "theoretical_loss": 3.400763811810394,
      "tokens_seen": 2465202176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025554663991975927,
      "loss": 1.1912,
      "theoretical_loss": 3.400756640739409,
      "tokens_seen": 2465267712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002555265797392177,
      "loss": 1.3859,
      "theoretical_loss": 3.400749469912431,
      "tokens_seen": 2465333248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025550651955867605,
      "loss": 1.2354,
      "theoretical_loss": 3.400742299329445,
      "tokens_seen": 2465398784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554864593781344,
      "loss": 1.2108,
      "theoretical_loss": 3.400735128990437,
      "tokens_seen": 2465464320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025546639919759277,
      "loss": 1.1971,
      "theoretical_loss": 3.4007279588953914,
      "tokens_seen": 2465529856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554463390170512,
      "loss": 1.3361,
      "theoretical_loss": 3.4007207890442936,
      "tokens_seen": 2465595392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025542627883650955,
      "loss": 1.2921,
      "theoretical_loss": 3.4007136194371292,
      "tokens_seen": 2465660928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554062186559679,
      "loss": 1.3289,
      "theoretical_loss": 3.4007064500738835,
      "tokens_seen": 2465726464
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553861584754263,
      "loss": 1.288,
      "theoretical_loss": 3.400699280954541,
      "tokens_seen": 2465792000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025536609829488464,
      "loss": 1.2986,
      "theoretical_loss": 3.400692112079087,
      "tokens_seen": 2465857536
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 897133,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2484326362609863,
      "objective/train/theoretical_loss": 3.4006849434475077,
      "objective/train/tokens_used": 836382176,
      "theoretical_loss": 3.4006849434475077,
      "tokens_seen": 2465923072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025534603811434305,
      "loss": 1.2703,
      "theoretical_loss": 3.4006849434475077,
      "tokens_seen": 2465923072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553259779338014,
      "loss": 1.2842,
      "theoretical_loss": 3.4006777750597874,
      "tokens_seen": 2465988608
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553059177532598,
      "loss": 1.3129,
      "theoretical_loss": 3.4006706069159116,
      "tokens_seen": 2466054144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025528585757271814,
      "loss": 1.3345,
      "theoretical_loss": 3.400663439015865,
      "tokens_seen": 2466119680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025526579739217656,
      "loss": 1.245,
      "theoretical_loss": 3.400656271359634,
      "tokens_seen": 2466185216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002552457372116349,
      "loss": 1.3945,
      "theoretical_loss": 3.4006491039472033,
      "tokens_seen": 2466250752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002552256770310933,
      "loss": 1.3552,
      "theoretical_loss": 3.4006419367785576,
      "tokens_seen": 2466316288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025520561685055164,
      "loss": 1.1685,
      "theoretical_loss": 3.4006347698536823,
      "tokens_seen": 2466381824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025518555667001006,
      "loss": 1.2597,
      "theoretical_loss": 3.4006276031725635,
      "tokens_seen": 2466447360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002551654964894684,
      "loss": 1.28,
      "theoretical_loss": 3.4006204367351853,
      "tokens_seen": 2466512896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002551454363089268,
      "loss": 1.2678,
      "theoretical_loss": 3.4006132705415335,
      "tokens_seen": 2466578432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025512537612838515,
      "loss": 1.2468,
      "theoretical_loss": 3.4006061045915934,
      "tokens_seen": 2466643968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002551053159478435,
      "loss": 1.2012,
      "theoretical_loss": 3.40059893888535,
      "tokens_seen": 2466709504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002550852557673019,
      "loss": 1.2066,
      "theoretical_loss": 3.4005917734227884,
      "tokens_seen": 2466775040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002550651955867603,
      "loss": 1.2797,
      "theoretical_loss": 3.4005846082038946,
      "tokens_seen": 2466840576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025504513540621865,
      "loss": 1.3193,
      "theoretical_loss": 3.400577443228653,
      "tokens_seen": 2466906112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000255025075225677,
      "loss": 1.427,
      "theoretical_loss": 3.400570278497049,
      "tokens_seen": 2466971648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025500501504513543,
      "loss": 1.2894,
      "theoretical_loss": 3.400563114009068,
      "tokens_seen": 2467037184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549849548645938,
      "loss": 1.3065,
      "theoretical_loss": 3.4005559497646956,
      "tokens_seen": 2467102720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025496489468405215,
      "loss": 1.2431,
      "theoretical_loss": 3.4005487857639167,
      "tokens_seen": 2467168256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549448345035105,
      "loss": 1.351,
      "theoretical_loss": 3.400541622006716,
      "tokens_seen": 2467233792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549247743229689,
      "loss": 1.3369,
      "theoretical_loss": 3.40053445849308,
      "tokens_seen": 2467299328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549047141424273,
      "loss": 1.1883,
      "theoretical_loss": 3.4005272952229926,
      "tokens_seen": 2467364864
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025488465396188566,
      "loss": 1.2606,
      "theoretical_loss": 3.40052013219644,
      "tokens_seen": 2467430400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254864593781344,
      "loss": 1.1122,
      "theoretical_loss": 3.4005129694134073,
      "tokens_seen": 2467495936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548445336008024,
      "loss": 1.2738,
      "theoretical_loss": 3.400505806873879,
      "tokens_seen": 2467561472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548244734202608,
      "loss": 1.3744,
      "theoretical_loss": 3.4004986445778416,
      "tokens_seen": 2467627008
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025480441323971916,
      "loss": 1.265,
      "theoretical_loss": 3.400491482525279,
      "tokens_seen": 2467692544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002547843530591775,
      "loss": 1.2596,
      "theoretical_loss": 3.4004843207161777,
      "tokens_seen": 2467758080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002547642928786359,
      "loss": 1.3174,
      "theoretical_loss": 3.4004771591505225,
      "tokens_seen": 2467823616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025474423269809425,
      "loss": 1.2611,
      "theoretical_loss": 3.4004699978282984,
      "tokens_seen": 2467889152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025472417251755266,
      "loss": 1.4096,
      "theoretical_loss": 3.4004628367494907,
      "tokens_seen": 2467954688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254704112337011,
      "loss": 1.3379,
      "theoretical_loss": 3.400455675914085,
      "tokens_seen": 2468020224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546840521564694,
      "loss": 1.2788,
      "theoretical_loss": 3.4004485153220663,
      "tokens_seen": 2468085760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025466399197592775,
      "loss": 1.2646,
      "theoretical_loss": 3.40044135497342,
      "tokens_seen": 2468151296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025464393179538616,
      "loss": 1.3432,
      "theoretical_loss": 3.400434194868131,
      "tokens_seen": 2468216832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025462387161484453,
      "loss": 1.2121,
      "theoretical_loss": 3.400427035006185,
      "tokens_seen": 2468282368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546038114343029,
      "loss": 1.3065,
      "theoretical_loss": 3.400419875387567,
      "tokens_seen": 2468347904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025458375125376125,
      "loss": 1.3534,
      "theoretical_loss": 3.400412716012263,
      "tokens_seen": 2468413440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545636910732196,
      "loss": 1.299,
      "theoretical_loss": 3.400405556880257,
      "tokens_seen": 2468478976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025454363089267803,
      "loss": 1.2512,
      "theoretical_loss": 3.4003983979915358,
      "tokens_seen": 2468544512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545235707121364,
      "loss": 1.3126,
      "theoretical_loss": 3.4003912393460833,
      "tokens_seen": 2468610048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545035105315948,
      "loss": 1.2881,
      "theoretical_loss": 3.4003840809438852,
      "tokens_seen": 2468675584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025448345035105317,
      "loss": 1.3258,
      "theoretical_loss": 3.4003769227849268,
      "tokens_seen": 2468741120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544633901705116,
      "loss": 1.3368,
      "theoretical_loss": 3.4003697648691937,
      "tokens_seen": 2468806656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025444332998996995,
      "loss": 1.2432,
      "theoretical_loss": 3.4003626071966706,
      "tokens_seen": 2468872192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544232698094283,
      "loss": 1.3649,
      "theoretical_loss": 3.4003554497673436,
      "tokens_seen": 2468937728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544032096288867,
      "loss": 1.2638,
      "theoretical_loss": 3.400348292581197,
      "tokens_seen": 2469003264
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025438314944834504,
      "loss": 1.1754,
      "theoretical_loss": 3.400341135638217,
      "tokens_seen": 2469068800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025436308926780345,
      "loss": 1.3857,
      "theoretical_loss": 3.400333978938388,
      "tokens_seen": 2469134336
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 898257,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.171244502067566,
      "objective/train/theoretical_loss": 3.4003268224816963,
      "objective/train/tokens_used": 839658976,
      "theoretical_loss": 3.4003268224816963,
      "tokens_seen": 2469199872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002543430290872618,
      "loss": 1.2643,
      "theoretical_loss": 3.4003268224816963,
      "tokens_seen": 2469199872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002543229689067202,
      "loss": 1.2239,
      "theoretical_loss": 3.4003196662681265,
      "tokens_seen": 2469265408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025430290872617854,
      "loss": 1.247,
      "theoretical_loss": 3.4003125102976637,
      "tokens_seen": 2469330944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025428284854563696,
      "loss": 1.22,
      "theoretical_loss": 3.400305354570294,
      "tokens_seen": 2469396480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542627883650953,
      "loss": 1.3195,
      "theoretical_loss": 3.4002981990860017,
      "tokens_seen": 2469462016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542427281845537,
      "loss": 1.305,
      "theoretical_loss": 3.400291043844773,
      "tokens_seen": 2469527552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025422266800401204,
      "loss": 1.3304,
      "theoretical_loss": 3.4002838888465927,
      "tokens_seen": 2469593088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025420260782347046,
      "loss": 1.3272,
      "theoretical_loss": 3.4002767340914457,
      "tokens_seen": 2469658624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541825476429288,
      "loss": 1.3424,
      "theoretical_loss": 3.4002695795793185,
      "tokens_seen": 2469724160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541624874623872,
      "loss": 1.1842,
      "theoretical_loss": 3.400262425310195,
      "tokens_seen": 2469789696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025414242728184555,
      "loss": 1.2724,
      "theoretical_loss": 3.4002552712840615,
      "tokens_seen": 2469855232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541223671013039,
      "loss": 1.2416,
      "theoretical_loss": 3.400248117500903,
      "tokens_seen": 2469920768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541023069207623,
      "loss": 1.363,
      "theoretical_loss": 3.400240963960705,
      "tokens_seen": 2469986304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002540822467402207,
      "loss": 1.3486,
      "theoretical_loss": 3.4002338106634524,
      "tokens_seen": 2470051840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025406218655967905,
      "loss": 1.447,
      "theoretical_loss": 3.4002266576091307,
      "tokens_seen": 2470117376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002540421263791374,
      "loss": 1.3055,
      "theoretical_loss": 3.4002195047977253,
      "tokens_seen": 2470182912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025402206619859583,
      "loss": 1.2536,
      "theoretical_loss": 3.400212352229221,
      "tokens_seen": 2470248448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002540020060180542,
      "loss": 1.4328,
      "theoretical_loss": 3.4002051999036036,
      "tokens_seen": 2470313984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025398194583751255,
      "loss": 1.4135,
      "theoretical_loss": 3.400198047820859,
      "tokens_seen": 2470379520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002539618856569709,
      "loss": 1.3562,
      "theoretical_loss": 3.400190895980971,
      "tokens_seen": 2470445056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002539418254764293,
      "loss": 1.3296,
      "theoretical_loss": 3.4001837443839262,
      "tokens_seen": 2470510592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002539217652958877,
      "loss": 1.3178,
      "theoretical_loss": 3.400176593029709,
      "tokens_seen": 2470576128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025390170511534606,
      "loss": 1.3027,
      "theoretical_loss": 3.4001694419183055,
      "tokens_seen": 2470641664
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538816449348044,
      "loss": 1.3103,
      "theoretical_loss": 3.4001622910497007,
      "tokens_seen": 2470707200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538615847542628,
      "loss": 1.4414,
      "theoretical_loss": 3.4001551404238795,
      "tokens_seen": 2470772736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538415245737212,
      "loss": 1.2714,
      "theoretical_loss": 3.400147990040828,
      "tokens_seen": 2470838272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025382146439317956,
      "loss": 1.2991,
      "theoretical_loss": 3.400140839900531,
      "tokens_seen": 2470903808
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538014042126379,
      "loss": 1.3553,
      "theoretical_loss": 3.400133690002974,
      "tokens_seen": 2470969344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537813440320963,
      "loss": 1.3477,
      "theoretical_loss": 3.4001265403481424,
      "tokens_seen": 2471034880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025376128385155465,
      "loss": 1.285,
      "theoretical_loss": 3.400119390936021,
      "tokens_seen": 2471100416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025374122367101306,
      "loss": 1.2024,
      "theoretical_loss": 3.4001122417665957,
      "tokens_seen": 2471165952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537211634904714,
      "loss": 1.3761,
      "theoretical_loss": 3.4001050928398517,
      "tokens_seen": 2471231488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537011033099298,
      "loss": 1.3542,
      "theoretical_loss": 3.400097944155774,
      "tokens_seen": 2471297024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025368104312938815,
      "loss": 1.1995,
      "theoretical_loss": 3.4000907957143487,
      "tokens_seen": 2471362560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025366098294884657,
      "loss": 1.3057,
      "theoretical_loss": 3.4000836475155602,
      "tokens_seen": 2471428096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025364092276830493,
      "loss": 1.3117,
      "theoretical_loss": 3.4000764995593946,
      "tokens_seen": 2471493632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536208625877633,
      "loss": 1.2502,
      "theoretical_loss": 3.4000693518458367,
      "tokens_seen": 2471559168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025360080240722165,
      "loss": 1.2898,
      "theoretical_loss": 3.400062204374872,
      "tokens_seen": 2471624704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025358074222668,
      "loss": 1.3149,
      "theoretical_loss": 3.400055057146486,
      "tokens_seen": 2471690240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025356068204613843,
      "loss": 1.3647,
      "theoretical_loss": 3.4000479101606635,
      "tokens_seen": 2471755776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535406218655968,
      "loss": 1.2987,
      "theoretical_loss": 3.4000407634173904,
      "tokens_seen": 2471821312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025352056168505516,
      "loss": 1.2208,
      "theoretical_loss": 3.400033616916652,
      "tokens_seen": 2471886848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535005015045135,
      "loss": 1.2578,
      "theoretical_loss": 3.4000264706584336,
      "tokens_seen": 2471952384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025348044132397193,
      "loss": 1.3675,
      "theoretical_loss": 3.40001932464272,
      "tokens_seen": 2472017920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534603811434303,
      "loss": 1.3747,
      "theoretical_loss": 3.4000121788694977,
      "tokens_seen": 2472083456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025344032096288866,
      "loss": 1.1592,
      "theoretical_loss": 3.4000050333387506,
      "tokens_seen": 2472148992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253420260782347,
      "loss": 1.2827,
      "theoretical_loss": 3.399997888050465,
      "tokens_seen": 2472214528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534002006018054,
      "loss": 1.3685,
      "theoretical_loss": 3.3999907430046266,
      "tokens_seen": 2472280064
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533801404212638,
      "loss": 1.4195,
      "theoretical_loss": 3.3999835982012194,
      "tokens_seen": 2472345600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025336008024072216,
      "loss": 1.3339,
      "theoretical_loss": 3.39997645364023,
      "tokens_seen": 2472411136
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 899541,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3970298767089844,
      "objective/train/theoretical_loss": 3.399969309321643,
      "objective/train/tokens_used": 842935776,
      "theoretical_loss": 3.399969309321643,
      "tokens_seen": 2472476672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533400200601805,
      "loss": 1.3013,
      "theoretical_loss": 3.399969309321643,
      "tokens_seen": 2472476672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533199598796389,
      "loss": 1.1648,
      "theoretical_loss": 3.399962165245444,
      "tokens_seen": 2472542208
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532998996990973,
      "loss": 1.1538,
      "theoretical_loss": 3.3999550214116185,
      "tokens_seen": 2472607744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025327983951855566,
      "loss": 1.2646,
      "theoretical_loss": 3.3999478778201517,
      "tokens_seen": 2472673280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253259779338014,
      "loss": 1.4066,
      "theoretical_loss": 3.399940734471029,
      "tokens_seen": 2472738816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532397191574724,
      "loss": 1.4237,
      "theoretical_loss": 3.3999335913642357,
      "tokens_seen": 2472804352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025321965897693075,
      "loss": 1.4168,
      "theoretical_loss": 3.399926448499757,
      "tokens_seen": 2472869888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025319959879638917,
      "loss": 1.3408,
      "theoretical_loss": 3.3999193058775785,
      "tokens_seen": 2472935424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025317953861584753,
      "loss": 1.233,
      "theoretical_loss": 3.3999121634976857,
      "tokens_seen": 2473000960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002531594784353059,
      "loss": 1.2979,
      "theoretical_loss": 3.3999050213600635,
      "tokens_seen": 2473066496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025313941825476425,
      "loss": 1.347,
      "theoretical_loss": 3.3998978794646977,
      "tokens_seen": 2473132032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025311935807422267,
      "loss": 1.3366,
      "theoretical_loss": 3.3998907378115737,
      "tokens_seen": 2473197568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025309929789368103,
      "loss": 1.3087,
      "theoretical_loss": 3.3998835964006764,
      "tokens_seen": 2473263104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002530792377131394,
      "loss": 1.2667,
      "theoretical_loss": 3.3998764552319916,
      "tokens_seen": 2473328640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025305917753259776,
      "loss": 1.307,
      "theoretical_loss": 3.399869314305504,
      "tokens_seen": 2473394176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002530391173520562,
      "loss": 1.2641,
      "theoretical_loss": 3.3998621736212002,
      "tokens_seen": 2473459712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025301905717151454,
      "loss": 1.3426,
      "theoretical_loss": 3.3998550331790645,
      "tokens_seen": 2473525248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025299899699097295,
      "loss": 1.2919,
      "theoretical_loss": 3.3998478929790825,
      "tokens_seen": 2473590784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529789368104313,
      "loss": 1.2603,
      "theoretical_loss": 3.39984075302124,
      "tokens_seen": 2473656320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529588766298897,
      "loss": 1.3741,
      "theoretical_loss": 3.3998336133055216,
      "tokens_seen": 2473721856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529388164493481,
      "loss": 1.2717,
      "theoretical_loss": 3.3998264738319133,
      "tokens_seen": 2473787392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025291875626880646,
      "loss": 1.3289,
      "theoretical_loss": 3.3998193346004006,
      "tokens_seen": 2473852928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528986960882648,
      "loss": 1.2296,
      "theoretical_loss": 3.399812195610968,
      "tokens_seen": 2473918464
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528786359077232,
      "loss": 1.2522,
      "theoretical_loss": 3.399805056863602,
      "tokens_seen": 2473984000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528585757271816,
      "loss": 1.2973,
      "theoretical_loss": 3.399797918358287,
      "tokens_seen": 2474049536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025283851554663996,
      "loss": 1.2837,
      "theoretical_loss": 3.399790780095009,
      "tokens_seen": 2474115072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528184553660983,
      "loss": 1.3227,
      "theoretical_loss": 3.399783642073753,
      "tokens_seen": 2474180608
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527983951855567,
      "loss": 1.287,
      "theoretical_loss": 3.399776504294505,
      "tokens_seen": 2474246144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025277833500501505,
      "loss": 1.2835,
      "theoretical_loss": 3.39976936675725,
      "tokens_seen": 2474311680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025275827482447346,
      "loss": 1.3138,
      "theoretical_loss": 3.399762229461973,
      "tokens_seen": 2474377216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527382146439318,
      "loss": 1.2274,
      "theoretical_loss": 3.3997550924086597,
      "tokens_seen": 2474442752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527181544633902,
      "loss": 1.3367,
      "theoretical_loss": 3.3997479555972956,
      "tokens_seen": 2474508288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025269809428284855,
      "loss": 1.3539,
      "theoretical_loss": 3.3997408190278664,
      "tokens_seen": 2474573824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025267803410230697,
      "loss": 1.3539,
      "theoretical_loss": 3.399733682700357,
      "tokens_seen": 2474639360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025265797392176533,
      "loss": 1.2803,
      "theoretical_loss": 3.3997265466147524,
      "tokens_seen": 2474704896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002526379137412237,
      "loss": 1.1842,
      "theoretical_loss": 3.399719410771039,
      "tokens_seen": 2474770432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025261785356068205,
      "loss": 1.3009,
      "theoretical_loss": 3.3997122751692013,
      "tokens_seen": 2474835968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525977933801404,
      "loss": 1.2362,
      "theoretical_loss": 3.3997051398092255,
      "tokens_seen": 2474901504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025257773319959883,
      "loss": 1.2856,
      "theoretical_loss": 3.3996980046910963,
      "tokens_seen": 2474967040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525576730190572,
      "loss": 1.2006,
      "theoretical_loss": 3.3996908698147994,
      "tokens_seen": 2475032576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025253761283851556,
      "loss": 1.3353,
      "theoretical_loss": 3.3996837351803206,
      "tokens_seen": 2475098112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525175526579739,
      "loss": 1.2818,
      "theoretical_loss": 3.3996766007876444,
      "tokens_seen": 2475163648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025249749247743233,
      "loss": 1.4091,
      "theoretical_loss": 3.399669466636757,
      "tokens_seen": 2475229184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002524774322968907,
      "loss": 1.3858,
      "theoretical_loss": 3.3996623327276434,
      "tokens_seen": 2475294720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025245737211634906,
      "loss": 1.296,
      "theoretical_loss": 3.3996551990602892,
      "tokens_seen": 2475360256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002524373119358074,
      "loss": 1.4147,
      "theoretical_loss": 3.3996480656346795,
      "tokens_seen": 2475425792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002524172517552658,
      "loss": 1.262,
      "theoretical_loss": 3.3996409324508,
      "tokens_seen": 2475491328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523971915747242,
      "loss": 1.2315,
      "theoretical_loss": 3.399633799508636,
      "tokens_seen": 2475556864
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025237713139418256,
      "loss": 1.3357,
      "theoretical_loss": 3.399626666808173,
      "tokens_seen": 2475622400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523570712136409,
      "loss": 1.3354,
      "theoretical_loss": 3.3996195343493962,
      "tokens_seen": 2475687936
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 901124,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2519432306289673,
      "objective/train/theoretical_loss": 3.3996124021322918,
      "objective/train/tokens_used": 846212576,
      "theoretical_loss": 3.3996124021322918,
      "tokens_seen": 2475753472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523370110330993,
      "loss": 1.3891,
      "theoretical_loss": 3.3996124021322918,
      "tokens_seen": 2475753472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523169508525577,
      "loss": 1.3378,
      "theoretical_loss": 3.3996052701568438,
      "tokens_seen": 2475819008
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025229689067201606,
      "loss": 1.2418,
      "theoretical_loss": 3.399598138423039,
      "tokens_seen": 2475884544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025227683049147443,
      "loss": 1.4269,
      "theoretical_loss": 3.3995910069308617,
      "tokens_seen": 2475950080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002522567703109328,
      "loss": 1.3235,
      "theoretical_loss": 3.399583875680298,
      "tokens_seen": 2476015616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025223671013039115,
      "loss": 1.2395,
      "theoretical_loss": 3.399576744671333,
      "tokens_seen": 2476081152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025221664994984957,
      "loss": 1.3774,
      "theoretical_loss": 3.3995696139039526,
      "tokens_seen": 2476146688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025219658976930793,
      "loss": 1.2903,
      "theoretical_loss": 3.3995624833781415,
      "tokens_seen": 2476212224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002521765295887663,
      "loss": 1.3205,
      "theoretical_loss": 3.399555353093886,
      "tokens_seen": 2476277760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025215646940822465,
      "loss": 1.2984,
      "theoretical_loss": 3.3995482230511707,
      "tokens_seen": 2476343296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025213640922768307,
      "loss": 1.4112,
      "theoretical_loss": 3.399541093249981,
      "tokens_seen": 2476408832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025211634904714143,
      "loss": 1.3454,
      "theoretical_loss": 3.3995339636903035,
      "tokens_seen": 2476474368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002520962888665998,
      "loss": 1.2686,
      "theoretical_loss": 3.3995268343721223,
      "tokens_seen": 2476539904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025207622868605816,
      "loss": 1.31,
      "theoretical_loss": 3.3995197052954236,
      "tokens_seen": 2476605440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002520561685055166,
      "loss": 1.2162,
      "theoretical_loss": 3.3995125764601926,
      "tokens_seen": 2476670976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025203610832497494,
      "loss": 1.3281,
      "theoretical_loss": 3.3995054478664146,
      "tokens_seen": 2476736512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002520160481444333,
      "loss": 1.3353,
      "theoretical_loss": 3.399498319514075,
      "tokens_seen": 2476802048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025199598796389166,
      "loss": 1.2379,
      "theoretical_loss": 3.39949119140316,
      "tokens_seen": 2476867584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025197592778335,
      "loss": 1.2788,
      "theoretical_loss": 3.3994840635336536,
      "tokens_seen": 2476933120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025195586760280844,
      "loss": 1.2848,
      "theoretical_loss": 3.3994769359055423,
      "tokens_seen": 2476998656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002519358074222668,
      "loss": 1.2879,
      "theoretical_loss": 3.3994698085188118,
      "tokens_seen": 2477064192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025191574724172516,
      "loss": 1.2709,
      "theoretical_loss": 3.3994626813734463,
      "tokens_seen": 2477129728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002518956870611835,
      "loss": 1.3468,
      "theoretical_loss": 3.3994555544694323,
      "tokens_seen": 2477195264
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025187562688064194,
      "loss": 1.3157,
      "theoretical_loss": 3.399448427806755,
      "tokens_seen": 2477260800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002518555667001003,
      "loss": 1.271,
      "theoretical_loss": 3.3994413013853997,
      "tokens_seen": 2477326336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025183550651955867,
      "loss": 1.4645,
      "theoretical_loss": 3.399434175205352,
      "tokens_seen": 2477391872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025181544633901703,
      "loss": 1.3035,
      "theoretical_loss": 3.3994270492665972,
      "tokens_seen": 2477457408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517953861584754,
      "loss": 1.3667,
      "theoretical_loss": 3.3994199235691203,
      "tokens_seen": 2477522944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517753259779338,
      "loss": 1.3674,
      "theoretical_loss": 3.399412798112908,
      "tokens_seen": 2477588480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025175526579739217,
      "loss": 1.3269,
      "theoretical_loss": 3.3994056728979443,
      "tokens_seen": 2477654016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025173520561685053,
      "loss": 1.2677,
      "theoretical_loss": 3.399398547924216,
      "tokens_seen": 2477719552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517151454363089,
      "loss": 1.272,
      "theoretical_loss": 3.3993914231917075,
      "tokens_seen": 2477785088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516950852557673,
      "loss": 1.223,
      "theoretical_loss": 3.399384298700405,
      "tokens_seen": 2477850624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516750250752257,
      "loss": 1.382,
      "theoretical_loss": 3.3993771744502927,
      "tokens_seen": 2477916160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025165496489468404,
      "loss": 1.4047,
      "theoretical_loss": 3.399370050441358,
      "tokens_seen": 2477981696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516349047141424,
      "loss": 1.3301,
      "theoretical_loss": 3.399362926673585,
      "tokens_seen": 2478047232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025161484453360076,
      "loss": 1.3455,
      "theoretical_loss": 3.3993558031469595,
      "tokens_seen": 2478112768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515947843530592,
      "loss": 1.2766,
      "theoretical_loss": 3.3993486798614665,
      "tokens_seen": 2478178304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025157472417251754,
      "loss": 1.3943,
      "theoretical_loss": 3.3993415568170926,
      "tokens_seen": 2478243840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515546639919759,
      "loss": 1.3411,
      "theoretical_loss": 3.399334434013822,
      "tokens_seen": 2478309376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025153460381143426,
      "loss": 1.2909,
      "theoretical_loss": 3.399327311451641,
      "tokens_seen": 2478374912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515145436308927,
      "loss": 1.3031,
      "theoretical_loss": 3.3993201891305347,
      "tokens_seen": 2478440448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514944834503511,
      "loss": 1.3338,
      "theoretical_loss": 3.3993130670504885,
      "tokens_seen": 2478505984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025147442326980946,
      "loss": 1.4301,
      "theoretical_loss": 3.399305945211488,
      "tokens_seen": 2478571520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514543630892678,
      "loss": 1.336,
      "theoretical_loss": 3.399298823613519,
      "tokens_seen": 2478637056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514343029087262,
      "loss": 1.34,
      "theoretical_loss": 3.3992917022565665,
      "tokens_seen": 2478702592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514142427281846,
      "loss": 1.3152,
      "theoretical_loss": 3.3992845811406163,
      "tokens_seen": 2478768128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025139418254764296,
      "loss": 1.2802,
      "theoretical_loss": 3.3992774602656537,
      "tokens_seen": 2478833664
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513741223671013,
      "loss": 1.5276,
      "theoretical_loss": 3.399270339631664,
      "tokens_seen": 2478899200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513540621865597,
      "loss": 1.2222,
      "theoretical_loss": 3.399263219238633,
      "tokens_seen": 2478964736
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 901918,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4798810482025146,
      "objective/train/theoretical_loss": 3.399256099086546,
      "objective/train/tokens_used": 849489376,
      "theoretical_loss": 3.399256099086546,
      "tokens_seen": 2479030272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513340020060181,
      "loss": 1.3852,
      "theoretical_loss": 3.399256099086546,
      "tokens_seen": 2479030272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025131394182547647,
      "loss": 1.178,
      "theoretical_loss": 3.3992489791753884,
      "tokens_seen": 2479095808
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025129388164493483,
      "loss": 1.3435,
      "theoretical_loss": 3.3992418595051457,
      "tokens_seen": 2479161344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002512738214643932,
      "loss": 1.2602,
      "theoretical_loss": 3.3992347400758036,
      "tokens_seen": 2479226880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025125376128385155,
      "loss": 1.3079,
      "theoretical_loss": 3.3992276208873475,
      "tokens_seen": 2479292416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025123370110330997,
      "loss": 1.1964,
      "theoretical_loss": 3.3992205019397628,
      "tokens_seen": 2479357952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025121364092276833,
      "loss": 1.3512,
      "theoretical_loss": 3.399213383233035,
      "tokens_seen": 2479423488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511935807422267,
      "loss": 1.3259,
      "theoretical_loss": 3.3992062647671495,
      "tokens_seen": 2479489024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025117352056168506,
      "loss": 1.3042,
      "theoretical_loss": 3.399199146542092,
      "tokens_seen": 2479554560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025115346038114347,
      "loss": 1.3641,
      "theoretical_loss": 3.399192028557848,
      "tokens_seen": 2479620096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025113340020060183,
      "loss": 1.1336,
      "theoretical_loss": 3.3991849108144025,
      "tokens_seen": 2479685632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511133400200602,
      "loss": 1.3666,
      "theoretical_loss": 3.3991777933117415,
      "tokens_seen": 2479751168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025109327983951856,
      "loss": 1.3548,
      "theoretical_loss": 3.3991706760498506,
      "tokens_seen": 2479816704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000251073219658977,
      "loss": 1.2685,
      "theoretical_loss": 3.3991635590287146,
      "tokens_seen": 2479882240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025105315947843534,
      "loss": 1.2976,
      "theoretical_loss": 3.3991564422483203,
      "tokens_seen": 2479947776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002510330992978937,
      "loss": 1.3874,
      "theoretical_loss": 3.399149325708651,
      "tokens_seen": 2480013312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025101303911735206,
      "loss": 1.3662,
      "theoretical_loss": 3.3991422094096944,
      "tokens_seen": 2480078848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002509929789368104,
      "loss": 1.3587,
      "theoretical_loss": 3.3991350933514353,
      "tokens_seen": 2480144384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025097291875626884,
      "loss": 1.2371,
      "theoretical_loss": 3.3991279775338583,
      "tokens_seen": 2480209920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002509528585757272,
      "loss": 1.3304,
      "theoretical_loss": 3.39912086195695,
      "tokens_seen": 2480275456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025093279839518556,
      "loss": 1.3329,
      "theoretical_loss": 3.3991137466206958,
      "tokens_seen": 2480340992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002509127382146439,
      "loss": 1.3344,
      "theoretical_loss": 3.3991066315250804,
      "tokens_seen": 2480406528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025089267803410234,
      "loss": 1.1925,
      "theoretical_loss": 3.3990995166700904,
      "tokens_seen": 2480472064
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508726178535607,
      "loss": 1.2723,
      "theoretical_loss": 3.39909240205571,
      "tokens_seen": 2480537600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025085255767301907,
      "loss": 1.2843,
      "theoretical_loss": 3.3990852876819257,
      "tokens_seen": 2480603136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025083249749247743,
      "loss": 1.2822,
      "theoretical_loss": 3.399078173548723,
      "tokens_seen": 2480668672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508124373119358,
      "loss": 1.2573,
      "theoretical_loss": 3.399071059656087,
      "tokens_seen": 2480734208
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507923771313942,
      "loss": 1.3705,
      "theoretical_loss": 3.3990639460040035,
      "tokens_seen": 2480799744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025077231695085257,
      "loss": 1.2406,
      "theoretical_loss": 3.3990568325924575,
      "tokens_seen": 2480865280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025075225677031093,
      "loss": 1.465,
      "theoretical_loss": 3.399049719421435,
      "tokens_seen": 2480930816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507321965897693,
      "loss": 1.3613,
      "theoretical_loss": 3.399042606490922,
      "tokens_seen": 2480996352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507121364092277,
      "loss": 1.3736,
      "theoretical_loss": 3.399035493800903,
      "tokens_seen": 2481061888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506920762286861,
      "loss": 1.3005,
      "theoretical_loss": 3.3990283813513633,
      "tokens_seen": 2481127424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025067201604814444,
      "loss": 1.2957,
      "theoretical_loss": 3.39902126914229,
      "tokens_seen": 2481192960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506519558676028,
      "loss": 1.3358,
      "theoretical_loss": 3.399014157173667,
      "tokens_seen": 2481258496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025063189568706116,
      "loss": 1.2478,
      "theoretical_loss": 3.3990070454454813,
      "tokens_seen": 2481324032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506118355065196,
      "loss": 1.2297,
      "theoretical_loss": 3.398999933957717,
      "tokens_seen": 2481389568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025059177532597794,
      "loss": 1.3491,
      "theoretical_loss": 3.39899282271036,
      "tokens_seen": 2481455104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505717151454363,
      "loss": 1.2844,
      "theoretical_loss": 3.3989857117033964,
      "tokens_seen": 2481520640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025055165496489466,
      "loss": 1.376,
      "theoretical_loss": 3.3989786009368115,
      "tokens_seen": 2481586176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505315947843531,
      "loss": 1.3269,
      "theoretical_loss": 3.3989714904105908,
      "tokens_seen": 2481651712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025051153460381144,
      "loss": 1.3084,
      "theoretical_loss": 3.3989643801247196,
      "tokens_seen": 2481717248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002504914744232698,
      "loss": 1.2167,
      "theoretical_loss": 3.3989572700791832,
      "tokens_seen": 2481782784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025047141424272817,
      "loss": 1.3676,
      "theoretical_loss": 3.3989501602739676,
      "tokens_seen": 2481848320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025045135406218653,
      "loss": 1.2946,
      "theoretical_loss": 3.3989430507090588,
      "tokens_seen": 2481913856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025043129388164495,
      "loss": 1.3691,
      "theoretical_loss": 3.3989359413844413,
      "tokens_seen": 2481979392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002504112337011033,
      "loss": 1.2622,
      "theoretical_loss": 3.398928832300101,
      "tokens_seen": 2482044928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025039117352056167,
      "loss": 1.2578,
      "theoretical_loss": 3.398921723456024,
      "tokens_seen": 2482110464
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025037111334002003,
      "loss": 1.4824,
      "theoretical_loss": 3.398914614852195,
      "tokens_seen": 2482176000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025035105315947845,
      "loss": 1.3281,
      "theoretical_loss": 3.3989075064886,
      "tokens_seen": 2482241536
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 903113,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3541759252548218,
      "objective/train/theoretical_loss": 3.3989003983652246,
      "objective/train/tokens_used": 852766176,
      "theoretical_loss": 3.3989003983652246,
      "tokens_seen": 2482307072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503309929789368,
      "loss": 1.3556,
      "theoretical_loss": 3.3989003983652246,
      "tokens_seen": 2482307072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503109327983952,
      "loss": 1.2891,
      "theoretical_loss": 3.3988932904820537,
      "tokens_seen": 2482372608
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025029087261785354,
      "loss": 1.2133,
      "theoretical_loss": 3.3988861828390737,
      "tokens_seen": 2482438144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502708124373119,
      "loss": 1.426,
      "theoretical_loss": 3.3988790754362697,
      "tokens_seen": 2482503680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502507522567703,
      "loss": 1.3232,
      "theoretical_loss": 3.3988719682736273,
      "tokens_seen": 2482569216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502306920762287,
      "loss": 1.3657,
      "theoretical_loss": 3.3988648613511323,
      "tokens_seen": 2482634752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025021063189568704,
      "loss": 1.2924,
      "theoretical_loss": 3.3988577546687697,
      "tokens_seen": 2482700288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501905717151454,
      "loss": 1.2758,
      "theoretical_loss": 3.3988506482265253,
      "tokens_seen": 2482765824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501705115346038,
      "loss": 1.2373,
      "theoretical_loss": 3.3988435420243848,
      "tokens_seen": 2482831360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501504513540622,
      "loss": 1.2455,
      "theoretical_loss": 3.3988364360623335,
      "tokens_seen": 2482896896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025013039117352054,
      "loss": 1.263,
      "theoretical_loss": 3.3988293303403574,
      "tokens_seen": 2482962432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501103309929789,
      "loss": 1.2571,
      "theoretical_loss": 3.3988222248584417,
      "tokens_seen": 2483027968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025009027081243727,
      "loss": 1.3233,
      "theoretical_loss": 3.398815119616572,
      "tokens_seen": 2483093504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002500702106318957,
      "loss": 1.3388,
      "theoretical_loss": 3.3988080146147337,
      "tokens_seen": 2483159040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025005015045135405,
      "loss": 1.4493,
      "theoretical_loss": 3.3988009098529126,
      "tokens_seen": 2483224576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002500300902708124,
      "loss": 1.3901,
      "theoretical_loss": 3.398793805331094,
      "tokens_seen": 2483290112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025001003009027077,
      "loss": 1.2823,
      "theoretical_loss": 3.3987867010492643,
      "tokens_seen": 2483355648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002499899699097292,
      "loss": 1.3034,
      "theoretical_loss": 3.3987795970074077,
      "tokens_seen": 2483421184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024996990972918755,
      "loss": 1.2635,
      "theoretical_loss": 3.398772493205511,
      "tokens_seen": 2483486720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024994984954864597,
      "loss": 1.3248,
      "theoretical_loss": 3.398765389643559,
      "tokens_seen": 2483552256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024992978936810433,
      "loss": 1.4511,
      "theoretical_loss": 3.398758286321537,
      "tokens_seen": 2483617792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002499097291875627,
      "loss": 1.3944,
      "theoretical_loss": 3.398751183239432,
      "tokens_seen": 2483683328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024988966900702105,
      "loss": 1.207,
      "theoretical_loss": 3.3987440803972278,
      "tokens_seen": 2483748864
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002498696088264794,
      "loss": 1.3166,
      "theoretical_loss": 3.3987369777949112,
      "tokens_seen": 2483814400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024984954864593783,
      "loss": 1.1807,
      "theoretical_loss": 3.3987298754324673,
      "tokens_seen": 2483879936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002498294884653962,
      "loss": 1.3815,
      "theoretical_loss": 3.3987227733098817,
      "tokens_seen": 2483945472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024980942828485455,
      "loss": 1.4362,
      "theoretical_loss": 3.39871567142714,
      "tokens_seen": 2484011008
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497893681043129,
      "loss": 1.3367,
      "theoretical_loss": 3.3987085697842274,
      "tokens_seen": 2484076544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024976930792377133,
      "loss": 1.2403,
      "theoretical_loss": 3.3987014683811303,
      "tokens_seen": 2484142080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497492477432297,
      "loss": 1.276,
      "theoretical_loss": 3.398694367217834,
      "tokens_seen": 2484207616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024972918756268806,
      "loss": 1.235,
      "theoretical_loss": 3.3986872662943233,
      "tokens_seen": 2484273152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497091273821464,
      "loss": 1.4189,
      "theoretical_loss": 3.3986801656105845,
      "tokens_seen": 2484338688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496890672016048,
      "loss": 1.4088,
      "theoretical_loss": 3.3986730651666033,
      "tokens_seen": 2484404224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496690070210632,
      "loss": 1.2772,
      "theoretical_loss": 3.398665964962365,
      "tokens_seen": 2484469760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024964894684052156,
      "loss": 1.3537,
      "theoretical_loss": 3.3986588649978553,
      "tokens_seen": 2484535296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496288866599799,
      "loss": 1.2367,
      "theoretical_loss": 3.3986517652730597,
      "tokens_seen": 2484600832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024960882647943834,
      "loss": 1.4176,
      "theoretical_loss": 3.3986446657879634,
      "tokens_seen": 2484666368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495887662988967,
      "loss": 1.4045,
      "theoretical_loss": 3.398637566542553,
      "tokens_seen": 2484731904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024956870611835506,
      "loss": 1.3051,
      "theoretical_loss": 3.3986304675368126,
      "tokens_seen": 2484797440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495486459378135,
      "loss": 1.3138,
      "theoretical_loss": 3.3986233687707292,
      "tokens_seen": 2484862976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024952858575727184,
      "loss": 1.2604,
      "theoretical_loss": 3.398616270244288,
      "tokens_seen": 2484928512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495085255767302,
      "loss": 1.1999,
      "theoretical_loss": 3.398609171957474,
      "tokens_seen": 2484994048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024948846539618857,
      "loss": 1.305,
      "theoretical_loss": 3.3986020739102734,
      "tokens_seen": 2485059584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024946840521564693,
      "loss": 1.3133,
      "theoretical_loss": 3.3985949761026717,
      "tokens_seen": 2485125120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024944834503510535,
      "loss": 1.2485,
      "theoretical_loss": 3.398587878534654,
      "tokens_seen": 2485190656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002494282848545637,
      "loss": 1.2223,
      "theoretical_loss": 3.3985807812062068,
      "tokens_seen": 2485256192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024940822467402207,
      "loss": 1.3108,
      "theoretical_loss": 3.3985736841173146,
      "tokens_seen": 2485321728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024938816449348043,
      "loss": 1.2937,
      "theoretical_loss": 3.3985665872679642,
      "tokens_seen": 2485387264
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024936810431293885,
      "loss": 1.2634,
      "theoretical_loss": 3.3985594906581404,
      "tokens_seen": 2485452800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002493480441323972,
      "loss": 1.2663,
      "theoretical_loss": 3.3985523942878286,
      "tokens_seen": 2485518336
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 904288,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2782502174377441,
      "objective/train/theoretical_loss": 3.3985452981570154,
      "objective/train/tokens_used": 856042976,
      "theoretical_loss": 3.3985452981570154,
      "tokens_seen": 2485583872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002493279839518556,
      "loss": 1.252,
      "theoretical_loss": 3.3985452981570154,
      "tokens_seen": 2485583872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024930792377131394,
      "loss": 1.3401,
      "theoretical_loss": 3.3985382022656854,
      "tokens_seen": 2485649408
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492878635907723,
      "loss": 1.4326,
      "theoretical_loss": 3.3985311066138246,
      "tokens_seen": 2485714944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492678034102307,
      "loss": 1.2876,
      "theoretical_loss": 3.398524011201419,
      "tokens_seen": 2485780480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492477432296891,
      "loss": 1.3128,
      "theoretical_loss": 3.3985169160284534,
      "tokens_seen": 2485846016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024922768304914744,
      "loss": 1.24,
      "theoretical_loss": 3.398509821094914,
      "tokens_seen": 2485911552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492076228686058,
      "loss": 1.186,
      "theoretical_loss": 3.3985027264007863,
      "tokens_seen": 2485977088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491875626880642,
      "loss": 1.2708,
      "theoretical_loss": 3.3984956319460555,
      "tokens_seen": 2486042624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491675025075226,
      "loss": 1.366,
      "theoretical_loss": 3.3984885377307075,
      "tokens_seen": 2486108160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024914744232698094,
      "loss": 1.2984,
      "theoretical_loss": 3.3984814437547284,
      "tokens_seen": 2486173696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491273821464393,
      "loss": 1.3309,
      "theoretical_loss": 3.398474350018103,
      "tokens_seen": 2486239232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024910732196589767,
      "loss": 1.2778,
      "theoretical_loss": 3.3984672565208176,
      "tokens_seen": 2486304768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490872617853561,
      "loss": 1.2828,
      "theoretical_loss": 3.3984601632628575,
      "tokens_seen": 2486370304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024906720160481445,
      "loss": 1.2744,
      "theoretical_loss": 3.3984530702442077,
      "tokens_seen": 2486435840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490471414242728,
      "loss": 1.2582,
      "theoretical_loss": 3.3984459774648554,
      "tokens_seen": 2486501376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024902708124373117,
      "loss": 1.3138,
      "theoretical_loss": 3.3984388849247846,
      "tokens_seen": 2486566912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490070210631896,
      "loss": 1.2444,
      "theoretical_loss": 3.398431792623982,
      "tokens_seen": 2486632448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024898696088264795,
      "loss": 1.3583,
      "theoretical_loss": 3.3984247005624324,
      "tokens_seen": 2486697984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002489669007021063,
      "loss": 1.4093,
      "theoretical_loss": 3.3984176087401217,
      "tokens_seen": 2486763520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002489468405215647,
      "loss": 1.2117,
      "theoretical_loss": 3.398410517157036,
      "tokens_seen": 2486829056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024892678034102304,
      "loss": 1.3196,
      "theoretical_loss": 3.398403425813161,
      "tokens_seen": 2486894592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024890672016048145,
      "loss": 1.3787,
      "theoretical_loss": 3.3983963347084813,
      "tokens_seen": 2486960128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488866599799398,
      "loss": 1.3075,
      "theoretical_loss": 3.398389243842983,
      "tokens_seen": 2487025664
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024886659979939823,
      "loss": 1.3273,
      "theoretical_loss": 3.398382153216652,
      "tokens_seen": 2487091200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488465396188566,
      "loss": 1.3336,
      "theoretical_loss": 3.398375062829474,
      "tokens_seen": 2487156736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024882647943831496,
      "loss": 1.3448,
      "theoretical_loss": 3.3983679726814344,
      "tokens_seen": 2487222272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488064192577733,
      "loss": 1.2555,
      "theoretical_loss": 3.398360882772519,
      "tokens_seen": 2487287808
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024878635907723173,
      "loss": 1.1584,
      "theoretical_loss": 3.398353793102713,
      "tokens_seen": 2487353344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487662988966901,
      "loss": 1.2994,
      "theoretical_loss": 3.3983467036720025,
      "tokens_seen": 2487418880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024874623871614846,
      "loss": 1.4172,
      "theoretical_loss": 3.398339614480373,
      "tokens_seen": 2487484416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487261785356068,
      "loss": 1.2021,
      "theoretical_loss": 3.3983325255278096,
      "tokens_seen": 2487549952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487061183550652,
      "loss": 1.3081,
      "theoretical_loss": 3.398325436814299,
      "tokens_seen": 2487615488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486860581745236,
      "loss": 1.2459,
      "theoretical_loss": 3.398318348339826,
      "tokens_seen": 2487681024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024866599799398196,
      "loss": 1.4061,
      "theoretical_loss": 3.398311260104377,
      "tokens_seen": 2487746560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486459378134403,
      "loss": 1.2761,
      "theoretical_loss": 3.398304172107937,
      "tokens_seen": 2487812096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486258776328987,
      "loss": 1.4075,
      "theoretical_loss": 3.3982970843504914,
      "tokens_seen": 2487877632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486058174523571,
      "loss": 1.2773,
      "theoretical_loss": 3.398289996832027,
      "tokens_seen": 2487943168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024858575727181546,
      "loss": 1.3485,
      "theoretical_loss": 3.398282909552528,
      "tokens_seen": 2488008704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002485656970912738,
      "loss": 1.2861,
      "theoretical_loss": 3.398275822511981,
      "tokens_seen": 2488074240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002485456369107322,
      "loss": 1.2942,
      "theoretical_loss": 3.3982687357103716,
      "tokens_seen": 2488139776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024852557673019055,
      "loss": 1.2823,
      "theoretical_loss": 3.398261649147685,
      "tokens_seen": 2488205312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024850551654964897,
      "loss": 1.2922,
      "theoretical_loss": 3.398254562823907,
      "tokens_seen": 2488270848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024848545636910733,
      "loss": 1.3095,
      "theoretical_loss": 3.398247476739024,
      "tokens_seen": 2488336384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002484653961885657,
      "loss": 1.3002,
      "theoretical_loss": 3.3982403908930205,
      "tokens_seen": 2488401920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024844533600802405,
      "loss": 1.4044,
      "theoretical_loss": 3.398233305285883,
      "tokens_seen": 2488467456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024842527582748247,
      "loss": 1.3391,
      "theoretical_loss": 3.3982262199175968,
      "tokens_seen": 2488532992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024840521564694083,
      "loss": 1.3447,
      "theoretical_loss": 3.3982191347881474,
      "tokens_seen": 2488598528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002483851554663992,
      "loss": 1.3976,
      "theoretical_loss": 3.398212049897521,
      "tokens_seen": 2488664064
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024836509528585756,
      "loss": 1.3185,
      "theoretical_loss": 3.3982049652457023,
      "tokens_seen": 2488729600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248345035105316,
      "loss": 1.3527,
      "theoretical_loss": 3.398197880832678,
      "tokens_seen": 2488795136
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 905609,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2062811851501465,
      "objective/train/theoretical_loss": 3.398190796658433,
      "objective/train/tokens_used": 859319776,
      "theoretical_loss": 3.398190796658433,
      "tokens_seen": 2488860672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024832497492477434,
      "loss": 1.2004,
      "theoretical_loss": 3.398190796658433,
      "tokens_seen": 2488860672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002483049147442327,
      "loss": 1.2462,
      "theoretical_loss": 3.398183712722954,
      "tokens_seen": 2488926208
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024828485456369106,
      "loss": 1.309,
      "theoretical_loss": 3.3981766290262256,
      "tokens_seen": 2488991744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482647943831494,
      "loss": 1.3463,
      "theoretical_loss": 3.3981695455682335,
      "tokens_seen": 2489057280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024824473420260784,
      "loss": 1.3427,
      "theoretical_loss": 3.398162462348964,
      "tokens_seen": 2489122816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482246740220662,
      "loss": 1.4135,
      "theoretical_loss": 3.398155379368403,
      "tokens_seen": 2489188352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024820461384152456,
      "loss": 1.3162,
      "theoretical_loss": 3.398148296626535,
      "tokens_seen": 2489253888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481845536609829,
      "loss": 1.3228,
      "theoretical_loss": 3.3981412141233465,
      "tokens_seen": 2489319424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024816449348044134,
      "loss": 1.2983,
      "theoretical_loss": 3.398134131858823,
      "tokens_seen": 2489384960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481444332998997,
      "loss": 1.37,
      "theoretical_loss": 3.39812704983295,
      "tokens_seen": 2489450496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024812437311935807,
      "loss": 1.3446,
      "theoretical_loss": 3.3981199680457137,
      "tokens_seen": 2489516032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481043129388165,
      "loss": 1.2452,
      "theoretical_loss": 3.398112886497099,
      "tokens_seen": 2489581568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024808425275827485,
      "loss": 1.3364,
      "theoretical_loss": 3.3981058051870923,
      "tokens_seen": 2489647104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002480641925777332,
      "loss": 1.2841,
      "theoretical_loss": 3.398098724115679,
      "tokens_seen": 2489712640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024804413239719157,
      "loss": 1.2488,
      "theoretical_loss": 3.3980916432828447,
      "tokens_seen": 2489778176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024802407221665,
      "loss": 1.2844,
      "theoretical_loss": 3.3980845626885756,
      "tokens_seen": 2489843712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024800401203610835,
      "loss": 1.425,
      "theoretical_loss": 3.398077482332856,
      "tokens_seen": 2489909248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002479839518555667,
      "loss": 1.4244,
      "theoretical_loss": 3.398070402215673,
      "tokens_seen": 2489974784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002479638916750251,
      "loss": 1.3454,
      "theoretical_loss": 3.398063322337012,
      "tokens_seen": 2490040320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024794383149448344,
      "loss": 1.2685,
      "theoretical_loss": 3.3980562426968586,
      "tokens_seen": 2490105856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024792377131394185,
      "loss": 1.315,
      "theoretical_loss": 3.3980491632951977,
      "tokens_seen": 2490171392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002479037111334002,
      "loss": 1.3586,
      "theoretical_loss": 3.3980420841320162,
      "tokens_seen": 2490236928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478836509528586,
      "loss": 1.2656,
      "theoretical_loss": 3.398035005207299,
      "tokens_seen": 2490302464
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024786359077231694,
      "loss": 1.2893,
      "theoretical_loss": 3.3980279265210322,
      "tokens_seen": 2490368000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024784353059177536,
      "loss": 1.3092,
      "theoretical_loss": 3.3980208480732013,
      "tokens_seen": 2490433536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478234704112337,
      "loss": 1.3621,
      "theoretical_loss": 3.398013769863792,
      "tokens_seen": 2490499072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478034102306921,
      "loss": 1.3002,
      "theoretical_loss": 3.39800669189279,
      "tokens_seen": 2490564608
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024778335005015044,
      "loss": 1.291,
      "theoretical_loss": 3.397999614160181,
      "tokens_seen": 2490630144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024776328986960886,
      "loss": 1.3411,
      "theoretical_loss": 3.397992536665951,
      "tokens_seen": 2490695680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477432296890672,
      "loss": 1.292,
      "theoretical_loss": 3.3979854594100853,
      "tokens_seen": 2490761216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477231695085256,
      "loss": 1.3295,
      "theoretical_loss": 3.3979783823925693,
      "tokens_seen": 2490826752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024770310932798395,
      "loss": 1.3149,
      "theoretical_loss": 3.39797130561339,
      "tokens_seen": 2490892288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476830491474423,
      "loss": 1.273,
      "theoretical_loss": 3.3979642290725316,
      "tokens_seen": 2490957824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476629889669007,
      "loss": 1.4394,
      "theoretical_loss": 3.3979571527699806,
      "tokens_seen": 2491023360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476429287863591,
      "loss": 1.2769,
      "theoretical_loss": 3.3979500767057225,
      "tokens_seen": 2491088896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024762286860581745,
      "loss": 1.2908,
      "theoretical_loss": 3.397943000879743,
      "tokens_seen": 2491154432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476028084252758,
      "loss": 1.3369,
      "theoretical_loss": 3.397935925292028,
      "tokens_seen": 2491219968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024758274824473423,
      "loss": 1.3049,
      "theoretical_loss": 3.397928849942563,
      "tokens_seen": 2491285504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475626880641926,
      "loss": 1.3192,
      "theoretical_loss": 3.3979217748313335,
      "tokens_seen": 2491351040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024754262788365095,
      "loss": 1.2839,
      "theoretical_loss": 3.397914699958326,
      "tokens_seen": 2491416576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475225677031093,
      "loss": 1.3115,
      "theoretical_loss": 3.3979076253235254,
      "tokens_seen": 2491482112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475025075225677,
      "loss": 1.4467,
      "theoretical_loss": 3.397900550926918,
      "tokens_seen": 2491547648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474824473420261,
      "loss": 1.3443,
      "theoretical_loss": 3.3978934767684885,
      "tokens_seen": 2491613184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024746238716148445,
      "loss": 1.2875,
      "theoretical_loss": 3.3978864028482243,
      "tokens_seen": 2491678720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474423269809428,
      "loss": 1.3229,
      "theoretical_loss": 3.3978793291661096,
      "tokens_seen": 2491744256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474222668004012,
      "loss": 1.2385,
      "theoretical_loss": 3.397872255722131,
      "tokens_seen": 2491809792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474022066198596,
      "loss": 1.3834,
      "theoretical_loss": 3.3978651825162736,
      "tokens_seen": 2491875328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024738214643931796,
      "loss": 1.2925,
      "theoretical_loss": 3.3978581095485234,
      "tokens_seen": 2491940864
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473620862587764,
      "loss": 1.3056,
      "theoretical_loss": 3.3978510368188664,
      "tokens_seen": 2492006400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024734202607823474,
      "loss": 1.285,
      "theoretical_loss": 3.3978439643272877,
      "tokens_seen": 2492071936
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 906987,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.148254156112671,
      "objective/train/theoretical_loss": 3.397836892073774,
      "objective/train/tokens_used": 862596576,
      "theoretical_loss": 3.397836892073774,
      "tokens_seen": 2492137472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473219658976931,
      "loss": 1.1904,
      "theoretical_loss": 3.397836892073774,
      "tokens_seen": 2492137472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024730190571715146,
      "loss": 1.323,
      "theoretical_loss": 3.39782982005831,
      "tokens_seen": 2492203008
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472818455366098,
      "loss": 1.5083,
      "theoretical_loss": 3.397822748280882,
      "tokens_seen": 2492268544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024726178535606824,
      "loss": 1.2224,
      "theoretical_loss": 3.3978156767414753,
      "tokens_seen": 2492334080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472417251755266,
      "loss": 1.372,
      "theoretical_loss": 3.3978086054400762,
      "tokens_seen": 2492399616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024722166499498496,
      "loss": 1.2508,
      "theoretical_loss": 3.3978015343766703,
      "tokens_seen": 2492465152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472016048144433,
      "loss": 1.275,
      "theoretical_loss": 3.397794463551243,
      "tokens_seen": 2492530688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024718154463390174,
      "loss": 1.2373,
      "theoretical_loss": 3.39778739296378,
      "tokens_seen": 2492596224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471614844533601,
      "loss": 1.2955,
      "theoretical_loss": 3.3977803226142673,
      "tokens_seen": 2492661760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024714142427281847,
      "loss": 1.2886,
      "theoretical_loss": 3.3977732525026907,
      "tokens_seen": 2492727296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024712136409227683,
      "loss": 1.2405,
      "theoretical_loss": 3.397766182629036,
      "tokens_seen": 2492792832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471013039117352,
      "loss": 1.3852,
      "theoretical_loss": 3.3977591129932883,
      "tokens_seen": 2492858368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470812437311936,
      "loss": 1.361,
      "theoretical_loss": 3.3977520435954345,
      "tokens_seen": 2492923904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024706118355065197,
      "loss": 1.1757,
      "theoretical_loss": 3.397744974435459,
      "tokens_seen": 2492989440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024704112337011033,
      "loss": 1.4564,
      "theoretical_loss": 3.3977379055133485,
      "tokens_seen": 2493054976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470210631895687,
      "loss": 1.3568,
      "theoretical_loss": 3.397730836829088,
      "tokens_seen": 2493120512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470010030090271,
      "loss": 1.3232,
      "theoretical_loss": 3.397723768382664,
      "tokens_seen": 2493186048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002469809428284855,
      "loss": 1.3393,
      "theoretical_loss": 3.3977167001740622,
      "tokens_seen": 2493251584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024696088264794384,
      "loss": 1.2738,
      "theoretical_loss": 3.397709632203268,
      "tokens_seen": 2493317120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002469408224674022,
      "loss": 1.2933,
      "theoretical_loss": 3.3977025644702668,
      "tokens_seen": 2493382656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024692076228686056,
      "loss": 1.2349,
      "theoretical_loss": 3.397695496975045,
      "tokens_seen": 2493448192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246900702106319,
      "loss": 1.3842,
      "theoretical_loss": 3.397688429717588,
      "tokens_seen": 2493513728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024688064192577734,
      "loss": 1.161,
      "theoretical_loss": 3.397681362697882,
      "tokens_seen": 2493579264
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468605817452357,
      "loss": 1.1998,
      "theoretical_loss": 3.3976742959159116,
      "tokens_seen": 2493644800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024684052156469406,
      "loss": 1.323,
      "theoretical_loss": 3.397667229371664,
      "tokens_seen": 2493710336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468204613841525,
      "loss": 1.1944,
      "theoretical_loss": 3.3976601630651246,
      "tokens_seen": 2493775872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024680040120361084,
      "loss": 1.217,
      "theoretical_loss": 3.3976530969962786,
      "tokens_seen": 2493841408
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467803410230692,
      "loss": 1.2876,
      "theoretical_loss": 3.3976460311651118,
      "tokens_seen": 2493906944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024676028084252757,
      "loss": 1.3305,
      "theoretical_loss": 3.3976389655716104,
      "tokens_seen": 2493972480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024674022066198593,
      "loss": 1.2719,
      "theoretical_loss": 3.3976319002157602,
      "tokens_seen": 2494038016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024672016048144435,
      "loss": 1.3139,
      "theoretical_loss": 3.397624835097546,
      "tokens_seen": 2494103552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467001003009027,
      "loss": 1.3514,
      "theoretical_loss": 3.3976177702169554,
      "tokens_seen": 2494169088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024668004012036107,
      "loss": 1.2691,
      "theoretical_loss": 3.3976107055739724,
      "tokens_seen": 2494234624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024665997993981943,
      "loss": 1.2836,
      "theoretical_loss": 3.3976036411685833,
      "tokens_seen": 2494300160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024663991975927785,
      "loss": 1.3656,
      "theoretical_loss": 3.3975965770007743,
      "tokens_seen": 2494365696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002466198595787362,
      "loss": 1.28,
      "theoretical_loss": 3.397589513070531,
      "tokens_seen": 2494431232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024659979939819463,
      "loss": 1.3324,
      "theoretical_loss": 3.3975824493778384,
      "tokens_seen": 2494496768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246579739217653,
      "loss": 1.3077,
      "theoretical_loss": 3.3975753859226834,
      "tokens_seen": 2494562304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024655967903711135,
      "loss": 1.3581,
      "theoretical_loss": 3.397568322705051,
      "tokens_seen": 2494627840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465396188565697,
      "loss": 1.3503,
      "theoretical_loss": 3.3975612597249274,
      "tokens_seen": 2494693376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465195586760281,
      "loss": 1.2005,
      "theoretical_loss": 3.3975541969822984,
      "tokens_seen": 2494758912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464994984954865,
      "loss": 1.3329,
      "theoretical_loss": 3.3975471344771493,
      "tokens_seen": 2494824448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024647943831494486,
      "loss": 1.1613,
      "theoretical_loss": 3.397540072209466,
      "tokens_seen": 2494889984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464593781344032,
      "loss": 1.3635,
      "theoretical_loss": 3.397533010179235,
      "tokens_seen": 2494955520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464393179538616,
      "loss": 1.3921,
      "theoretical_loss": 3.3975259483864413,
      "tokens_seen": 2495021056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024641925777332,
      "loss": 1.3385,
      "theoretical_loss": 3.3975188868310706,
      "tokens_seen": 2495086592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024639919759277836,
      "loss": 1.2526,
      "theoretical_loss": 3.397511825513109,
      "tokens_seen": 2495152128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463791374122367,
      "loss": 1.1815,
      "theoretical_loss": 3.397504764432543,
      "tokens_seen": 2495217664
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463590772316951,
      "loss": 1.2979,
      "theoretical_loss": 3.397497703589357,
      "tokens_seen": 2495283200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024633901705115345,
      "loss": 1.3793,
      "theoretical_loss": 3.397490642983538,
      "tokens_seen": 2495348736
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 908214,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2217057943344116,
      "objective/train/theoretical_loss": 3.3974835826150707,
      "objective/train/tokens_used": 865873376,
      "theoretical_loss": 3.3974835826150707,
      "tokens_seen": 2495414272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024631895687061186,
      "loss": 1.2668,
      "theoretical_loss": 3.3974835826150707,
      "tokens_seen": 2495414272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462988966900702,
      "loss": 1.2666,
      "theoretical_loss": 3.3974765224839416,
      "tokens_seen": 2495479808
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462788365095286,
      "loss": 1.2652,
      "theoretical_loss": 3.397469462590136,
      "tokens_seen": 2495545344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024625877632898695,
      "loss": 1.2481,
      "theoretical_loss": 3.3974624029336407,
      "tokens_seen": 2495610880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024623871614844536,
      "loss": 1.2666,
      "theoretical_loss": 3.39745534351444,
      "tokens_seen": 2495676416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462186559679037,
      "loss": 1.2957,
      "theoretical_loss": 3.397448284332521,
      "tokens_seen": 2495741952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461985957873621,
      "loss": 1.3204,
      "theoretical_loss": 3.397441225387869,
      "tokens_seen": 2495807488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024617853560682045,
      "loss": 1.4021,
      "theoretical_loss": 3.3974341666804695,
      "tokens_seen": 2495873024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461584754262788,
      "loss": 1.3076,
      "theoretical_loss": 3.3974271082103087,
      "tokens_seen": 2495938560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024613841524573723,
      "loss": 1.3235,
      "theoretical_loss": 3.397420049977372,
      "tokens_seen": 2496004096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461183550651956,
      "loss": 1.2981,
      "theoretical_loss": 3.397412991981646,
      "tokens_seen": 2496069632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024609829488465395,
      "loss": 1.3519,
      "theoretical_loss": 3.3974059342231158,
      "tokens_seen": 2496135168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460782347041123,
      "loss": 1.4067,
      "theoretical_loss": 3.397398876701767,
      "tokens_seen": 2496200704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024605817452357073,
      "loss": 1.3553,
      "theoretical_loss": 3.397391819417586,
      "tokens_seen": 2496266240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460381143430291,
      "loss": 1.4182,
      "theoretical_loss": 3.3973847623705584,
      "tokens_seen": 2496331776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024601805416248746,
      "loss": 1.2988,
      "theoretical_loss": 3.39737770556067,
      "tokens_seen": 2496397312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459979939819458,
      "loss": 1.2761,
      "theoretical_loss": 3.3973706489879065,
      "tokens_seen": 2496462848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459779338014042,
      "loss": 1.2759,
      "theoretical_loss": 3.397363592652254,
      "tokens_seen": 2496528384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459578736208626,
      "loss": 1.3149,
      "theoretical_loss": 3.3973565365536977,
      "tokens_seen": 2496593920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024593781344032096,
      "loss": 1.266,
      "theoretical_loss": 3.397349480692224,
      "tokens_seen": 2496659456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459177532597793,
      "loss": 1.4351,
      "theoretical_loss": 3.3973424250678184,
      "tokens_seen": 2496724992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458976930792377,
      "loss": 1.2678,
      "theoretical_loss": 3.397335369680467,
      "tokens_seen": 2496790528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458776328986961,
      "loss": 1.2867,
      "theoretical_loss": 3.3973283145301556,
      "tokens_seen": 2496856064
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024585757271815446,
      "loss": 1.359,
      "theoretical_loss": 3.39732125961687,
      "tokens_seen": 2496921600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458375125376129,
      "loss": 1.3574,
      "theoretical_loss": 3.3973142049405953,
      "tokens_seen": 2496987136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024581745235707124,
      "loss": 1.362,
      "theoretical_loss": 3.397307150501318,
      "tokens_seen": 2497052672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457973921765296,
      "loss": 1.2774,
      "theoretical_loss": 3.3973000962990243,
      "tokens_seen": 2497118208
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024577733199598797,
      "loss": 1.257,
      "theoretical_loss": 3.397293042333699,
      "tokens_seen": 2497183744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024575727181544633,
      "loss": 1.2144,
      "theoretical_loss": 3.397285988605329,
      "tokens_seen": 2497249280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024573721163490475,
      "loss": 1.2676,
      "theoretical_loss": 3.3972789351138992,
      "tokens_seen": 2497314816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457171514543631,
      "loss": 1.4925,
      "theoretical_loss": 3.397271881859396,
      "tokens_seen": 2497380352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024569709127382147,
      "loss": 1.4294,
      "theoretical_loss": 3.397264828841805,
      "tokens_seen": 2497445888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024567703109327983,
      "loss": 1.357,
      "theoretical_loss": 3.3972577760611116,
      "tokens_seen": 2497511424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024565697091273825,
      "loss": 1.3072,
      "theoretical_loss": 3.3972507235173026,
      "tokens_seen": 2497576960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002456369107321966,
      "loss": 1.4013,
      "theoretical_loss": 3.397243671210363,
      "tokens_seen": 2497642496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245616850551655,
      "loss": 1.2713,
      "theoretical_loss": 3.397236619140279,
      "tokens_seen": 2497708032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024559679037111334,
      "loss": 1.2894,
      "theoretical_loss": 3.3972295673070363,
      "tokens_seen": 2497773568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455767301905717,
      "loss": 1.3014,
      "theoretical_loss": 3.397222515710621,
      "tokens_seen": 2497839104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455566700100301,
      "loss": 1.3756,
      "theoretical_loss": 3.3972154643510186,
      "tokens_seen": 2497904640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455366098294885,
      "loss": 1.303,
      "theoretical_loss": 3.397208413228215,
      "tokens_seen": 2497970176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024551654964894684,
      "loss": 1.2442,
      "theoretical_loss": 3.397201362342196,
      "tokens_seen": 2498035712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454964894684052,
      "loss": 1.2907,
      "theoretical_loss": 3.397194311692948,
      "tokens_seen": 2498101248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454764292878636,
      "loss": 1.3495,
      "theoretical_loss": 3.3971872612804566,
      "tokens_seen": 2498166784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245456369107322,
      "loss": 1.3344,
      "theoretical_loss": 3.3971802111047067,
      "tokens_seen": 2498232320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024543630892678034,
      "loss": 1.3165,
      "theoretical_loss": 3.397173161165685,
      "tokens_seen": 2498297856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454162487462387,
      "loss": 1.2856,
      "theoretical_loss": 3.3971661114633775,
      "tokens_seen": 2498363392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024539618856569707,
      "loss": 1.2515,
      "theoretical_loss": 3.3971590619977694,
      "tokens_seen": 2498428928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002453761283851555,
      "loss": 1.1653,
      "theoretical_loss": 3.397152012768847,
      "tokens_seen": 2498494464
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024535606820461385,
      "loss": 1.3751,
      "theoretical_loss": 3.3971449637765962,
      "tokens_seen": 2498560000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002453360080240722,
      "loss": 1.3467,
      "theoretical_loss": 3.3971379150210024,
      "tokens_seen": 2498625536
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 909311,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1775472164154053,
      "objective/train/theoretical_loss": 3.397130866502052,
      "objective/train/tokens_used": 869150176,
      "theoretical_loss": 3.397130866502052,
      "tokens_seen": 2498691072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024531594784353057,
      "loss": 1.243,
      "theoretical_loss": 3.397130866502052,
      "tokens_seen": 2498691072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245295887662989,
      "loss": 1.3678,
      "theoretical_loss": 3.3971238182197303,
      "tokens_seen": 2498756608
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024527582748244735,
      "loss": 1.2951,
      "theoretical_loss": 3.3971167701740237,
      "tokens_seen": 2498822144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002452557673019057,
      "loss": 1.1966,
      "theoretical_loss": 3.3971097223649176,
      "tokens_seen": 2498887680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002452357071213641,
      "loss": 1.3145,
      "theoretical_loss": 3.397102674792398,
      "tokens_seen": 2498953216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024521564694082244,
      "loss": 1.3083,
      "theoretical_loss": 3.3970956274564506,
      "tokens_seen": 2499018752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024519558676028085,
      "loss": 1.4205,
      "theoretical_loss": 3.3970885803570616,
      "tokens_seen": 2499084288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451755265797392,
      "loss": 1.1542,
      "theoretical_loss": 3.397081533494217,
      "tokens_seen": 2499149824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451554663991976,
      "loss": 1.44,
      "theoretical_loss": 3.397074486867902,
      "tokens_seen": 2499215360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024513540621865594,
      "loss": 1.4321,
      "theoretical_loss": 3.3970674404781027,
      "tokens_seen": 2499280896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024511534603811435,
      "loss": 1.3213,
      "theoretical_loss": 3.3970603943248054,
      "tokens_seen": 2499346432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002450952858575727,
      "loss": 1.2879,
      "theoretical_loss": 3.3970533484079954,
      "tokens_seen": 2499411968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024507522567703113,
      "loss": 1.3041,
      "theoretical_loss": 3.397046302727659,
      "tokens_seen": 2499477504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002450551654964895,
      "loss": 1.2256,
      "theoretical_loss": 3.3970392572837818,
      "tokens_seen": 2499543040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024503510531594786,
      "loss": 1.2722,
      "theoretical_loss": 3.3970322120763496,
      "tokens_seen": 2499608576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002450150451354062,
      "loss": 1.3751,
      "theoretical_loss": 3.3970251671053484,
      "tokens_seen": 2499674112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002449949849548646,
      "loss": 1.318,
      "theoretical_loss": 3.3970181223707643,
      "tokens_seen": 2499739648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000244974924774323,
      "loss": 1.2402,
      "theoretical_loss": 3.3970110778725826,
      "tokens_seen": 2499805184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024495486459378136,
      "loss": 1.3678,
      "theoretical_loss": 3.3970040336107896,
      "tokens_seen": 2499870720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002449348044132397,
      "loss": 1.2922,
      "theoretical_loss": 3.396996989585371,
      "tokens_seen": 2499936256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002449147442326981,
      "loss": 1.2874,
      "theoretical_loss": 3.396989945796313,
      "tokens_seen": 2500001792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448946840521565,
      "loss": 1.2734,
      "theoretical_loss": 3.3969829022436007,
      "tokens_seen": 2500067328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024487462387161486,
      "loss": 1.189,
      "theoretical_loss": 3.3969758589272208,
      "tokens_seen": 2500132864
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448545636910732,
      "loss": 1.2168,
      "theoretical_loss": 3.396968815847159,
      "tokens_seen": 2500198400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448345035105316,
      "loss": 1.3108,
      "theoretical_loss": 3.396961773003401,
      "tokens_seen": 2500263936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024481444332998995,
      "loss": 1.3047,
      "theoretical_loss": 3.396954730395932,
      "tokens_seen": 2500329472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024479438314944837,
      "loss": 1.2748,
      "theoretical_loss": 3.3969476880247393,
      "tokens_seen": 2500395008
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024477432296890673,
      "loss": 1.2455,
      "theoretical_loss": 3.396940645889808,
      "tokens_seen": 2500460544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002447542627883651,
      "loss": 1.3413,
      "theoretical_loss": 3.3969336039911244,
      "tokens_seen": 2500526080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024473420260782345,
      "loss": 1.1791,
      "theoretical_loss": 3.3969265623286735,
      "tokens_seen": 2500591616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024471414242728187,
      "loss": 1.2259,
      "theoretical_loss": 3.3969195209024416,
      "tokens_seen": 2500657152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024469408224674023,
      "loss": 1.4091,
      "theoretical_loss": 3.396912479712415,
      "tokens_seen": 2500722688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446740220661986,
      "loss": 1.3666,
      "theoretical_loss": 3.3969054387585795,
      "tokens_seen": 2500788224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024465396188565696,
      "loss": 1.2633,
      "theoretical_loss": 3.3968983980409204,
      "tokens_seen": 2500853760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446339017051154,
      "loss": 1.1678,
      "theoretical_loss": 3.396891357559424,
      "tokens_seen": 2500919296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024461384152457374,
      "loss": 1.3439,
      "theoretical_loss": 3.3968843173140764,
      "tokens_seen": 2500984832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445937813440321,
      "loss": 1.3315,
      "theoretical_loss": 3.3968772773048634,
      "tokens_seen": 2501050368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024457372116349046,
      "loss": 1.3109,
      "theoretical_loss": 3.39687023753177,
      "tokens_seen": 2501115904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445536609829488,
      "loss": 1.3383,
      "theoretical_loss": 3.396863197994784,
      "tokens_seen": 2501181440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024453360080240724,
      "loss": 1.2633,
      "theoretical_loss": 3.396856158693889,
      "tokens_seen": 2501246976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445135406218656,
      "loss": 1.2642,
      "theoretical_loss": 3.3968491196290724,
      "tokens_seen": 2501312512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024449348044132396,
      "loss": 1.3262,
      "theoretical_loss": 3.3968420808003197,
      "tokens_seen": 2501378048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002444734202607823,
      "loss": 1.2824,
      "theoretical_loss": 3.3968350422076172,
      "tokens_seen": 2501443584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024445336008024074,
      "loss": 1.3248,
      "theoretical_loss": 3.39682800385095,
      "tokens_seen": 2501509120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002444332998996991,
      "loss": 1.2704,
      "theoretical_loss": 3.3968209657303046,
      "tokens_seen": 2501574656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024441323971915747,
      "loss": 1.3109,
      "theoretical_loss": 3.396813927845667,
      "tokens_seen": 2501640192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024439317953861583,
      "loss": 1.2614,
      "theoretical_loss": 3.396806890197022,
      "tokens_seen": 2501705728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443731193580742,
      "loss": 1.3496,
      "theoretical_loss": 3.396799852784357,
      "tokens_seen": 2501771264
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443530591775326,
      "loss": 1.3203,
      "theoretical_loss": 3.396792815607657,
      "tokens_seen": 2501836800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024433299899699097,
      "loss": 1.2492,
      "theoretical_loss": 3.3967857786669082,
      "tokens_seen": 2501902336
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 910572,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3356267213821411,
      "objective/train/theoretical_loss": 3.3967787419620965,
      "objective/train/tokens_used": 872426976,
      "theoretical_loss": 3.3967787419620965,
      "tokens_seen": 2501967872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443129388164494,
      "loss": 1.2816,
      "theoretical_loss": 3.3967787419620965,
      "tokens_seen": 2501967872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024429287863590775,
      "loss": 1.2678,
      "theoretical_loss": 3.396771705493208,
      "tokens_seen": 2502033408
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442728184553661,
      "loss": 1.2394,
      "theoretical_loss": 3.3967646692602282,
      "tokens_seen": 2502098944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442527582748245,
      "loss": 1.2657,
      "theoretical_loss": 3.3967576332631433,
      "tokens_seen": 2502164480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024423269809428284,
      "loss": 1.1568,
      "theoretical_loss": 3.3967505975019385,
      "tokens_seen": 2502230016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024421263791374125,
      "loss": 1.3706,
      "theoretical_loss": 3.396743561976601,
      "tokens_seen": 2502295552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002441925777331996,
      "loss": 1.2704,
      "theoretical_loss": 3.396736526687116,
      "tokens_seen": 2502361088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000244172517552658,
      "loss": 1.2622,
      "theoretical_loss": 3.396729491633469,
      "tokens_seen": 2502426624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024415245737211634,
      "loss": 1.2789,
      "theoretical_loss": 3.396722456815647,
      "tokens_seen": 2502492160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024413239719157473,
      "loss": 1.2568,
      "theoretical_loss": 3.3967154222336347,
      "tokens_seen": 2502557696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024411233701103312,
      "loss": 1.2607,
      "theoretical_loss": 3.396708387887419,
      "tokens_seen": 2502623232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024409227683049148,
      "loss": 1.1742,
      "theoretical_loss": 3.3967013537769852,
      "tokens_seen": 2502688768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024407221664994987,
      "loss": 1.2899,
      "theoretical_loss": 3.3966943199023194,
      "tokens_seen": 2502754304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024405215646940823,
      "loss": 1.341,
      "theoretical_loss": 3.3966872862634077,
      "tokens_seen": 2502819840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002440320962888666,
      "loss": 1.1911,
      "theoretical_loss": 3.396680252860236,
      "tokens_seen": 2502885376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024401203610832498,
      "loss": 1.4833,
      "theoretical_loss": 3.39667321969279,
      "tokens_seen": 2502950912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024399197592778335,
      "loss": 1.245,
      "theoretical_loss": 3.396666186761056,
      "tokens_seen": 2503016448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024397191574724173,
      "loss": 1.2568,
      "theoretical_loss": 3.3966591540650195,
      "tokens_seen": 2503081984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002439518555667001,
      "loss": 1.2373,
      "theoretical_loss": 3.3966521216046663,
      "tokens_seen": 2503147520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024393179538615849,
      "loss": 1.3586,
      "theoretical_loss": 3.396645089379983,
      "tokens_seen": 2503213056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024391173520561685,
      "loss": 1.345,
      "theoretical_loss": 3.3966380573909554,
      "tokens_seen": 2503278592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024389167502507524,
      "loss": 1.304,
      "theoretical_loss": 3.3966310256375687,
      "tokens_seen": 2503344128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002438716148445336,
      "loss": 1.3118,
      "theoretical_loss": 3.3966239941198095,
      "tokens_seen": 2503409664
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024385155466399196,
      "loss": 1.2612,
      "theoretical_loss": 3.3966169628376637,
      "tokens_seen": 2503475200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024383149448345035,
      "loss": 1.1515,
      "theoretical_loss": 3.3966099317911174,
      "tokens_seen": 2503540736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024381143430290871,
      "loss": 1.3744,
      "theoretical_loss": 3.396602900980156,
      "tokens_seen": 2503606272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002437913741223671,
      "loss": 1.3877,
      "theoretical_loss": 3.3965958704047656,
      "tokens_seen": 2503671808
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024377131394182547,
      "loss": 1.3577,
      "theoretical_loss": 3.3965888400649322,
      "tokens_seen": 2503737344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024375125376128385,
      "loss": 1.3073,
      "theoretical_loss": 3.396581809960642,
      "tokens_seen": 2503802880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024373119358074222,
      "loss": 1.182,
      "theoretical_loss": 3.3965747800918806,
      "tokens_seen": 2503868416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002437111334002006,
      "loss": 1.1259,
      "theoretical_loss": 3.396567750458634,
      "tokens_seen": 2503933952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024369107321965897,
      "loss": 1.2835,
      "theoretical_loss": 3.3965607210608884,
      "tokens_seen": 2503999488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024367101303911736,
      "loss": 1.3467,
      "theoretical_loss": 3.3965536918986294,
      "tokens_seen": 2504065024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024365095285857572,
      "loss": 1.2089,
      "theoretical_loss": 3.3965466629718435,
      "tokens_seen": 2504130560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024363089267803408,
      "loss": 1.3103,
      "theoretical_loss": 3.396539634280516,
      "tokens_seen": 2504196096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002436108324974925,
      "loss": 1.1838,
      "theoretical_loss": 3.396532605824633,
      "tokens_seen": 2504261632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024359077231695086,
      "loss": 1.3384,
      "theoretical_loss": 3.3965255776041805,
      "tokens_seen": 2504327168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024357071213640925,
      "loss": 1.3258,
      "theoretical_loss": 3.3965185496191443,
      "tokens_seen": 2504392704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002435506519558676,
      "loss": 1.2457,
      "theoretical_loss": 3.3965115218695114,
      "tokens_seen": 2504458240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000243530591775326,
      "loss": 1.2626,
      "theoretical_loss": 3.3965044943552662,
      "tokens_seen": 2504523776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024351053159478436,
      "loss": 1.3249,
      "theoretical_loss": 3.396497467076396,
      "tokens_seen": 2504589312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024349047141424275,
      "loss": 1.2892,
      "theoretical_loss": 3.3964904400328857,
      "tokens_seen": 2504654848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024347041123370112,
      "loss": 1.258,
      "theoretical_loss": 3.396483413224722,
      "tokens_seen": 2504720384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024345035105315948,
      "loss": 1.223,
      "theoretical_loss": 3.3964763866518903,
      "tokens_seen": 2504785920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024343029087261787,
      "loss": 1.3423,
      "theoretical_loss": 3.3964693603143767,
      "tokens_seen": 2504851456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024341023069207623,
      "loss": 1.231,
      "theoretical_loss": 3.396462334212168,
      "tokens_seen": 2504916992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024339017051153462,
      "loss": 1.2587,
      "theoretical_loss": 3.3964553083452484,
      "tokens_seen": 2504982528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024337011033099298,
      "loss": 1.2877,
      "theoretical_loss": 3.3964482827136058,
      "tokens_seen": 2505048064
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024335005015045137,
      "loss": 1.1973,
      "theoretical_loss": 3.3964412573172247,
      "tokens_seen": 2505113600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024332998996990973,
      "loss": 1.189,
      "theoretical_loss": 3.396434232156092,
      "tokens_seen": 2505179136
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 911801,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2049455642700195,
      "objective/train/theoretical_loss": 3.396427207230193,
      "objective/train/tokens_used": 875703776,
      "theoretical_loss": 3.396427207230193,
      "tokens_seen": 2505244672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024330992978936812,
      "loss": 1.255,
      "theoretical_loss": 3.396427207230193,
      "tokens_seen": 2505244672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024328986960882648,
      "loss": 1.3085,
      "theoretical_loss": 3.3964201825395146,
      "tokens_seen": 2505310208
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024326980942828487,
      "loss": 1.3258,
      "theoretical_loss": 3.3964131580840418,
      "tokens_seen": 2505375744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024324974924774324,
      "loss": 1.2469,
      "theoretical_loss": 3.396406133863761,
      "tokens_seen": 2505441280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002432296890672016,
      "loss": 1.277,
      "theoretical_loss": 3.396399109878658,
      "tokens_seen": 2505506816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024320962888666,
      "loss": 1.2912,
      "theoretical_loss": 3.396392086128719,
      "tokens_seen": 2505572352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024318956870611835,
      "loss": 1.1993,
      "theoretical_loss": 3.3963850626139296,
      "tokens_seen": 2505637888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024316950852557674,
      "loss": 1.3634,
      "theoretical_loss": 3.396378039334276,
      "tokens_seen": 2505703424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002431494483450351,
      "loss": 1.3874,
      "theoretical_loss": 3.3963710162897445,
      "tokens_seen": 2505768960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002431293881644935,
      "loss": 1.3488,
      "theoretical_loss": 3.3963639934803207,
      "tokens_seen": 2505834496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024310932798395185,
      "loss": 1.3272,
      "theoretical_loss": 3.3963569709059906,
      "tokens_seen": 2505900032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024308926780341024,
      "loss": 1.3915,
      "theoretical_loss": 3.3963499485667406,
      "tokens_seen": 2505965568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002430692076228686,
      "loss": 1.2889,
      "theoretical_loss": 3.396342926462556,
      "tokens_seen": 2506031104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024304914744232697,
      "loss": 1.1886,
      "theoretical_loss": 3.396335904593423,
      "tokens_seen": 2506096640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024302908726178536,
      "loss": 1.2166,
      "theoretical_loss": 3.3963288829593283,
      "tokens_seen": 2506162176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024300902708124372,
      "loss": 1.2082,
      "theoretical_loss": 3.396321861560257,
      "tokens_seen": 2506227712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002429889669007021,
      "loss": 1.3306,
      "theoretical_loss": 3.396314840396195,
      "tokens_seen": 2506293248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024296890672016047,
      "loss": 1.2791,
      "theoretical_loss": 3.396307819467129,
      "tokens_seen": 2506358784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024294884653961886,
      "loss": 1.3418,
      "theoretical_loss": 3.3963007987730447,
      "tokens_seen": 2506424320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024292878635907722,
      "loss": 1.2467,
      "theoretical_loss": 3.396293778313928,
      "tokens_seen": 2506489856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002429087261785356,
      "loss": 1.3428,
      "theoretical_loss": 3.396286758089765,
      "tokens_seen": 2506555392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024288866599799397,
      "loss": 1.1408,
      "theoretical_loss": 3.3962797381005414,
      "tokens_seen": 2506620928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024286860581745234,
      "loss": 1.3193,
      "theoretical_loss": 3.396272718346244,
      "tokens_seen": 2506686464
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024284854563691075,
      "loss": 1.2052,
      "theoretical_loss": 3.396265698826858,
      "tokens_seen": 2506752000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024282848545636911,
      "loss": 1.2707,
      "theoretical_loss": 3.3962586795423695,
      "tokens_seen": 2506817536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002428084252758275,
      "loss": 1.3113,
      "theoretical_loss": 3.3962516604927644,
      "tokens_seen": 2506883072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024278836509528587,
      "loss": 1.2016,
      "theoretical_loss": 3.396244641678029,
      "tokens_seen": 2506948608
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024276830491474426,
      "loss": 1.2756,
      "theoretical_loss": 3.3962376230981497,
      "tokens_seen": 2507014144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024274824473420262,
      "loss": 1.2539,
      "theoretical_loss": 3.3962306047531117,
      "tokens_seen": 2507079680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000242728184553661,
      "loss": 1.3405,
      "theoretical_loss": 3.3962235866429014,
      "tokens_seen": 2507145216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024270812437311937,
      "loss": 1.2471,
      "theoretical_loss": 3.3962165687675046,
      "tokens_seen": 2507210752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024268806419257776,
      "loss": 1.3524,
      "theoretical_loss": 3.3962095511269075,
      "tokens_seen": 2507276288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024266800401203612,
      "loss": 1.1967,
      "theoretical_loss": 3.3962025337210964,
      "tokens_seen": 2507341824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024264794383149448,
      "loss": 1.4228,
      "theoretical_loss": 3.3961955165500566,
      "tokens_seen": 2507407360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024262788365095287,
      "loss": 1.2913,
      "theoretical_loss": 3.396188499613775,
      "tokens_seen": 2507472896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024260782347041123,
      "loss": 1.2601,
      "theoretical_loss": 3.3961814829122363,
      "tokens_seen": 2507538432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024258776328986962,
      "loss": 1.3025,
      "theoretical_loss": 3.3961744664454274,
      "tokens_seen": 2507603968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024256770310932799,
      "loss": 1.2971,
      "theoretical_loss": 3.3961674502133348,
      "tokens_seen": 2507669504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024254764292878638,
      "loss": 1.1873,
      "theoretical_loss": 3.3961604342159437,
      "tokens_seen": 2507735040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024252758274824474,
      "loss": 1.3039,
      "theoretical_loss": 3.39615341845324,
      "tokens_seen": 2507800576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024250752256770313,
      "loss": 1.2806,
      "theoretical_loss": 3.3961464029252104,
      "tokens_seen": 2507866112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002424874623871615,
      "loss": 1.4146,
      "theoretical_loss": 3.3961393876318406,
      "tokens_seen": 2507931648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024246740220661985,
      "loss": 1.2456,
      "theoretical_loss": 3.3961323725731165,
      "tokens_seen": 2507997184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024244734202607824,
      "loss": 1.2534,
      "theoretical_loss": 3.3961253577490242,
      "tokens_seen": 2508062720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002424272818455366,
      "loss": 1.2875,
      "theoretical_loss": 3.3961183431595496,
      "tokens_seen": 2508128256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000242407221664995,
      "loss": 1.267,
      "theoretical_loss": 3.396111328804679,
      "tokens_seen": 2508193792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024238716148445335,
      "loss": 1.241,
      "theoretical_loss": 3.3961043146843983,
      "tokens_seen": 2508259328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024236710130391174,
      "loss": 1.3231,
      "theoretical_loss": 3.3960973007986937,
      "tokens_seen": 2508324864
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002423470411233701,
      "loss": 1.3257,
      "theoretical_loss": 3.3960902871475507,
      "tokens_seen": 2508390400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002423269809428285,
      "loss": 1.2854,
      "theoretical_loss": 3.3960832737309556,
      "tokens_seen": 2508455936
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 912960,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2665480375289917,
      "objective/train/theoretical_loss": 3.3960762605488948,
      "objective/train/tokens_used": 878980576,
      "theoretical_loss": 3.3960762605488948,
      "tokens_seen": 2508521472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024230692076228686,
      "loss": 1.3118,
      "theoretical_loss": 3.3960762605488948,
      "tokens_seen": 2508521472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024228686058174522,
      "loss": 1.2865,
      "theoretical_loss": 3.396069247601354,
      "tokens_seen": 2508587008
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002422668004012036,
      "loss": 1.2653,
      "theoretical_loss": 3.396062234888319,
      "tokens_seen": 2508652544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024224674022066197,
      "loss": 1.2423,
      "theoretical_loss": 3.3960552224097764,
      "tokens_seen": 2508718080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024222668004012036,
      "loss": 1.2996,
      "theoretical_loss": 3.396048210165712,
      "tokens_seen": 2508783616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024220661985957872,
      "loss": 1.1542,
      "theoretical_loss": 3.3960411981561114,
      "tokens_seen": 2508849152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002421865596790371,
      "loss": 1.2604,
      "theoretical_loss": 3.396034186380961,
      "tokens_seen": 2508914688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024216649949849547,
      "loss": 1.2149,
      "theoretical_loss": 3.3960271748402473,
      "tokens_seen": 2508980224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024214643931795386,
      "loss": 1.3407,
      "theoretical_loss": 3.3960201635339553,
      "tokens_seen": 2509045760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024212637913741223,
      "loss": 1.2862,
      "theoretical_loss": 3.396013152462072,
      "tokens_seen": 2509111296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024210631895687064,
      "loss": 1.2362,
      "theoretical_loss": 3.3960061416245826,
      "tokens_seen": 2509176832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000242086258776329,
      "loss": 1.3402,
      "theoretical_loss": 3.395999131021474,
      "tokens_seen": 2509242368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024206619859578737,
      "loss": 1.2018,
      "theoretical_loss": 3.3959921206527315,
      "tokens_seen": 2509307904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024204613841524576,
      "loss": 1.4198,
      "theoretical_loss": 3.3959851105183416,
      "tokens_seen": 2509373440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024202607823470412,
      "loss": 1.3456,
      "theoretical_loss": 3.3959781006182905,
      "tokens_seen": 2509438976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002420060180541625,
      "loss": 1.2254,
      "theoretical_loss": 3.395971090952564,
      "tokens_seen": 2509504512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024198595787362087,
      "loss": 1.3242,
      "theoretical_loss": 3.395964081521148,
      "tokens_seen": 2509570048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024196589769307926,
      "loss": 1.258,
      "theoretical_loss": 3.395957072324028,
      "tokens_seen": 2509635584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024194583751253762,
      "loss": 1.2676,
      "theoretical_loss": 3.3959500633611914,
      "tokens_seen": 2509701120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000241925777331996,
      "loss": 1.2625,
      "theoretical_loss": 3.3959430546326237,
      "tokens_seen": 2509766656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024190571715145437,
      "loss": 1.3453,
      "theoretical_loss": 3.3959360461383103,
      "tokens_seen": 2509832192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024188565697091274,
      "loss": 1.1878,
      "theoretical_loss": 3.395929037878238,
      "tokens_seen": 2509897728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024186559679037113,
      "loss": 1.1101,
      "theoretical_loss": 3.395922029852393,
      "tokens_seen": 2509963264
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002418455366098295,
      "loss": 1.265,
      "theoretical_loss": 3.395915022060761,
      "tokens_seen": 2510028800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024182547642928788,
      "loss": 1.3557,
      "theoretical_loss": 3.3959080145033274,
      "tokens_seen": 2510094336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024180541624874624,
      "loss": 1.209,
      "theoretical_loss": 3.395901007180079,
      "tokens_seen": 2510159872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024178535606820463,
      "loss": 1.2192,
      "theoretical_loss": 3.395894000091002,
      "tokens_seen": 2510225408
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000241765295887663,
      "loss": 1.3312,
      "theoretical_loss": 3.3958869932360822,
      "tokens_seen": 2510290944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024174523570712138,
      "loss": 1.2973,
      "theoretical_loss": 3.3958799866153058,
      "tokens_seen": 2510356480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024172517552657974,
      "loss": 1.1978,
      "theoretical_loss": 3.3958729802286585,
      "tokens_seen": 2510422016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024170511534603813,
      "loss": 1.2592,
      "theoretical_loss": 3.3958659740761266,
      "tokens_seen": 2510487552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002416850551654965,
      "loss": 1.2362,
      "theoretical_loss": 3.3958589681576963,
      "tokens_seen": 2510553088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024166499498495486,
      "loss": 1.2761,
      "theoretical_loss": 3.395851962473354,
      "tokens_seen": 2510618624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024164493480441325,
      "loss": 1.2353,
      "theoretical_loss": 3.3958449570230846,
      "tokens_seen": 2510684160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002416248746238716,
      "loss": 1.2603,
      "theoretical_loss": 3.3958379518068753,
      "tokens_seen": 2510749696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024160481444333,
      "loss": 1.2835,
      "theoretical_loss": 3.3958309468247116,
      "tokens_seen": 2510815232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024158475426278836,
      "loss": 1.3098,
      "theoretical_loss": 3.3958239420765794,
      "tokens_seen": 2510880768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024156469408224675,
      "loss": 1.2894,
      "theoretical_loss": 3.395816937562466,
      "tokens_seen": 2510946304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002415446339017051,
      "loss": 1.2669,
      "theoretical_loss": 3.3958099332823557,
      "tokens_seen": 2511011840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002415245737211635,
      "loss": 1.3402,
      "theoretical_loss": 3.3958029292362357,
      "tokens_seen": 2511077376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024150451354062186,
      "loss": 1.2344,
      "theoretical_loss": 3.395795925424092,
      "tokens_seen": 2511142912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024148445336008022,
      "loss": 1.2514,
      "theoretical_loss": 3.3957889218459103,
      "tokens_seen": 2511208448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024146439317953861,
      "loss": 1.3256,
      "theoretical_loss": 3.3957819185016773,
      "tokens_seen": 2511273984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024144433299899698,
      "loss": 1.3806,
      "theoretical_loss": 3.395774915391378,
      "tokens_seen": 2511339520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024142427281845537,
      "loss": 1.2892,
      "theoretical_loss": 3.3957679125149998,
      "tokens_seen": 2511405056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024140421263791373,
      "loss": 1.3458,
      "theoretical_loss": 3.3957609098725277,
      "tokens_seen": 2511470592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024138415245737212,
      "loss": 1.3419,
      "theoretical_loss": 3.3957539074639485,
      "tokens_seen": 2511536128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002413640922768305,
      "loss": 1.2786,
      "theoretical_loss": 3.395746905289248,
      "tokens_seen": 2511601664
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002413440320962889,
      "loss": 1.2827,
      "theoretical_loss": 3.395739903348412,
      "tokens_seen": 2511667200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024132397191574726,
      "loss": 1.2896,
      "theoretical_loss": 3.395732901641427,
      "tokens_seen": 2511732736
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 914176,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3319772481918335,
      "objective/train/theoretical_loss": 3.395725900168279,
      "objective/train/tokens_used": 882257376,
      "theoretical_loss": 3.395725900168279,
      "tokens_seen": 2511798272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024130391173520562,
      "loss": 1.3164,
      "theoretical_loss": 3.395725900168279,
      "tokens_seen": 2511798272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000241283851554664,
      "loss": 1.4194,
      "theoretical_loss": 3.395718898928954,
      "tokens_seen": 2511863808
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024126379137412237,
      "loss": 1.3157,
      "theoretical_loss": 3.395711897923438,
      "tokens_seen": 2511929344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024124373119358076,
      "loss": 1.2739,
      "theoretical_loss": 3.3957048971517176,
      "tokens_seen": 2511994880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024122367101303912,
      "loss": 1.4492,
      "theoretical_loss": 3.395697896613778,
      "tokens_seen": 2512060416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002412036108324975,
      "loss": 1.3712,
      "theoretical_loss": 3.395690896309606,
      "tokens_seen": 2512125952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024118355065195587,
      "loss": 1.325,
      "theoretical_loss": 3.3956838962391878,
      "tokens_seen": 2512191488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024116349047141426,
      "loss": 1.2067,
      "theoretical_loss": 3.395676896402509,
      "tokens_seen": 2512257024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024114343029087263,
      "loss": 1.224,
      "theoretical_loss": 3.395669896799556,
      "tokens_seen": 2512322560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024112337011033102,
      "loss": 1.2227,
      "theoretical_loss": 3.3956628974303147,
      "tokens_seen": 2512388096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024110330992978938,
      "loss": 1.2728,
      "theoretical_loss": 3.3956558982947715,
      "tokens_seen": 2512453632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024108324974924774,
      "loss": 1.2426,
      "theoretical_loss": 3.395648899392912,
      "tokens_seen": 2512519168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024106318956870613,
      "loss": 1.2978,
      "theoretical_loss": 3.395641900724723,
      "tokens_seen": 2512584704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002410431293881645,
      "loss": 1.257,
      "theoretical_loss": 3.39563490229019,
      "tokens_seen": 2512650240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024102306920762288,
      "loss": 1.3434,
      "theoretical_loss": 3.3956279040892996,
      "tokens_seen": 2512715776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024100300902708124,
      "loss": 1.4236,
      "theoretical_loss": 3.395620906122037,
      "tokens_seen": 2512781312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024098294884653963,
      "loss": 1.3639,
      "theoretical_loss": 3.3956139083883894,
      "tokens_seen": 2512846848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000240962888665998,
      "loss": 1.3421,
      "theoretical_loss": 3.3956069108883424,
      "tokens_seen": 2512912384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024094282848545638,
      "loss": 1.2332,
      "theoretical_loss": 3.395599913621882,
      "tokens_seen": 2512977920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024092276830491475,
      "loss": 1.1876,
      "theoretical_loss": 3.3955929165889946,
      "tokens_seen": 2513043456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002409027081243731,
      "loss": 1.3529,
      "theoretical_loss": 3.3955859197896663,
      "tokens_seen": 2513108992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002408826479438315,
      "loss": 1.362,
      "theoretical_loss": 3.3955789232238827,
      "tokens_seen": 2513174528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024086258776328986,
      "loss": 1.3722,
      "theoretical_loss": 3.395571926891631,
      "tokens_seen": 2513240064
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024084252758274825,
      "loss": 1.2987,
      "theoretical_loss": 3.395564930792896,
      "tokens_seen": 2513305600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002408224674022066,
      "loss": 1.3985,
      "theoretical_loss": 3.395557934927665,
      "tokens_seen": 2513371136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000240802407221665,
      "loss": 1.1562,
      "theoretical_loss": 3.395550939295923,
      "tokens_seen": 2513436672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024078234704112336,
      "loss": 1.3234,
      "theoretical_loss": 3.395543943897657,
      "tokens_seen": 2513502208
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024076228686058175,
      "loss": 1.3481,
      "theoretical_loss": 3.395536948732852,
      "tokens_seen": 2513567744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024074222668004012,
      "loss": 1.2679,
      "theoretical_loss": 3.395529953801496,
      "tokens_seen": 2513633280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024072216649949848,
      "loss": 1.3197,
      "theoretical_loss": 3.3955229591035736,
      "tokens_seen": 2513698816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024070210631895687,
      "loss": 1.1287,
      "theoretical_loss": 3.3955159646390713,
      "tokens_seen": 2513764352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024068204613841523,
      "loss": 1.3856,
      "theoretical_loss": 3.3955089704079753,
      "tokens_seen": 2513829888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024066198595787362,
      "loss": 1.3083,
      "theoretical_loss": 3.395501976410272,
      "tokens_seen": 2513895424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024064192577733198,
      "loss": 1.3458,
      "theoretical_loss": 3.395494982645947,
      "tokens_seen": 2513960960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024062186559679037,
      "loss": 1.2991,
      "theoretical_loss": 3.3954879891149865,
      "tokens_seen": 2514026496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024060180541624876,
      "loss": 1.3509,
      "theoretical_loss": 3.395480995817377,
      "tokens_seen": 2514092032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024058174523570715,
      "loss": 1.223,
      "theoretical_loss": 3.3954740027531045,
      "tokens_seen": 2514157568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002405616850551655,
      "loss": 1.2719,
      "theoretical_loss": 3.395467009922155,
      "tokens_seen": 2514223104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002405416248746239,
      "loss": 1.1867,
      "theoretical_loss": 3.395460017324515,
      "tokens_seen": 2514288640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024052156469408226,
      "loss": 1.3767,
      "theoretical_loss": 3.39545302496017,
      "tokens_seen": 2514354176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024050150451354062,
      "loss": 1.2269,
      "theoretical_loss": 3.3954460328291067,
      "tokens_seen": 2514419712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024048144433299901,
      "loss": 1.2896,
      "theoretical_loss": 3.395439040931311,
      "tokens_seen": 2514485248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024046138415245738,
      "loss": 1.3091,
      "theoretical_loss": 3.395432049266769,
      "tokens_seen": 2514550784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024044132397191577,
      "loss": 1.3215,
      "theoretical_loss": 3.3954250578354666,
      "tokens_seen": 2514616320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024042126379137413,
      "loss": 1.2738,
      "theoretical_loss": 3.3954180666373905,
      "tokens_seen": 2514681856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024040120361083252,
      "loss": 1.2625,
      "theoretical_loss": 3.395411075672526,
      "tokens_seen": 2514747392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024038114343029088,
      "loss": 1.244,
      "theoretical_loss": 3.3954040849408607,
      "tokens_seen": 2514812928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024036108324974927,
      "loss": 1.2857,
      "theoretical_loss": 3.3953970944423797,
      "tokens_seen": 2514878464
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024034102306920763,
      "loss": 1.2971,
      "theoretical_loss": 3.395390104177069,
      "tokens_seen": 2514944000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000240320962888666,
      "loss": 1.2655,
      "theoretical_loss": 3.395383114144915,
      "tokens_seen": 2515009536
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 915361,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.244124174118042,
      "objective/train/theoretical_loss": 3.395376124345904,
      "objective/train/tokens_used": 885534176,
      "theoretical_loss": 3.395376124345904,
      "tokens_seen": 2515075072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024030090270812438,
      "loss": 1.2465,
      "theoretical_loss": 3.395376124345904,
      "tokens_seen": 2515075072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024028084252758274,
      "loss": 1.2714,
      "theoretical_loss": 3.3953691347800223,
      "tokens_seen": 2515140608
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024026078234704113,
      "loss": 1.249,
      "theoretical_loss": 3.3953621454472556,
      "tokens_seen": 2515206144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002402407221664995,
      "loss": 1.2827,
      "theoretical_loss": 3.3953551563475903,
      "tokens_seen": 2515271680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024022066198595789,
      "loss": 1.1938,
      "theoretical_loss": 3.3953481674810124,
      "tokens_seen": 2515337216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024020060180541625,
      "loss": 1.1646,
      "theoretical_loss": 3.395341178847508,
      "tokens_seen": 2515402752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024018054162487464,
      "loss": 1.3404,
      "theoretical_loss": 3.3953341904470635,
      "tokens_seen": 2515468288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000240160481444333,
      "loss": 1.1998,
      "theoretical_loss": 3.3953272022796654,
      "tokens_seen": 2515533824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024014042126379136,
      "loss": 1.3259,
      "theoretical_loss": 3.395320214345299,
      "tokens_seen": 2515599360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024012036108324975,
      "loss": 1.3914,
      "theoretical_loss": 3.395313226643951,
      "tokens_seen": 2515664896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024010030090270811,
      "loss": 1.3053,
      "theoretical_loss": 3.3953062391756075,
      "tokens_seen": 2515730432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002400802407221665,
      "loss": 1.3496,
      "theoretical_loss": 3.3952992519402545,
      "tokens_seen": 2515795968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024006018054162487,
      "loss": 1.2803,
      "theoretical_loss": 3.3952922649378783,
      "tokens_seen": 2515861504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024004012036108325,
      "loss": 1.3036,
      "theoretical_loss": 3.395285278168465,
      "tokens_seen": 2515927040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024002006018054162,
      "loss": 1.3084,
      "theoretical_loss": 3.395278291632001,
      "tokens_seen": 2515992576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024,
      "loss": 1.2416,
      "theoretical_loss": 3.395271305328472,
      "tokens_seen": 2516058112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00023997993981945837,
      "loss": 1.3358,
      "theoretical_loss": 3.3952643192578646,
      "tokens_seen": 2516123648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00023995987963891676,
      "loss": 1.3276,
      "theoretical_loss": 3.395257333420165,
      "tokens_seen": 2516189184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023993981945837512,
      "loss": 1.2792,
      "theoretical_loss": 3.3952503478153586,
      "tokens_seen": 2516254720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023991975927783348,
      "loss": 1.3809,
      "theoretical_loss": 3.3952433624434324,
      "tokens_seen": 2516320256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023989969909729187,
      "loss": 1.1999,
      "theoretical_loss": 3.3952363773043723,
      "tokens_seen": 2516385792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023987963891675023,
      "loss": 1.1577,
      "theoretical_loss": 3.3952293923981647,
      "tokens_seen": 2516451328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023985957873620865,
      "loss": 1.3156,
      "theoretical_loss": 3.3952224077247957,
      "tokens_seen": 2516516864
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239839518555667,
      "loss": 1.2288,
      "theoretical_loss": 3.3952154232842506,
      "tokens_seen": 2516582400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002398194583751254,
      "loss": 1.3285,
      "theoretical_loss": 3.3952084390765167,
      "tokens_seen": 2516647936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023979939819458376,
      "loss": 1.3744,
      "theoretical_loss": 3.3952014551015797,
      "tokens_seen": 2516713472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023977933801404215,
      "loss": 1.2413,
      "theoretical_loss": 3.395194471359426,
      "tokens_seen": 2516779008
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023975927783350052,
      "loss": 1.2416,
      "theoretical_loss": 3.3951874878500417,
      "tokens_seen": 2516844544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023973921765295888,
      "loss": 1.2554,
      "theoretical_loss": 3.395180504573413,
      "tokens_seen": 2516910080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023971915747241727,
      "loss": 1.3108,
      "theoretical_loss": 3.3951735215295256,
      "tokens_seen": 2516975616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023969909729187563,
      "loss": 1.2718,
      "theoretical_loss": 3.3951665387183665,
      "tokens_seen": 2517041152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023967903711133402,
      "loss": 1.2574,
      "theoretical_loss": 3.3951595561399213,
      "tokens_seen": 2517106688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023965897693079238,
      "loss": 1.252,
      "theoretical_loss": 3.395152573794176,
      "tokens_seen": 2517172224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023963891675025077,
      "loss": 1.2349,
      "theoretical_loss": 3.3951455916811177,
      "tokens_seen": 2517237760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023961885656970913,
      "loss": 1.3255,
      "theoretical_loss": 3.395138609800732,
      "tokens_seen": 2517303296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023959879638916752,
      "loss": 1.3004,
      "theoretical_loss": 3.395131628153005,
      "tokens_seen": 2517368832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023957873620862588,
      "loss": 1.305,
      "theoretical_loss": 3.395124646737923,
      "tokens_seen": 2517434368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023955867602808427,
      "loss": 1.2907,
      "theoretical_loss": 3.3951176655554716,
      "tokens_seen": 2517499904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023953861584754264,
      "loss": 1.4413,
      "theoretical_loss": 3.3951106846056383,
      "tokens_seen": 2517565440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239518555667001,
      "loss": 1.2821,
      "theoretical_loss": 3.3951037038884087,
      "tokens_seen": 2517630976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002394984954864594,
      "loss": 1.4294,
      "theoretical_loss": 3.395096723403768,
      "tokens_seen": 2517696512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023947843530591775,
      "loss": 1.2547,
      "theoretical_loss": 3.395089743151704,
      "tokens_seen": 2517762048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023945837512537614,
      "loss": 1.3577,
      "theoretical_loss": 3.3950827631322023,
      "tokens_seen": 2517827584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002394383149448345,
      "loss": 1.3869,
      "theoretical_loss": 3.3950757833452485,
      "tokens_seen": 2517893120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002394182547642929,
      "loss": 1.2721,
      "theoretical_loss": 3.3950688037908296,
      "tokens_seen": 2517958656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023939819458375125,
      "loss": 1.3566,
      "theoretical_loss": 3.3950618244689315,
      "tokens_seen": 2518024192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023937813440320964,
      "loss": 1.2574,
      "theoretical_loss": 3.3950548453795406,
      "tokens_seen": 2518089728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239358074222668,
      "loss": 1.2537,
      "theoretical_loss": 3.3950478665226425,
      "tokens_seen": 2518155264
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023933801404212637,
      "loss": 1.2403,
      "theoretical_loss": 3.3950408878982237,
      "tokens_seen": 2518220800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023931795386158476,
      "loss": 1.4048,
      "theoretical_loss": 3.395033909506271,
      "tokens_seen": 2518286336
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 916685,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2623939514160156,
      "objective/train/theoretical_loss": 3.395026931346769,
      "objective/train/tokens_used": 888810976,
      "theoretical_loss": 3.395026931346769,
      "tokens_seen": 2518351872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023929789368104312,
      "loss": 1.2771,
      "theoretical_loss": 3.395026931346769,
      "tokens_seen": 2518351872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002392778335005015,
      "loss": 1.2298,
      "theoretical_loss": 3.3950199534197063,
      "tokens_seen": 2518417408
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023925777331995987,
      "loss": 1.1146,
      "theoretical_loss": 3.3950129757250673,
      "tokens_seen": 2518482944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023923771313941826,
      "loss": 1.283,
      "theoretical_loss": 3.3950059982628384,
      "tokens_seen": 2518548480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023921765295887662,
      "loss": 1.2702,
      "theoretical_loss": 3.3949990210330068,
      "tokens_seen": 2518614016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239197592778335,
      "loss": 1.2369,
      "theoretical_loss": 3.3949920440355577,
      "tokens_seen": 2518679552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023917753259779337,
      "loss": 1.229,
      "theoretical_loss": 3.3949850672704773,
      "tokens_seen": 2518745088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023915747241725174,
      "loss": 1.3067,
      "theoretical_loss": 3.394978090737753,
      "tokens_seen": 2518810624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023913741223671012,
      "loss": 1.359,
      "theoretical_loss": 3.3949711144373698,
      "tokens_seen": 2518876160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002391173520561685,
      "loss": 1.3256,
      "theoretical_loss": 3.394964138369314,
      "tokens_seen": 2518941696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002390972918756269,
      "loss": 1.2956,
      "theoretical_loss": 3.3949571625335726,
      "tokens_seen": 2519007232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023907723169508527,
      "loss": 1.2952,
      "theoretical_loss": 3.394950186930131,
      "tokens_seen": 2519072768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023905717151454365,
      "loss": 1.2274,
      "theoretical_loss": 3.394943211558976,
      "tokens_seen": 2519138304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023903711133400202,
      "loss": 1.2334,
      "theoretical_loss": 3.3949362364200937,
      "tokens_seen": 2519203840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002390170511534604,
      "loss": 1.2963,
      "theoretical_loss": 3.39492926151347,
      "tokens_seen": 2519269376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023899699097291877,
      "loss": 1.3311,
      "theoretical_loss": 3.3949222868390914,
      "tokens_seen": 2519334912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023897693079237716,
      "loss": 1.3855,
      "theoretical_loss": 3.394915312396944,
      "tokens_seen": 2519400448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023895687061183552,
      "loss": 1.3365,
      "theoretical_loss": 3.3949083381870144,
      "tokens_seen": 2519465984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023893681043129388,
      "loss": 1.2198,
      "theoretical_loss": 3.394901364209288,
      "tokens_seen": 2519531520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023891675025075227,
      "loss": 1.3557,
      "theoretical_loss": 3.394894390463752,
      "tokens_seen": 2519597056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023889669007021063,
      "loss": 1.3104,
      "theoretical_loss": 3.394887416950392,
      "tokens_seen": 2519662592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023887662988966902,
      "loss": 1.2074,
      "theoretical_loss": 3.3948804436691944,
      "tokens_seen": 2519728128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023885656970912739,
      "loss": 1.4014,
      "theoretical_loss": 3.3948734706201456,
      "tokens_seen": 2519793664
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023883650952858577,
      "loss": 1.3765,
      "theoretical_loss": 3.394866497803232,
      "tokens_seen": 2519859200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023881644934804414,
      "loss": 1.2059,
      "theoretical_loss": 3.394859525218439,
      "tokens_seen": 2519924736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023879638916750253,
      "loss": 1.2375,
      "theoretical_loss": 3.3948525528657534,
      "tokens_seen": 2519990272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002387763289869609,
      "loss": 1.189,
      "theoretical_loss": 3.3948455807451614,
      "tokens_seen": 2520055808
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023875626880641925,
      "loss": 1.3316,
      "theoretical_loss": 3.3948386088566496,
      "tokens_seen": 2520121344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023873620862587764,
      "loss": 1.2478,
      "theoretical_loss": 3.3948316372002036,
      "tokens_seen": 2520186880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000238716148445336,
      "loss": 1.2374,
      "theoretical_loss": 3.3948246657758094,
      "tokens_seen": 2520252416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386960882647944,
      "loss": 1.3467,
      "theoretical_loss": 3.3948176945834545,
      "tokens_seen": 2520317952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023867602808425275,
      "loss": 1.4208,
      "theoretical_loss": 3.394810723623124,
      "tokens_seen": 2520383488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023865596790371114,
      "loss": 1.1307,
      "theoretical_loss": 3.3948037528948047,
      "tokens_seen": 2520449024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386359077231695,
      "loss": 1.3776,
      "theoretical_loss": 3.394796782398483,
      "tokens_seen": 2520514560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386158475426279,
      "loss": 1.2455,
      "theoretical_loss": 3.3947898121341447,
      "tokens_seen": 2520580096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023859578736208626,
      "loss": 1.3522,
      "theoretical_loss": 3.3947828421017756,
      "tokens_seen": 2520645632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023857572718154462,
      "loss": 1.3456,
      "theoretical_loss": 3.3947758723013632,
      "tokens_seen": 2520711168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000238555667001003,
      "loss": 1.3175,
      "theoretical_loss": 3.394768902732893,
      "tokens_seen": 2520776704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023853560682046137,
      "loss": 1.2439,
      "theoretical_loss": 3.394761933396351,
      "tokens_seen": 2520842240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023851554663991976,
      "loss": 1.3561,
      "theoretical_loss": 3.394754964291724,
      "tokens_seen": 2520907776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023849548645937812,
      "loss": 1.2372,
      "theoretical_loss": 3.3947479954189976,
      "tokens_seen": 2520973312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002384754262788365,
      "loss": 1.2974,
      "theoretical_loss": 3.394741026778159,
      "tokens_seen": 2521038848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023845536609829487,
      "loss": 1.3473,
      "theoretical_loss": 3.394734058369194,
      "tokens_seen": 2521104384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023843530591775326,
      "loss": 1.2912,
      "theoretical_loss": 3.394727090192089,
      "tokens_seen": 2521169920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023841524573721163,
      "loss": 1.368,
      "theoretical_loss": 3.3947201222468295,
      "tokens_seen": 2521235456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023839518555667002,
      "loss": 1.2389,
      "theoretical_loss": 3.3947131545334024,
      "tokens_seen": 2521300992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023837512537612838,
      "loss": 1.2624,
      "theoretical_loss": 3.3947061870517943,
      "tokens_seen": 2521366528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023835506519558677,
      "loss": 1.4093,
      "theoretical_loss": 3.394699219801991,
      "tokens_seen": 2521432064
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023833500501504516,
      "loss": 1.2918,
      "theoretical_loss": 3.3946922527839782,
      "tokens_seen": 2521497600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023831494483450352,
      "loss": 1.3125,
      "theoretical_loss": 3.3946852859977437,
      "tokens_seen": 2521563136
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 917863,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2736694812774658,
      "objective/train/theoretical_loss": 3.394678319443272,
      "objective/train/tokens_used": 892087776,
      "theoretical_loss": 3.394678319443272,
      "tokens_seen": 2521628672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382948846539619,
      "loss": 1.2443,
      "theoretical_loss": 3.394678319443272,
      "tokens_seen": 2521628672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023827482447342027,
      "loss": 1.3449,
      "theoretical_loss": 3.3946713531205504,
      "tokens_seen": 2521694208
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023825476429287866,
      "loss": 1.2789,
      "theoretical_loss": 3.3946643870295654,
      "tokens_seen": 2521759744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023823470411233702,
      "loss": 1.2904,
      "theoretical_loss": 3.3946574211703027,
      "tokens_seen": 2521825280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382146439317954,
      "loss": 1.3102,
      "theoretical_loss": 3.3946504555427484,
      "tokens_seen": 2521890816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023819458375125377,
      "loss": 1.3501,
      "theoretical_loss": 3.3946434901468896,
      "tokens_seen": 2521956352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023817452357071214,
      "loss": 1.2194,
      "theoretical_loss": 3.3946365249827117,
      "tokens_seen": 2522021888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023815446339017052,
      "loss": 1.3184,
      "theoretical_loss": 3.3946295600502014,
      "tokens_seen": 2522087424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002381344032096289,
      "loss": 1.2719,
      "theoretical_loss": 3.394622595349345,
      "tokens_seen": 2522152960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023811434302908728,
      "loss": 1.3278,
      "theoretical_loss": 3.3946156308801285,
      "tokens_seen": 2522218496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023809428284854564,
      "loss": 1.3833,
      "theoretical_loss": 3.3946086666425384,
      "tokens_seen": 2522284032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023807422266800403,
      "loss": 1.2255,
      "theoretical_loss": 3.394601702636561,
      "tokens_seen": 2522349568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002380541624874624,
      "loss": 1.2968,
      "theoretical_loss": 3.3945947388621827,
      "tokens_seen": 2522415104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023803410230692078,
      "loss": 1.3707,
      "theoretical_loss": 3.394587775319389,
      "tokens_seen": 2522480640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023801404212637914,
      "loss": 1.3019,
      "theoretical_loss": 3.3945808120081677,
      "tokens_seen": 2522546176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023799398194583753,
      "loss": 1.2885,
      "theoretical_loss": 3.394573848928504,
      "tokens_seen": 2522611712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002379739217652959,
      "loss": 1.3273,
      "theoretical_loss": 3.3945668860803835,
      "tokens_seen": 2522677248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023795386158475426,
      "loss": 1.3853,
      "theoretical_loss": 3.394559923463794,
      "tokens_seen": 2522742784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023793380140421265,
      "loss": 1.3099,
      "theoretical_loss": 3.3945529610787215,
      "tokens_seen": 2522808320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000237913741223671,
      "loss": 1.3246,
      "theoretical_loss": 3.3945459989251514,
      "tokens_seen": 2522873856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378936810431294,
      "loss": 1.1867,
      "theoretical_loss": 3.3945390370030704,
      "tokens_seen": 2522939392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023787362086258776,
      "loss": 1.2945,
      "theoretical_loss": 3.394532075312465,
      "tokens_seen": 2523004928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023785356068204615,
      "loss": 1.3469,
      "theoretical_loss": 3.3945251138533212,
      "tokens_seen": 2523070464
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378335005015045,
      "loss": 1.258,
      "theoretical_loss": 3.394518152625626,
      "tokens_seen": 2523136000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378134403209629,
      "loss": 1.2601,
      "theoretical_loss": 3.394511191629365,
      "tokens_seen": 2523201536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023779338014042126,
      "loss": 1.2274,
      "theoretical_loss": 3.3945042308645244,
      "tokens_seen": 2523267072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023777331995987962,
      "loss": 1.3227,
      "theoretical_loss": 3.394497270331091,
      "tokens_seen": 2523332608
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023775325977933801,
      "loss": 1.246,
      "theoretical_loss": 3.3944903100290507,
      "tokens_seen": 2523398144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023773319959879638,
      "loss": 1.3277,
      "theoretical_loss": 3.39448334995839,
      "tokens_seen": 2523463680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023771313941825477,
      "loss": 1.2785,
      "theoretical_loss": 3.3944763901190953,
      "tokens_seen": 2523529216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023769307923771313,
      "loss": 1.4128,
      "theoretical_loss": 3.394469430511153,
      "tokens_seen": 2523594752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023767301905717152,
      "loss": 1.3937,
      "theoretical_loss": 3.3944624711345486,
      "tokens_seen": 2523660288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023765295887662988,
      "loss": 1.2275,
      "theoretical_loss": 3.394455511989269,
      "tokens_seen": 2523725824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023763289869608827,
      "loss": 1.3531,
      "theoretical_loss": 3.394448553075301,
      "tokens_seen": 2523791360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023761283851554663,
      "loss": 1.2761,
      "theoretical_loss": 3.39444159439263,
      "tokens_seen": 2523856896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023759277833500502,
      "loss": 1.3607,
      "theoretical_loss": 3.3944346359412423,
      "tokens_seen": 2523922432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002375727181544634,
      "loss": 1.3944,
      "theoretical_loss": 3.3944276777211253,
      "tokens_seen": 2523987968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023755265797392177,
      "loss": 1.2724,
      "theoretical_loss": 3.3944207197322642,
      "tokens_seen": 2524053504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023753259779338016,
      "loss": 1.3558,
      "theoretical_loss": 3.3944137619746457,
      "tokens_seen": 2524119040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023751253761283852,
      "loss": 1.2184,
      "theoretical_loss": 3.3944068044482565,
      "tokens_seen": 2524184576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002374924774322969,
      "loss": 1.3499,
      "theoretical_loss": 3.394399847153082,
      "tokens_seen": 2524250112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023747241725175527,
      "loss": 1.3423,
      "theoretical_loss": 3.3943928900891094,
      "tokens_seen": 2524315648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023745235707121366,
      "loss": 1.342,
      "theoretical_loss": 3.3943859332563244,
      "tokens_seen": 2524381184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023743229689067203,
      "loss": 1.3008,
      "theoretical_loss": 3.394378976654714,
      "tokens_seen": 2524446720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023741223671013042,
      "loss": 1.3382,
      "theoretical_loss": 3.3943720202842638,
      "tokens_seen": 2524512256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023739217652958878,
      "loss": 1.2082,
      "theoretical_loss": 3.39436506414496,
      "tokens_seen": 2524577792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023737211634904714,
      "loss": 1.3622,
      "theoretical_loss": 3.3943581082367897,
      "tokens_seen": 2524643328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023735205616850553,
      "loss": 1.539,
      "theoretical_loss": 3.3943511525597394,
      "tokens_seen": 2524708864
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002373319959879639,
      "loss": 1.3451,
      "theoretical_loss": 3.394344197113794,
      "tokens_seen": 2524774400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023731193580742228,
      "loss": 1.3228,
      "theoretical_loss": 3.394337241898941,
      "tokens_seen": 2524839936
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 918983,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3001699447631836,
      "objective/train/theoretical_loss": 3.394330286915167,
      "objective/train/tokens_used": 895364576,
      "theoretical_loss": 3.394330286915167,
      "tokens_seen": 2524905472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023729187562688064,
      "loss": 1.3305,
      "theoretical_loss": 3.394330286915167,
      "tokens_seen": 2524905472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023727181544633903,
      "loss": 1.3807,
      "theoretical_loss": 3.394323332162457,
      "tokens_seen": 2524971008
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002372517552657974,
      "loss": 1.3349,
      "theoretical_loss": 3.3943163776407985,
      "tokens_seen": 2525036544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023723169508525578,
      "loss": 1.3456,
      "theoretical_loss": 3.394309423350177,
      "tokens_seen": 2525102080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023721163490471415,
      "loss": 1.3051,
      "theoretical_loss": 3.3943024692905794,
      "tokens_seen": 2525167616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002371915747241725,
      "loss": 1.4065,
      "theoretical_loss": 3.394295515461992,
      "tokens_seen": 2525233152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002371715145436309,
      "loss": 1.3745,
      "theoretical_loss": 3.3942885618644008,
      "tokens_seen": 2525298688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023715145436308926,
      "loss": 1.3283,
      "theoretical_loss": 3.3942816084977925,
      "tokens_seen": 2525364224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023713139418254765,
      "loss": 1.3107,
      "theoretical_loss": 3.394274655362153,
      "tokens_seen": 2525429760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000237111334002006,
      "loss": 1.314,
      "theoretical_loss": 3.3942677024574692,
      "tokens_seen": 2525495296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002370912738214644,
      "loss": 1.3154,
      "theoretical_loss": 3.3942607497837267,
      "tokens_seen": 2525560832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023707121364092276,
      "loss": 1.2463,
      "theoretical_loss": 3.3942537973409124,
      "tokens_seen": 2525626368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023705115346038115,
      "loss": 1.2468,
      "theoretical_loss": 3.394246845129013,
      "tokens_seen": 2525691904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023703109327983952,
      "loss": 1.3303,
      "theoretical_loss": 3.3942398931480136,
      "tokens_seen": 2525757440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023701103309929788,
      "loss": 1.387,
      "theoretical_loss": 3.3942329413979015,
      "tokens_seen": 2525822976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023699097291875627,
      "loss": 1.325,
      "theoretical_loss": 3.394225989878663,
      "tokens_seen": 2525888512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023697091273821463,
      "loss": 1.3172,
      "theoretical_loss": 3.3942190385902844,
      "tokens_seen": 2525954048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023695085255767302,
      "loss": 1.2369,
      "theoretical_loss": 3.3942120875327513,
      "tokens_seen": 2526019584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023693079237713138,
      "loss": 1.457,
      "theoretical_loss": 3.3942051367060513,
      "tokens_seen": 2526085120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023691073219658977,
      "loss": 1.2633,
      "theoretical_loss": 3.3941981861101698,
      "tokens_seen": 2526150656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023689067201604813,
      "loss": 1.246,
      "theoretical_loss": 3.3941912357450934,
      "tokens_seen": 2526216192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023687061183550652,
      "loss": 1.2951,
      "theoretical_loss": 3.394184285610809,
      "tokens_seen": 2526281728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002368505516549649,
      "loss": 1.3114,
      "theoretical_loss": 3.3941773357073015,
      "tokens_seen": 2526347264
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002368304914744233,
      "loss": 1.2443,
      "theoretical_loss": 3.394170386034559,
      "tokens_seen": 2526412800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023681043129388166,
      "loss": 1.306,
      "theoretical_loss": 3.3941634365925664,
      "tokens_seen": 2526478336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023679037111334002,
      "loss": 1.2768,
      "theoretical_loss": 3.394156487381311,
      "tokens_seen": 2526543872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023677031093279841,
      "loss": 1.3769,
      "theoretical_loss": 3.394149538400779,
      "tokens_seen": 2526609408
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023675025075225678,
      "loss": 1.27,
      "theoretical_loss": 3.3941425896509565,
      "tokens_seen": 2526674944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023673019057171517,
      "loss": 1.341,
      "theoretical_loss": 3.3941356411318298,
      "tokens_seen": 2526740480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023671013039117353,
      "loss": 1.3168,
      "theoretical_loss": 3.3941286928433856,
      "tokens_seen": 2526806016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023669007021063192,
      "loss": 1.3324,
      "theoretical_loss": 3.3941217447856102,
      "tokens_seen": 2526871552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023667001003009028,
      "loss": 1.2687,
      "theoretical_loss": 3.39411479695849,
      "tokens_seen": 2526937088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023664994984954867,
      "loss": 1.292,
      "theoretical_loss": 3.3941078493620105,
      "tokens_seen": 2527002624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023662988966900703,
      "loss": 1.2878,
      "theoretical_loss": 3.3941009019961594,
      "tokens_seen": 2527068160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002366098294884654,
      "loss": 1.3061,
      "theoretical_loss": 3.394093954860922,
      "tokens_seen": 2527133696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023658976930792378,
      "loss": 1.3248,
      "theoretical_loss": 3.3940870079562853,
      "tokens_seen": 2527199232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023656970912738214,
      "loss": 1.3349,
      "theoretical_loss": 3.3940800612822355,
      "tokens_seen": 2527264768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023654964894684053,
      "loss": 1.1914,
      "theoretical_loss": 3.3940731148387586,
      "tokens_seen": 2527330304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002365295887662989,
      "loss": 1.3029,
      "theoretical_loss": 3.3940661686258418,
      "tokens_seen": 2527395840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023650952858575729,
      "loss": 1.3017,
      "theoretical_loss": 3.3940592226434707,
      "tokens_seen": 2527461376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023648946840521565,
      "loss": 1.2516,
      "theoretical_loss": 3.3940522768916317,
      "tokens_seen": 2527526912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023646940822467404,
      "loss": 1.3855,
      "theoretical_loss": 3.394045331370312,
      "tokens_seen": 2527592448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002364493480441324,
      "loss": 1.1429,
      "theoretical_loss": 3.394038386079497,
      "tokens_seen": 2527657984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023642928786359076,
      "loss": 1.2563,
      "theoretical_loss": 3.394031441019173,
      "tokens_seen": 2527723520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023640922768304915,
      "loss": 1.3341,
      "theoretical_loss": 3.3940244961893278,
      "tokens_seen": 2527789056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002363891675025075,
      "loss": 1.2451,
      "theoretical_loss": 3.394017551589946,
      "tokens_seen": 2527854592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002363691073219659,
      "loss": 1.3222,
      "theoretical_loss": 3.3940106072210154,
      "tokens_seen": 2527920128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023634904714142426,
      "loss": 1.2938,
      "theoretical_loss": 3.394003663082521,
      "tokens_seen": 2527985664
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023632898696088265,
      "loss": 1.2194,
      "theoretical_loss": 3.3939967191744507,
      "tokens_seen": 2528051200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023630892678034102,
      "loss": 1.3052,
      "theoretical_loss": 3.39398977549679,
      "tokens_seen": 2528116736
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 920165,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1944116353988647,
      "objective/train/theoretical_loss": 3.3939828320495247,
      "objective/train/tokens_used": 898641376,
      "theoretical_loss": 3.3939828320495247,
      "tokens_seen": 2528182272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002362888665997994,
      "loss": 1.19,
      "theoretical_loss": 3.3939828320495247,
      "tokens_seen": 2528182272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023626880641925777,
      "loss": 1.2717,
      "theoretical_loss": 3.3939758888326423,
      "tokens_seen": 2528247808
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023624874623871616,
      "loss": 1.4383,
      "theoretical_loss": 3.393968945846129,
      "tokens_seen": 2528313344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023622868605817452,
      "loss": 1.3235,
      "theoretical_loss": 3.3939620030899706,
      "tokens_seen": 2528378880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023620862587763288,
      "loss": 1.1821,
      "theoretical_loss": 3.393955060564154,
      "tokens_seen": 2528444416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023618856569709127,
      "loss": 1.3457,
      "theoretical_loss": 3.393948118268666,
      "tokens_seen": 2528509952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023616850551654963,
      "loss": 1.2929,
      "theoretical_loss": 3.3939411762034917,
      "tokens_seen": 2528575488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023614844533600802,
      "loss": 1.292,
      "theoretical_loss": 3.393934234368618,
      "tokens_seen": 2528641024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023612838515546639,
      "loss": 1.3213,
      "theoretical_loss": 3.3939272927640323,
      "tokens_seen": 2528706560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002361083249749248,
      "loss": 1.235,
      "theoretical_loss": 3.3939203513897196,
      "tokens_seen": 2528772096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023608826479438316,
      "loss": 1.3484,
      "theoretical_loss": 3.393913410245667,
      "tokens_seen": 2528837632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023606820461384155,
      "loss": 1.2588,
      "theoretical_loss": 3.393906469331861,
      "tokens_seen": 2528903168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023604814443329992,
      "loss": 1.3215,
      "theoretical_loss": 3.393899528648287,
      "tokens_seen": 2528968704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023602808425275828,
      "loss": 1.3271,
      "theoretical_loss": 3.3938925881949333,
      "tokens_seen": 2529034240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023600802407221667,
      "loss": 1.1908,
      "theoretical_loss": 3.3938856479717847,
      "tokens_seen": 2529099776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023598796389167503,
      "loss": 1.2718,
      "theoretical_loss": 3.393878707978828,
      "tokens_seen": 2529165312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023596790371113342,
      "loss": 1.302,
      "theoretical_loss": 3.3938717682160497,
      "tokens_seen": 2529230848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023594784353059178,
      "loss": 1.3865,
      "theoretical_loss": 3.393864828683436,
      "tokens_seen": 2529296384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023592778335005017,
      "loss": 1.3541,
      "theoretical_loss": 3.393857889380974,
      "tokens_seen": 2529361920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023590772316950853,
      "loss": 1.3214,
      "theoretical_loss": 3.393850950308649,
      "tokens_seen": 2529427456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023588766298896692,
      "loss": 1.3339,
      "theoretical_loss": 3.393844011466448,
      "tokens_seen": 2529492992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023586760280842528,
      "loss": 1.3494,
      "theoretical_loss": 3.3938370728543577,
      "tokens_seen": 2529558528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023584754262788367,
      "loss": 1.2793,
      "theoretical_loss": 3.3938301344723643,
      "tokens_seen": 2529624064
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023582748244734204,
      "loss": 1.2974,
      "theoretical_loss": 3.3938231963204535,
      "tokens_seen": 2529689600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002358074222668004,
      "loss": 1.3017,
      "theoretical_loss": 3.393816258398613,
      "tokens_seen": 2529755136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357873620862588,
      "loss": 1.3097,
      "theoretical_loss": 3.3938093207068283,
      "tokens_seen": 2529820672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023576730190571715,
      "loss": 1.229,
      "theoretical_loss": 3.3938023832450854,
      "tokens_seen": 2529886208
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023574724172517554,
      "loss": 1.3584,
      "theoretical_loss": 3.3937954460133724,
      "tokens_seen": 2529951744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357271815446339,
      "loss": 1.2374,
      "theoretical_loss": 3.393788509011674,
      "tokens_seen": 2530017280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357071213640923,
      "loss": 1.3936,
      "theoretical_loss": 3.393781572239978,
      "tokens_seen": 2530082816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023568706118355065,
      "loss": 1.2647,
      "theoretical_loss": 3.393774635698269,
      "tokens_seen": 2530148352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023566700100300904,
      "loss": 1.3391,
      "theoretical_loss": 3.3937676993865353,
      "tokens_seen": 2530213888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002356469408224674,
      "loss": 1.29,
      "theoretical_loss": 3.3937607633047624,
      "tokens_seen": 2530279424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023562688064192577,
      "loss": 1.2838,
      "theoretical_loss": 3.3937538274529366,
      "tokens_seen": 2530344960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023560682046138416,
      "loss": 1.2893,
      "theoretical_loss": 3.3937468918310447,
      "tokens_seen": 2530410496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023558676028084252,
      "loss": 1.2798,
      "theoretical_loss": 3.3937399564390733,
      "tokens_seen": 2530476032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002355667001003009,
      "loss": 1.2985,
      "theoretical_loss": 3.3937330212770083,
      "tokens_seen": 2530541568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023554663991975927,
      "loss": 1.316,
      "theoretical_loss": 3.393726086344836,
      "tokens_seen": 2530607104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023552657973921766,
      "loss": 1.3255,
      "theoretical_loss": 3.393719151642544,
      "tokens_seen": 2530672640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023550651955867602,
      "loss": 1.2738,
      "theoretical_loss": 3.3937122171701173,
      "tokens_seen": 2530738176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354864593781344,
      "loss": 1.3738,
      "theoretical_loss": 3.393705282927543,
      "tokens_seen": 2530803712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023546639919759277,
      "loss": 1.3518,
      "theoretical_loss": 3.3936983489148074,
      "tokens_seen": 2530869248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023544633901705113,
      "loss": 1.254,
      "theoretical_loss": 3.3936914151318973,
      "tokens_seen": 2530934784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023542627883650952,
      "loss": 1.4147,
      "theoretical_loss": 3.3936844815787985,
      "tokens_seen": 2531000320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354062186559679,
      "loss": 1.3344,
      "theoretical_loss": 3.3936775482554977,
      "tokens_seen": 2531065856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023538615847542628,
      "loss": 1.3206,
      "theoretical_loss": 3.3936706151619815,
      "tokens_seen": 2531131392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023536609829488464,
      "loss": 1.2576,
      "theoretical_loss": 3.3936636822982367,
      "tokens_seen": 2531196928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023534603811434305,
      "loss": 1.3308,
      "theoretical_loss": 3.3936567496642485,
      "tokens_seen": 2531262464
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023532597793380142,
      "loss": 1.2024,
      "theoretical_loss": 3.3936498172600045,
      "tokens_seen": 2531328000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002353059177532598,
      "loss": 1.3267,
      "theoretical_loss": 3.393642885085491,
      "tokens_seen": 2531393536
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 921348,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2266067266464233,
      "objective/train/theoretical_loss": 3.3936359531406937,
      "objective/train/tokens_used": 901918176,
      "theoretical_loss": 3.3936359531406937,
      "tokens_seen": 2531459072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023528585757271817,
      "loss": 1.2252,
      "theoretical_loss": 3.3936359531406937,
      "tokens_seen": 2531459072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023526579739217656,
      "loss": 1.2849,
      "theoretical_loss": 3.3936290214255997,
      "tokens_seen": 2531524608
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023524573721163492,
      "loss": 1.421,
      "theoretical_loss": 3.3936220899401954,
      "tokens_seen": 2531590144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023522567703109328,
      "loss": 1.2169,
      "theoretical_loss": 3.393615158684467,
      "tokens_seen": 2531655680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023520561685055167,
      "loss": 1.3707,
      "theoretical_loss": 3.393608227658401,
      "tokens_seen": 2531721216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023518555667001003,
      "loss": 1.4081,
      "theoretical_loss": 3.3936012968619838,
      "tokens_seen": 2531786752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023516549648946842,
      "loss": 1.32,
      "theoretical_loss": 3.393594366295202,
      "tokens_seen": 2531852288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023514543630892679,
      "loss": 1.3821,
      "theoretical_loss": 3.393587435958042,
      "tokens_seen": 2531917824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023512537612838517,
      "loss": 1.3061,
      "theoretical_loss": 3.39358050585049,
      "tokens_seen": 2531983360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023510531594784354,
      "loss": 1.3634,
      "theoretical_loss": 3.393573575972533,
      "tokens_seen": 2532048896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023508525576730193,
      "loss": 1.2712,
      "theoretical_loss": 3.393566646324157,
      "tokens_seen": 2532114432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002350651955867603,
      "loss": 1.3174,
      "theoretical_loss": 3.393559716905349,
      "tokens_seen": 2532179968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023504513540621865,
      "loss": 1.3451,
      "theoretical_loss": 3.3935527877160947,
      "tokens_seen": 2532245504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023502507522567704,
      "loss": 1.377,
      "theoretical_loss": 3.3935458587563807,
      "tokens_seen": 2532311040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002350050150451354,
      "loss": 1.3501,
      "theoretical_loss": 3.3935389300261942,
      "tokens_seen": 2532376576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002349849548645938,
      "loss": 1.247,
      "theoretical_loss": 3.3935320015255206,
      "tokens_seen": 2532442112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023496489468405215,
      "loss": 1.2696,
      "theoretical_loss": 3.393525073254347,
      "tokens_seen": 2532507648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023494483450351054,
      "loss": 1.2861,
      "theoretical_loss": 3.39351814521266,
      "tokens_seen": 2532573184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002349247743229689,
      "loss": 1.3149,
      "theoretical_loss": 3.3935112174004454,
      "tokens_seen": 2532638720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002349047141424273,
      "loss": 1.2406,
      "theoretical_loss": 3.39350428981769,
      "tokens_seen": 2532704256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023488465396188566,
      "loss": 1.3341,
      "theoretical_loss": 3.3934973624643807,
      "tokens_seen": 2532769792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023486459378134402,
      "loss": 1.306,
      "theoretical_loss": 3.393490435340503,
      "tokens_seen": 2532835328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002348445336008024,
      "loss": 1.239,
      "theoretical_loss": 3.3934835084460446,
      "tokens_seen": 2532900864
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023482447342026077,
      "loss": 1.444,
      "theoretical_loss": 3.393476581780991,
      "tokens_seen": 2532966400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023480441323971916,
      "loss": 1.3266,
      "theoretical_loss": 3.393469655345329,
      "tokens_seen": 2533031936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023478435305917752,
      "loss": 1.2563,
      "theoretical_loss": 3.3934627291390456,
      "tokens_seen": 2533097472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002347642928786359,
      "loss": 1.2682,
      "theoretical_loss": 3.393455803162126,
      "tokens_seen": 2533163008
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023474423269809427,
      "loss": 1.2032,
      "theoretical_loss": 3.3934488774145577,
      "tokens_seen": 2533228544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023472417251755266,
      "loss": 1.2747,
      "theoretical_loss": 3.393441951896327,
      "tokens_seen": 2533294080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023470411233701103,
      "loss": 1.3323,
      "theoretical_loss": 3.3934350266074196,
      "tokens_seen": 2533359616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023468405215646942,
      "loss": 1.2968,
      "theoretical_loss": 3.3934281015478227,
      "tokens_seen": 2533425152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023466399197592778,
      "loss": 1.305,
      "theoretical_loss": 3.3934211767175233,
      "tokens_seen": 2533490688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023464393179538614,
      "loss": 1.3011,
      "theoretical_loss": 3.3934142521165067,
      "tokens_seen": 2533556224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023462387161484453,
      "loss": 1.3367,
      "theoretical_loss": 3.3934073277447605,
      "tokens_seen": 2533621760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023460381143430292,
      "loss": 1.319,
      "theoretical_loss": 3.39340040360227,
      "tokens_seen": 2533687296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002345837512537613,
      "loss": 1.3371,
      "theoretical_loss": 3.393393479689023,
      "tokens_seen": 2533752832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023456369107321967,
      "loss": 1.3563,
      "theoretical_loss": 3.3933865560050043,
      "tokens_seen": 2533818368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023454363089267806,
      "loss": 1.3537,
      "theoretical_loss": 3.393379632550202,
      "tokens_seen": 2533883904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023452357071213642,
      "loss": 1.2425,
      "theoretical_loss": 3.393372709324602,
      "tokens_seen": 2533949440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002345035105315948,
      "loss": 1.2524,
      "theoretical_loss": 3.3933657863281903,
      "tokens_seen": 2534014976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023448345035105317,
      "loss": 1.3222,
      "theoretical_loss": 3.3933588635609544,
      "tokens_seen": 2534080512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023446339017051154,
      "loss": 1.3306,
      "theoretical_loss": 3.39335194102288,
      "tokens_seen": 2534146048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023444332998996992,
      "loss": 1.2474,
      "theoretical_loss": 3.3933450187139536,
      "tokens_seen": 2534211584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002344232698094283,
      "loss": 1.2812,
      "theoretical_loss": 3.393338096634162,
      "tokens_seen": 2534277120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023440320962888668,
      "loss": 1.2894,
      "theoretical_loss": 3.3933311747834916,
      "tokens_seen": 2534342656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023438314944834504,
      "loss": 1.3491,
      "theoretical_loss": 3.3933242531619285,
      "tokens_seen": 2534408192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023436308926780343,
      "loss": 1.3092,
      "theoretical_loss": 3.3933173317694596,
      "tokens_seen": 2534473728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002343430290872618,
      "loss": 1.2819,
      "theoretical_loss": 3.393310410606072,
      "tokens_seen": 2534539264
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023432296890672018,
      "loss": 1.2693,
      "theoretical_loss": 3.3933034896717507,
      "tokens_seen": 2534604800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023430290872617854,
      "loss": 1.3747,
      "theoretical_loss": 3.3932965689664836,
      "tokens_seen": 2534670336
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 922493,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4441323280334473,
      "objective/train/theoretical_loss": 3.3932896484902564,
      "objective/train/tokens_used": 905194976,
      "theoretical_loss": 3.3932896484902564,
      "tokens_seen": 2534735872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002342828485456369,
      "loss": 1.3382,
      "theoretical_loss": 3.3932896484902564,
      "tokens_seen": 2534735872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002342627883650953,
      "loss": 1.2646,
      "theoretical_loss": 3.393282728243056,
      "tokens_seen": 2534801408
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023424272818455366,
      "loss": 1.3002,
      "theoretical_loss": 3.3932758082248684,
      "tokens_seen": 2534866944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023422266800401204,
      "loss": 1.2672,
      "theoretical_loss": 3.393268888435681,
      "tokens_seen": 2534932480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002342026078234704,
      "loss": 1.2629,
      "theoretical_loss": 3.3932619688754793,
      "tokens_seen": 2534998016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341825476429288,
      "loss": 1.2561,
      "theoretical_loss": 3.3932550495442504,
      "tokens_seen": 2535063552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023416248746238716,
      "loss": 1.3855,
      "theoretical_loss": 3.3932481304419806,
      "tokens_seen": 2535129088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023414242728184555,
      "loss": 1.2565,
      "theoretical_loss": 3.393241211568656,
      "tokens_seen": 2535194624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341223671013039,
      "loss": 1.2771,
      "theoretical_loss": 3.393234292924264,
      "tokens_seen": 2535260160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341023069207623,
      "loss": 1.3158,
      "theoretical_loss": 3.3932273745087906,
      "tokens_seen": 2535325696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023408224674022066,
      "loss": 1.3222,
      "theoretical_loss": 3.3932204563222226,
      "tokens_seen": 2535391232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023406218655967902,
      "loss": 1.2322,
      "theoretical_loss": 3.393213538364546,
      "tokens_seen": 2535456768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002340421263791374,
      "loss": 1.3096,
      "theoretical_loss": 3.393206620635748,
      "tokens_seen": 2535522304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023402206619859578,
      "loss": 1.3569,
      "theoretical_loss": 3.393199703135814,
      "tokens_seen": 2535587840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023400200601805416,
      "loss": 1.2447,
      "theoretical_loss": 3.3931927858647315,
      "tokens_seen": 2535653376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023398194583751253,
      "loss": 1.1389,
      "theoretical_loss": 3.3931858688224867,
      "tokens_seen": 2535718912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023396188565697092,
      "loss": 1.2994,
      "theoretical_loss": 3.3931789520090665,
      "tokens_seen": 2535784448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023394182547642928,
      "loss": 1.268,
      "theoretical_loss": 3.393172035424457,
      "tokens_seen": 2535849984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023392176529588767,
      "loss": 1.2489,
      "theoretical_loss": 3.3931651190686445,
      "tokens_seen": 2535915520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023390170511534603,
      "loss": 1.2966,
      "theoretical_loss": 3.393158202941616,
      "tokens_seen": 2535981056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002338816449348044,
      "loss": 1.2845,
      "theoretical_loss": 3.3931512870433576,
      "tokens_seen": 2536046592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023386158475426278,
      "loss": 1.2369,
      "theoretical_loss": 3.393144371373856,
      "tokens_seen": 2536112128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023384152457372117,
      "loss": 1.3497,
      "theoretical_loss": 3.393137455933098,
      "tokens_seen": 2536177664
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023382146439317956,
      "loss": 1.3591,
      "theoretical_loss": 3.39313054072107,
      "tokens_seen": 2536243200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023380140421263792,
      "loss": 1.3728,
      "theoretical_loss": 3.3931236257377577,
      "tokens_seen": 2536308736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002337813440320963,
      "loss": 1.3139,
      "theoretical_loss": 3.393116710983149,
      "tokens_seen": 2536374272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023376128385155467,
      "loss": 1.245,
      "theoretical_loss": 3.3931097964572294,
      "tokens_seen": 2536439808
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023374122367101306,
      "loss": 1.3315,
      "theoretical_loss": 3.393102882159986,
      "tokens_seen": 2536505344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023372116349047143,
      "loss": 1.2661,
      "theoretical_loss": 3.393095968091405,
      "tokens_seen": 2536570880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023370110330992982,
      "loss": 1.1864,
      "theoretical_loss": 3.3930890542514733,
      "tokens_seen": 2536636416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023368104312938818,
      "loss": 1.2417,
      "theoretical_loss": 3.393082140640177,
      "tokens_seen": 2536701952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023366098294884654,
      "loss": 1.2916,
      "theoretical_loss": 3.393075227257503,
      "tokens_seen": 2536767488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023364092276830493,
      "loss": 1.308,
      "theoretical_loss": 3.3930683141034375,
      "tokens_seen": 2536833024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002336208625877633,
      "loss": 1.377,
      "theoretical_loss": 3.3930614011779667,
      "tokens_seen": 2536898560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023360080240722168,
      "loss": 1.2584,
      "theoretical_loss": 3.3930544884810785,
      "tokens_seen": 2536964096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023358074222668004,
      "loss": 1.1891,
      "theoretical_loss": 3.393047576012758,
      "tokens_seen": 2537029632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023356068204613843,
      "loss": 1.2358,
      "theoretical_loss": 3.393040663772992,
      "tokens_seen": 2537095168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002335406218655968,
      "loss": 1.2922,
      "theoretical_loss": 3.3930337517617684,
      "tokens_seen": 2537160704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023352056168505518,
      "loss": 1.2685,
      "theoretical_loss": 3.393026839979072,
      "tokens_seen": 2537226240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023350050150451355,
      "loss": 1.2198,
      "theoretical_loss": 3.39301992842489,
      "tokens_seen": 2537291776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002334804413239719,
      "loss": 1.2906,
      "theoretical_loss": 3.393013017099209,
      "tokens_seen": 2537357312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002334603811434303,
      "loss": 1.3042,
      "theoretical_loss": 3.3930061060020154,
      "tokens_seen": 2537422848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023344032096288866,
      "loss": 1.2958,
      "theoretical_loss": 3.392999195133296,
      "tokens_seen": 2537488384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023342026078234705,
      "loss": 1.2147,
      "theoretical_loss": 3.392992284493037,
      "tokens_seen": 2537553920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002334002006018054,
      "loss": 1.244,
      "theoretical_loss": 3.3929853740812255,
      "tokens_seen": 2537619456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002333801404212638,
      "loss": 1.2696,
      "theoretical_loss": 3.3929784638978475,
      "tokens_seen": 2537684992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023336008024072216,
      "loss": 1.3701,
      "theoretical_loss": 3.39297155394289,
      "tokens_seen": 2537750528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023334002006018055,
      "loss": 1.4499,
      "theoretical_loss": 3.3929646442163386,
      "tokens_seen": 2537816064
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023331995987963891,
      "loss": 1.2674,
      "theoretical_loss": 3.392957734718181,
      "tokens_seen": 2537881600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023329989969909728,
      "loss": 1.1724,
      "theoretical_loss": 3.3929508254484033,
      "tokens_seen": 2537947136
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 923645,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3264087438583374,
      "objective/train/theoretical_loss": 3.392943916406992,
      "objective/train/tokens_used": 908471776,
      "theoretical_loss": 3.392943916406992,
      "tokens_seen": 2538012672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023327983951855567,
      "loss": 1.3633,
      "theoretical_loss": 3.392943916406992,
      "tokens_seen": 2538012672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023325977933801403,
      "loss": 1.2765,
      "theoretical_loss": 3.392937007593934,
      "tokens_seen": 2538078208
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023323971915747242,
      "loss": 1.1971,
      "theoretical_loss": 3.392930099009215,
      "tokens_seen": 2538143744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023321965897693078,
      "loss": 1.2715,
      "theoretical_loss": 3.392923190652822,
      "tokens_seen": 2538209280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023319959879638917,
      "loss": 1.2952,
      "theoretical_loss": 3.3929162825247423,
      "tokens_seen": 2538274816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023317953861584753,
      "loss": 1.1082,
      "theoretical_loss": 3.3929093746249617,
      "tokens_seen": 2538340352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023315947843530592,
      "loss": 1.3841,
      "theoretical_loss": 3.392902466953467,
      "tokens_seen": 2538405888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023313941825476428,
      "loss": 1.1833,
      "theoretical_loss": 3.3928955595102446,
      "tokens_seen": 2538471424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023311935807422267,
      "loss": 1.2684,
      "theoretical_loss": 3.392888652295281,
      "tokens_seen": 2538536960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023309929789368106,
      "loss": 1.1557,
      "theoretical_loss": 3.3928817453085625,
      "tokens_seen": 2538602496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023307923771313942,
      "loss": 1.2416,
      "theoretical_loss": 3.3928748385500764,
      "tokens_seen": 2538668032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023305917753259781,
      "loss": 1.3206,
      "theoretical_loss": 3.392867932019809,
      "tokens_seen": 2538733568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023303911735205618,
      "loss": 1.2026,
      "theoretical_loss": 3.392861025717747,
      "tokens_seen": 2538799104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023301905717151457,
      "loss": 1.2755,
      "theoretical_loss": 3.3928541196438764,
      "tokens_seen": 2538864640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023299899699097293,
      "loss": 1.306,
      "theoretical_loss": 3.392847213798184,
      "tokens_seen": 2538930176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023297893681043132,
      "loss": 1.3124,
      "theoretical_loss": 3.392840308180657,
      "tokens_seen": 2538995712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023295887662988968,
      "loss": 1.2694,
      "theoretical_loss": 3.3928334027912816,
      "tokens_seen": 2539061248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023293881644934807,
      "loss": 1.3038,
      "theoretical_loss": 3.3928264976300437,
      "tokens_seen": 2539126784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023291875626880643,
      "loss": 1.3387,
      "theoretical_loss": 3.3928195926969305,
      "tokens_seen": 2539192320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002328986960882648,
      "loss": 1.2893,
      "theoretical_loss": 3.3928126879919285,
      "tokens_seen": 2539257856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023287863590772318,
      "loss": 1.4551,
      "theoretical_loss": 3.3928057835150245,
      "tokens_seen": 2539323392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023285857572718154,
      "loss": 1.3237,
      "theoretical_loss": 3.3927988792662047,
      "tokens_seen": 2539388928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023283851554663993,
      "loss": 1.2466,
      "theoretical_loss": 3.392791975245456,
      "tokens_seen": 2539454464
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002328184553660983,
      "loss": 1.2544,
      "theoretical_loss": 3.3927850714527645,
      "tokens_seen": 2539520000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023279839518555669,
      "loss": 1.3134,
      "theoretical_loss": 3.3927781678881175,
      "tokens_seen": 2539585536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023277833500501505,
      "loss": 1.1559,
      "theoretical_loss": 3.3927712645515005,
      "tokens_seen": 2539651072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023275827482447344,
      "loss": 1.2282,
      "theoretical_loss": 3.3927643614429015,
      "tokens_seen": 2539716608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002327382146439318,
      "loss": 1.3168,
      "theoretical_loss": 3.392757458562306,
      "tokens_seen": 2539782144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023271815446339016,
      "loss": 1.243,
      "theoretical_loss": 3.392750555909701,
      "tokens_seen": 2539847680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023269809428284855,
      "loss": 1.3382,
      "theoretical_loss": 3.392743653485073,
      "tokens_seen": 2539913216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002326780341023069,
      "loss": 1.3677,
      "theoretical_loss": 3.3927367512884086,
      "tokens_seen": 2539978752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002326579739217653,
      "loss": 1.2787,
      "theoretical_loss": 3.3927298493196947,
      "tokens_seen": 2540044288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023263791374122366,
      "loss": 1.1864,
      "theoretical_loss": 3.392722947578917,
      "tokens_seen": 2540109824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023261785356068205,
      "loss": 1.363,
      "theoretical_loss": 3.392716046066063,
      "tokens_seen": 2540175360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023259779338014042,
      "loss": 1.4196,
      "theoretical_loss": 3.392709144781119,
      "tokens_seen": 2540240896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002325777331995988,
      "loss": 1.2948,
      "theoretical_loss": 3.3927022437240715,
      "tokens_seen": 2540306432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023255767301905717,
      "loss": 1.2873,
      "theoretical_loss": 3.3926953428949074,
      "tokens_seen": 2540371968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023253761283851556,
      "loss": 1.337,
      "theoretical_loss": 3.3926884422936125,
      "tokens_seen": 2540437504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023251755265797392,
      "loss": 1.3602,
      "theoretical_loss": 3.3926815419201746,
      "tokens_seen": 2540503040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023249749247743228,
      "loss": 1.1958,
      "theoretical_loss": 3.3926746417745792,
      "tokens_seen": 2540568576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023247743229689067,
      "loss": 1.3665,
      "theoretical_loss": 3.3926677418568136,
      "tokens_seen": 2540634112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023245737211634903,
      "loss": 1.237,
      "theoretical_loss": 3.392660842166864,
      "tokens_seen": 2540699648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023243731193580742,
      "loss": 1.2613,
      "theoretical_loss": 3.3926539427047175,
      "tokens_seen": 2540765184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023241725175526578,
      "loss": 1.2628,
      "theoretical_loss": 3.39264704347036,
      "tokens_seen": 2540830720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023239719157472417,
      "loss": 1.2362,
      "theoretical_loss": 3.3926401444637784,
      "tokens_seen": 2540896256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023237713139418254,
      "loss": 1.2816,
      "theoretical_loss": 3.3926332456849595,
      "tokens_seen": 2540961792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023235707121364095,
      "loss": 1.2716,
      "theoretical_loss": 3.39262634713389,
      "tokens_seen": 2541027328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023233701103309932,
      "loss": 1.2402,
      "theoretical_loss": 3.392619448810556,
      "tokens_seen": 2541092864
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023231695085255768,
      "loss": 1.4142,
      "theoretical_loss": 3.3926125507149445,
      "tokens_seen": 2541158400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023229689067201607,
      "loss": 1.388,
      "theoretical_loss": 3.3926056528470423,
      "tokens_seen": 2541223936
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 924855,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3667056560516357,
      "objective/train/theoretical_loss": 3.3925987552068353,
      "objective/train/tokens_used": 911748576,
      "theoretical_loss": 3.3925987552068353,
      "tokens_seen": 2541289472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023227683049147443,
      "loss": 1.2384,
      "theoretical_loss": 3.3925987552068353,
      "tokens_seen": 2541289472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023225677031093282,
      "loss": 1.2457,
      "theoretical_loss": 3.3925918577943106,
      "tokens_seen": 2541355008
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023223671013039118,
      "loss": 1.3912,
      "theoretical_loss": 3.392584960609455,
      "tokens_seen": 2541420544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023221664994984957,
      "loss": 1.2465,
      "theoretical_loss": 3.3925780636522544,
      "tokens_seen": 2541486080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023219658976930793,
      "loss": 1.1891,
      "theoretical_loss": 3.3925711669226963,
      "tokens_seen": 2541551616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023217652958876632,
      "loss": 1.4094,
      "theoretical_loss": 3.3925642704207664,
      "tokens_seen": 2541617152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023215646940822468,
      "loss": 1.3169,
      "theoretical_loss": 3.392557374146452,
      "tokens_seen": 2541682688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023213640922768307,
      "loss": 1.2887,
      "theoretical_loss": 3.3925504780997398,
      "tokens_seen": 2541748224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023211634904714144,
      "loss": 1.3069,
      "theoretical_loss": 3.392543582280616,
      "tokens_seen": 2541813760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002320962888665998,
      "loss": 1.2283,
      "theoretical_loss": 3.392536686689067,
      "tokens_seen": 2541879296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002320762286860582,
      "loss": 1.41,
      "theoretical_loss": 3.3925297913250807,
      "tokens_seen": 2541944832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023205616850551655,
      "loss": 1.2219,
      "theoretical_loss": 3.392522896188642,
      "tokens_seen": 2542010368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023203610832497494,
      "loss": 1.348,
      "theoretical_loss": 3.3925160012797386,
      "tokens_seen": 2542075904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002320160481444333,
      "loss": 1.1912,
      "theoretical_loss": 3.392509106598357,
      "tokens_seen": 2542141440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002319959879638917,
      "loss": 1.2898,
      "theoretical_loss": 3.392502212144483,
      "tokens_seen": 2542206976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023197592778335005,
      "loss": 1.2949,
      "theoretical_loss": 3.392495317918105,
      "tokens_seen": 2542272512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023195586760280844,
      "loss": 1.376,
      "theoretical_loss": 3.392488423919208,
      "tokens_seen": 2542338048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002319358074222668,
      "loss": 1.3114,
      "theoretical_loss": 3.392481530147779,
      "tokens_seen": 2542403584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023191574724172517,
      "loss": 1.3126,
      "theoretical_loss": 3.392474636603805,
      "tokens_seen": 2542469120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023189568706118356,
      "loss": 1.3893,
      "theoretical_loss": 3.3924677432872725,
      "tokens_seen": 2542534656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023187562688064192,
      "loss": 1.2242,
      "theoretical_loss": 3.392460850198168,
      "tokens_seen": 2542600192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002318555667001003,
      "loss": 1.3638,
      "theoretical_loss": 3.3924539573364783,
      "tokens_seen": 2542665728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023183550651955867,
      "loss": 1.2766,
      "theoretical_loss": 3.3924470647021896,
      "tokens_seen": 2542731264
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023181544633901706,
      "loss": 1.3009,
      "theoretical_loss": 3.3924401722952893,
      "tokens_seen": 2542796800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023179538615847542,
      "loss": 1.366,
      "theoretical_loss": 3.3924332801157635,
      "tokens_seen": 2542862336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002317753259779338,
      "loss": 1.3684,
      "theoretical_loss": 3.392426388163599,
      "tokens_seen": 2542927872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023175526579739217,
      "loss": 1.3311,
      "theoretical_loss": 3.392419496438782,
      "tokens_seen": 2542993408
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023173520561685053,
      "loss": 1.3037,
      "theoretical_loss": 3.3924126049413,
      "tokens_seen": 2543058944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023171514543630892,
      "loss": 1.2468,
      "theoretical_loss": 3.392405713671139,
      "tokens_seen": 2543124480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023169508525576729,
      "loss": 1.2259,
      "theoretical_loss": 3.392398822628286,
      "tokens_seen": 2543190016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023167502507522568,
      "loss": 1.2246,
      "theoretical_loss": 3.3923919318127274,
      "tokens_seen": 2543255552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023165496489468404,
      "loss": 1.2855,
      "theoretical_loss": 3.3923850412244496,
      "tokens_seen": 2543321088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023163490471414243,
      "loss": 1.2819,
      "theoretical_loss": 3.3923781508634403,
      "tokens_seen": 2543386624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002316148445336008,
      "loss": 1.3187,
      "theoretical_loss": 3.3923712607296848,
      "tokens_seen": 2543452160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002315947843530592,
      "loss": 1.3124,
      "theoretical_loss": 3.3923643708231705,
      "tokens_seen": 2543517696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023157472417251757,
      "loss": 1.253,
      "theoretical_loss": 3.392357481143884,
      "tokens_seen": 2543583232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023155466399197596,
      "loss": 1.4152,
      "theoretical_loss": 3.3923505916918115,
      "tokens_seen": 2543648768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023153460381143432,
      "loss": 1.404,
      "theoretical_loss": 3.3923437024669405,
      "tokens_seen": 2543714304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023151454363089268,
      "loss": 1.2665,
      "theoretical_loss": 3.392336813469257,
      "tokens_seen": 2543779840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023149448345035107,
      "loss": 1.2368,
      "theoretical_loss": 3.392329924698748,
      "tokens_seen": 2543845376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023147442326980943,
      "loss": 1.3588,
      "theoretical_loss": 3.3923230361553998,
      "tokens_seen": 2543910912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023145436308926782,
      "loss": 1.2957,
      "theoretical_loss": 3.392316147839199,
      "tokens_seen": 2543976448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023143430290872619,
      "loss": 1.3076,
      "theoretical_loss": 3.392309259750133,
      "tokens_seen": 2544041984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023141424272818457,
      "loss": 1.2901,
      "theoretical_loss": 3.3923023718881877,
      "tokens_seen": 2544107520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023139418254764294,
      "loss": 1.2834,
      "theoretical_loss": 3.39229548425335,
      "tokens_seen": 2544173056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023137412236710133,
      "loss": 1.2669,
      "theoretical_loss": 3.392288596845607,
      "tokens_seen": 2544238592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002313540621865597,
      "loss": 1.2922,
      "theoretical_loss": 3.3922817096649442,
      "tokens_seen": 2544304128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023133400200601805,
      "loss": 1.3144,
      "theoretical_loss": 3.3922748227113497,
      "tokens_seen": 2544369664
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023131394182547644,
      "loss": 1.2929,
      "theoretical_loss": 3.392267935984809,
      "tokens_seen": 2544435200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312938816449348,
      "loss": 1.3303,
      "theoretical_loss": 3.3922610494853096,
      "tokens_seen": 2544500736
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 925876,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3360546827316284,
      "objective/train/theoretical_loss": 3.3922541632128373,
      "objective/train/tokens_used": 915025376,
      "theoretical_loss": 3.3922541632128373,
      "tokens_seen": 2544566272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312738214643932,
      "loss": 1.3073,
      "theoretical_loss": 3.3922541632128373,
      "tokens_seen": 2544566272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023125376128385155,
      "loss": 1.2513,
      "theoretical_loss": 3.39224727716738,
      "tokens_seen": 2544631808
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023123370110330994,
      "loss": 1.2782,
      "theoretical_loss": 3.392240391348923,
      "tokens_seen": 2544697344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312136409227683,
      "loss": 1.2373,
      "theoretical_loss": 3.392233505757454,
      "tokens_seen": 2544762880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002311935807422267,
      "loss": 1.2918,
      "theoretical_loss": 3.392226620392959,
      "tokens_seen": 2544828416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023117352056168506,
      "loss": 1.3595,
      "theoretical_loss": 3.3922197352554253,
      "tokens_seen": 2544893952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023115346038114342,
      "loss": 1.3242,
      "theoretical_loss": 3.392212850344839,
      "tokens_seen": 2544959488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002311334002006018,
      "loss": 1.2397,
      "theoretical_loss": 3.392205965661187,
      "tokens_seen": 2545025024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023111334002006017,
      "loss": 1.2491,
      "theoretical_loss": 3.3921990812044562,
      "tokens_seen": 2545090560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023109327983951856,
      "loss": 1.1877,
      "theoretical_loss": 3.392192196974633,
      "tokens_seen": 2545156096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023107321965897692,
      "loss": 1.2053,
      "theoretical_loss": 3.392185312971704,
      "tokens_seen": 2545221632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002310531594784353,
      "loss": 1.187,
      "theoretical_loss": 3.3921784291956563,
      "tokens_seen": 2545287168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023103309929789367,
      "loss": 1.1851,
      "theoretical_loss": 3.3921715456464763,
      "tokens_seen": 2545352704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023101303911735206,
      "loss": 1.3078,
      "theoretical_loss": 3.392164662324151,
      "tokens_seen": 2545418240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023099297893681043,
      "loss": 1.2729,
      "theoretical_loss": 3.392157779228666,
      "tokens_seen": 2545483776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023097291875626881,
      "loss": 1.3199,
      "theoretical_loss": 3.3921508963600093,
      "tokens_seen": 2545549312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023095285857572718,
      "loss": 1.3278,
      "theoretical_loss": 3.3921440137181667,
      "tokens_seen": 2545614848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023093279839518554,
      "loss": 1.1528,
      "theoretical_loss": 3.392137131303125,
      "tokens_seen": 2545680384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023091273821464393,
      "loss": 1.323,
      "theoretical_loss": 3.3921302491148717,
      "tokens_seen": 2545745920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002308926780341023,
      "loss": 1.2429,
      "theoretical_loss": 3.392123367153393,
      "tokens_seen": 2545811456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023087261785356068,
      "loss": 1.2823,
      "theoretical_loss": 3.392116485418675,
      "tokens_seen": 2545876992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023085255767301907,
      "loss": 1.1817,
      "theoretical_loss": 3.392109603910705,
      "tokens_seen": 2545942528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023083249749247746,
      "loss": 1.3439,
      "theoretical_loss": 3.3921027226294695,
      "tokens_seen": 2546008064
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023081243731193582,
      "loss": 1.3047,
      "theoretical_loss": 3.392095841574956,
      "tokens_seen": 2546073600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307923771313942,
      "loss": 1.2534,
      "theoretical_loss": 3.3920889607471496,
      "tokens_seen": 2546139136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023077231695085257,
      "loss": 1.3664,
      "theoretical_loss": 3.3920820801460385,
      "tokens_seen": 2546204672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023075225677031094,
      "loss": 1.3099,
      "theoretical_loss": 3.3920751997716083,
      "tokens_seen": 2546270208
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023073219658976932,
      "loss": 1.1967,
      "theoretical_loss": 3.392068319623846,
      "tokens_seen": 2546335744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307121364092277,
      "loss": 1.271,
      "theoretical_loss": 3.392061439702739,
      "tokens_seen": 2546401280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023069207622868608,
      "loss": 1.3284,
      "theoretical_loss": 3.3920545600082734,
      "tokens_seen": 2546466816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023067201604814444,
      "loss": 1.2488,
      "theoretical_loss": 3.392047680540436,
      "tokens_seen": 2546532352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023065195586760283,
      "loss": 1.3052,
      "theoretical_loss": 3.392040801299213,
      "tokens_seen": 2546597888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002306318956870612,
      "loss": 1.4035,
      "theoretical_loss": 3.392033922284592,
      "tokens_seen": 2546663424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023061183550651958,
      "loss": 1.3251,
      "theoretical_loss": 3.392027043496559,
      "tokens_seen": 2546728960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023059177532597794,
      "loss": 1.2543,
      "theoretical_loss": 3.392020164935101,
      "tokens_seen": 2546794496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002305717151454363,
      "loss": 1.2776,
      "theoretical_loss": 3.392013286600205,
      "tokens_seen": 2546860032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002305516549648947,
      "loss": 1.1164,
      "theoretical_loss": 3.392006408491857,
      "tokens_seen": 2546925568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023053159478435306,
      "loss": 1.3665,
      "theoretical_loss": 3.391999530610044,
      "tokens_seen": 2546991104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023051153460381144,
      "loss": 1.276,
      "theoretical_loss": 3.3919926529547535,
      "tokens_seen": 2547056640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304914744232698,
      "loss": 1.3972,
      "theoretical_loss": 3.391985775525971,
      "tokens_seen": 2547122176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304714142427282,
      "loss": 1.2194,
      "theoretical_loss": 3.391978898323684,
      "tokens_seen": 2547187712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023045135406218656,
      "loss": 1.321,
      "theoretical_loss": 3.3919720213478786,
      "tokens_seen": 2547253248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023043129388164495,
      "loss": 1.2579,
      "theoretical_loss": 3.391965144598542,
      "tokens_seen": 2547318784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304112337011033,
      "loss": 1.3014,
      "theoretical_loss": 3.391958268075661,
      "tokens_seen": 2547384320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002303911735205617,
      "loss": 1.2547,
      "theoretical_loss": 3.391951391779222,
      "tokens_seen": 2547449856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023037111334002006,
      "loss": 1.2699,
      "theoretical_loss": 3.391944515709212,
      "tokens_seen": 2547515392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023035105315947842,
      "loss": 1.3736,
      "theoretical_loss": 3.391937639865617,
      "tokens_seen": 2547580928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002303309929789368,
      "loss": 1.2853,
      "theoretical_loss": 3.3919307642484244,
      "tokens_seen": 2547646464
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023031093279839518,
      "loss": 1.1486,
      "theoretical_loss": 3.391923888857621,
      "tokens_seen": 2547712000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023029087261785356,
      "loss": 1.2565,
      "theoretical_loss": 3.3919170136931935,
      "tokens_seen": 2547777536
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 927107,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2384496927261353,
      "objective/train/theoretical_loss": 3.391910138755128,
      "objective/train/tokens_used": 918302176,
      "theoretical_loss": 3.391910138755128,
      "tokens_seen": 2547843072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023027081243731193,
      "loss": 1.2026,
      "theoretical_loss": 3.391910138755128,
      "tokens_seen": 2547843072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023025075225677032,
      "loss": 1.2926,
      "theoretical_loss": 3.3919032640434117,
      "tokens_seen": 2547908608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023023069207622868,
      "loss": 1.2563,
      "theoretical_loss": 3.391896389558031,
      "tokens_seen": 2547974144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023021063189568707,
      "loss": 1.1939,
      "theoretical_loss": 3.3918895152989736,
      "tokens_seen": 2548039680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023019057171514543,
      "loss": 1.201,
      "theoretical_loss": 3.3918826412662253,
      "tokens_seen": 2548105216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002301705115346038,
      "loss": 1.2752,
      "theoretical_loss": 3.391875767459773,
      "tokens_seen": 2548170752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023015045135406218,
      "loss": 1.2772,
      "theoretical_loss": 3.3918688938796033,
      "tokens_seen": 2548236288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023013039117352054,
      "loss": 1.3354,
      "theoretical_loss": 3.3918620205257035,
      "tokens_seen": 2548301824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023011033099297893,
      "loss": 1.2105,
      "theoretical_loss": 3.3918551473980596,
      "tokens_seen": 2548367360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023009027081243732,
      "loss": 1.2032,
      "theoretical_loss": 3.3918482744966587,
      "tokens_seen": 2548432896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002300702106318957,
      "loss": 1.3504,
      "theoretical_loss": 3.391841401821488,
      "tokens_seen": 2548498432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023005015045135407,
      "loss": 1.2992,
      "theoretical_loss": 3.391834529372533,
      "tokens_seen": 2548563968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023003009027081246,
      "loss": 1.3862,
      "theoretical_loss": 3.391827657149782,
      "tokens_seen": 2548629504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023001003009027083,
      "loss": 1.2772,
      "theoretical_loss": 3.3918207851532203,
      "tokens_seen": 2548695040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022998996990972922,
      "loss": 1.3028,
      "theoretical_loss": 3.3918139133828356,
      "tokens_seen": 2548760576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022996990972918758,
      "loss": 1.4012,
      "theoretical_loss": 3.3918070418386144,
      "tokens_seen": 2548826112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022994984954864594,
      "loss": 1.275,
      "theoretical_loss": 3.391800170520543,
      "tokens_seen": 2548891648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022992978936810433,
      "loss": 1.3267,
      "theoretical_loss": 3.391793299428609,
      "tokens_seen": 2548957184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002299097291875627,
      "loss": 1.2164,
      "theoretical_loss": 3.391786428562798,
      "tokens_seen": 2549022720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022988966900702108,
      "loss": 1.2215,
      "theoretical_loss": 3.3917795579230976,
      "tokens_seen": 2549088256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022986960882647944,
      "loss": 1.2221,
      "theoretical_loss": 3.3917726875094947,
      "tokens_seen": 2549153792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022984954864593783,
      "loss": 1.342,
      "theoretical_loss": 3.3917658173219754,
      "tokens_seen": 2549219328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002298294884653962,
      "loss": 1.1806,
      "theoretical_loss": 3.3917589473605267,
      "tokens_seen": 2549284864
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022980942828485458,
      "loss": 1.3192,
      "theoretical_loss": 3.3917520776251355,
      "tokens_seen": 2549350400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022978936810431295,
      "loss": 1.3126,
      "theoretical_loss": 3.3917452081157884,
      "tokens_seen": 2549415936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002297693079237713,
      "loss": 1.2251,
      "theoretical_loss": 3.391738338832472,
      "tokens_seen": 2549481472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002297492477432297,
      "loss": 1.3276,
      "theoretical_loss": 3.3917314697751735,
      "tokens_seen": 2549547008
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022972918756268806,
      "loss": 1.2749,
      "theoretical_loss": 3.391724600943879,
      "tokens_seen": 2549612544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022970912738214645,
      "loss": 1.2326,
      "theoretical_loss": 3.391717732338576,
      "tokens_seen": 2549678080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002296890672016048,
      "loss": 1.2217,
      "theoretical_loss": 3.3917108639592506,
      "tokens_seen": 2549743616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002296690070210632,
      "loss": 1.3089,
      "theoretical_loss": 3.3917039958058903,
      "tokens_seen": 2549809152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022964894684052156,
      "loss": 1.2547,
      "theoretical_loss": 3.3916971278784813,
      "tokens_seen": 2549874688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022962888665997995,
      "loss": 1.3146,
      "theoretical_loss": 3.3916902601770103,
      "tokens_seen": 2549940224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022960882647943831,
      "loss": 1.2868,
      "theoretical_loss": 3.3916833927014642,
      "tokens_seen": 2550005760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022958876629889668,
      "loss": 1.2445,
      "theoretical_loss": 3.39167652545183,
      "tokens_seen": 2550071296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022956870611835507,
      "loss": 1.2995,
      "theoretical_loss": 3.391669658428094,
      "tokens_seen": 2550136832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022954864593781343,
      "loss": 1.331,
      "theoretical_loss": 3.3916627916302433,
      "tokens_seen": 2550202368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022952858575727182,
      "loss": 1.2807,
      "theoretical_loss": 3.3916559250582647,
      "tokens_seen": 2550267904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022950852557673018,
      "loss": 1.3521,
      "theoretical_loss": 3.391649058712145,
      "tokens_seen": 2550333440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022948846539618857,
      "loss": 1.2518,
      "theoretical_loss": 3.3916421925918705,
      "tokens_seen": 2550398976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022946840521564693,
      "loss": 1.2686,
      "theoretical_loss": 3.3916353266974286,
      "tokens_seen": 2550464512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022944834503510532,
      "loss": 1.2378,
      "theoretical_loss": 3.391628461028805,
      "tokens_seen": 2550530048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022942828485456368,
      "loss": 1.2267,
      "theoretical_loss": 3.3916215955859883,
      "tokens_seen": 2550595584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022940822467402207,
      "loss": 1.2759,
      "theoretical_loss": 3.3916147303689637,
      "tokens_seen": 2550661120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022938816449348043,
      "loss": 1.3356,
      "theoretical_loss": 3.3916078653777184,
      "tokens_seen": 2550726656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002293681043129388,
      "loss": 1.2648,
      "theoretical_loss": 3.3916010006122397,
      "tokens_seen": 2550792192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002293480441323972,
      "loss": 1.1256,
      "theoretical_loss": 3.3915941360725133,
      "tokens_seen": 2550857728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022932798395185558,
      "loss": 1.2613,
      "theoretical_loss": 3.3915872717585267,
      "tokens_seen": 2550923264
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022930792377131397,
      "loss": 1.3427,
      "theoretical_loss": 3.3915804076702667,
      "tokens_seen": 2550988800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022928786359077233,
      "loss": 1.2914,
      "theoretical_loss": 3.39157354380772,
      "tokens_seen": 2551054336
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 928186,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2692723274230957,
      "objective/train/theoretical_loss": 3.3915666801708735,
      "objective/train/tokens_used": 921578976,
      "theoretical_loss": 3.3915666801708735,
      "tokens_seen": 2551119872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022926780341023072,
      "loss": 1.3094,
      "theoretical_loss": 3.3915666801708735,
      "tokens_seen": 2551119872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022924774322968908,
      "loss": 1.2983,
      "theoretical_loss": 3.3915598167597136,
      "tokens_seen": 2551185408
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022922768304914747,
      "loss": 1.3001,
      "theoretical_loss": 3.3915529535742275,
      "tokens_seen": 2551250944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022920762286860583,
      "loss": 1.2585,
      "theoretical_loss": 3.3915460906144017,
      "tokens_seen": 2551316480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002291875626880642,
      "loss": 1.2536,
      "theoretical_loss": 3.391539227880223,
      "tokens_seen": 2551382016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022916750250752258,
      "loss": 1.353,
      "theoretical_loss": 3.391532365371678,
      "tokens_seen": 2551447552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022914744232698094,
      "loss": 1.1998,
      "theoretical_loss": 3.3915255030887543,
      "tokens_seen": 2551513088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022912738214643933,
      "loss": 1.2663,
      "theoretical_loss": 3.3915186410314377,
      "tokens_seen": 2551578624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002291073219658977,
      "loss": 1.3219,
      "theoretical_loss": 3.3915117791997154,
      "tokens_seen": 2551644160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022908726178535609,
      "loss": 1.2352,
      "theoretical_loss": 3.3915049175935743,
      "tokens_seen": 2551709696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022906720160481445,
      "loss": 1.337,
      "theoretical_loss": 3.3914980562130013,
      "tokens_seen": 2551775232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022904714142427284,
      "loss": 1.2546,
      "theoretical_loss": 3.3914911950579825,
      "tokens_seen": 2551840768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002290270812437312,
      "loss": 1.2496,
      "theoretical_loss": 3.3914843341285055,
      "tokens_seen": 2551906304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022900702106318956,
      "loss": 1.1115,
      "theoretical_loss": 3.391477473424557,
      "tokens_seen": 2551971840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022898696088264795,
      "loss": 1.2667,
      "theoretical_loss": 3.391470612946123,
      "tokens_seen": 2552037376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002289669007021063,
      "loss": 1.122,
      "theoretical_loss": 3.3914637526931917,
      "tokens_seen": 2552102912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002289468405215647,
      "loss": 1.2336,
      "theoretical_loss": 3.3914568926657482,
      "tokens_seen": 2552168448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022892678034102306,
      "loss": 1.2311,
      "theoretical_loss": 3.391450032863781,
      "tokens_seen": 2552233984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022890672016048145,
      "loss": 1.3124,
      "theoretical_loss": 3.3914431732872754,
      "tokens_seen": 2552299520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022888665997993982,
      "loss": 1.3253,
      "theoretical_loss": 3.391436313936219,
      "tokens_seen": 2552365056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002288665997993982,
      "loss": 1.2424,
      "theoretical_loss": 3.3914294548105985,
      "tokens_seen": 2552430592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022884653961885657,
      "loss": 1.2629,
      "theoretical_loss": 3.3914225959104005,
      "tokens_seen": 2552496128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022882647943831496,
      "loss": 1.2983,
      "theoretical_loss": 3.3914157372356124,
      "tokens_seen": 2552561664
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022880641925777332,
      "loss": 1.3037,
      "theoretical_loss": 3.39140887878622,
      "tokens_seen": 2552627200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022878635907723168,
      "loss": 1.2505,
      "theoretical_loss": 3.391402020562211,
      "tokens_seen": 2552692736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022876629889669007,
      "loss": 1.308,
      "theoretical_loss": 3.3913951625635717,
      "tokens_seen": 2552758272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022874623871614843,
      "loss": 1.2239,
      "theoretical_loss": 3.3913883047902895,
      "tokens_seen": 2552823808
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022872617853560682,
      "loss": 1.2598,
      "theoretical_loss": 3.3913814472423507,
      "tokens_seen": 2552889344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022870611835506518,
      "loss": 1.2808,
      "theoretical_loss": 3.3913745899197423,
      "tokens_seen": 2552954880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022868605817452357,
      "loss": 1.2019,
      "theoretical_loss": 3.3913677328224505,
      "tokens_seen": 2553020416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022866599799398194,
      "loss": 1.2037,
      "theoretical_loss": 3.391360875950463,
      "tokens_seen": 2553085952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022864593781344033,
      "loss": 1.1416,
      "theoretical_loss": 3.3913540193037663,
      "tokens_seen": 2553151488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002286258776328987,
      "loss": 1.2984,
      "theoretical_loss": 3.391347162882347,
      "tokens_seen": 2553217024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022860581745235705,
      "loss": 1.233,
      "theoretical_loss": 3.391340306686192,
      "tokens_seen": 2553282560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022858575727181547,
      "loss": 1.194,
      "theoretical_loss": 3.3913334507152886,
      "tokens_seen": 2553348096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022856569709127383,
      "loss": 1.2404,
      "theoretical_loss": 3.3913265949696227,
      "tokens_seen": 2553413632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022854563691073222,
      "loss": 1.2232,
      "theoretical_loss": 3.391319739449182,
      "tokens_seen": 2553479168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022852557673019058,
      "loss": 1.2146,
      "theoretical_loss": 3.391312884153953,
      "tokens_seen": 2553544704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022850551654964897,
      "loss": 1.1606,
      "theoretical_loss": 3.3913060290839225,
      "tokens_seen": 2553610240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022848545636910733,
      "loss": 1.3204,
      "theoretical_loss": 3.391299174239077,
      "tokens_seen": 2553675776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022846539618856572,
      "loss": 1.2898,
      "theoretical_loss": 3.391292319619404,
      "tokens_seen": 2553741312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022844533600802408,
      "loss": 1.3152,
      "theoretical_loss": 3.3912854652248896,
      "tokens_seen": 2553806848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022842527582748247,
      "loss": 1.2375,
      "theoretical_loss": 3.391278611055521,
      "tokens_seen": 2553872384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022840521564694084,
      "loss": 1.2807,
      "theoretical_loss": 3.3912717571112854,
      "tokens_seen": 2553937920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283851554663992,
      "loss": 1.2345,
      "theoretical_loss": 3.391264903392169,
      "tokens_seen": 2554003456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283650952858576,
      "loss": 1.2174,
      "theoretical_loss": 3.391258049898159,
      "tokens_seen": 2554068992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022834503510531595,
      "loss": 1.4115,
      "theoretical_loss": 3.3912511966292422,
      "tokens_seen": 2554134528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022832497492477434,
      "loss": 1.2741,
      "theoretical_loss": 3.391244343585405,
      "tokens_seen": 2554200064
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283049147442327,
      "loss": 1.2154,
      "theoretical_loss": 3.3912374907666347,
      "tokens_seen": 2554265600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002282848545636911,
      "loss": 1.1808,
      "theoretical_loss": 3.391230638172918,
      "tokens_seen": 2554331136
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 929369,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2488348484039307,
      "objective/train/theoretical_loss": 3.3912237858042413,
      "objective/train/tokens_used": 924855776,
      "theoretical_loss": 3.3912237858042413,
      "tokens_seen": 2554396672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022826479438314945,
      "loss": 1.2962,
      "theoretical_loss": 3.3912237858042413,
      "tokens_seen": 2554396672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022824473420260784,
      "loss": 1.2765,
      "theoretical_loss": 3.3912169336605924,
      "tokens_seen": 2554462208
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002282246740220662,
      "loss": 1.2384,
      "theoretical_loss": 3.3912100817419577,
      "tokens_seen": 2554527744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022820461384152457,
      "loss": 1.1954,
      "theoretical_loss": 3.3912032300483235,
      "tokens_seen": 2554593280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022818455366098296,
      "loss": 1.3625,
      "theoretical_loss": 3.3911963785796777,
      "tokens_seen": 2554658816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022816449348044132,
      "loss": 1.2302,
      "theoretical_loss": 3.3911895273360058,
      "tokens_seen": 2554724352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002281444332998997,
      "loss": 1.294,
      "theoretical_loss": 3.391182676317296,
      "tokens_seen": 2554789888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022812437311935807,
      "loss": 1.1749,
      "theoretical_loss": 3.391175825523534,
      "tokens_seen": 2554855424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022810431293881646,
      "loss": 1.2554,
      "theoretical_loss": 3.3911689749547076,
      "tokens_seen": 2554920960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022808425275827482,
      "loss": 1.281,
      "theoretical_loss": 3.391162124610803,
      "tokens_seen": 2554986496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002280641925777332,
      "loss": 1.3678,
      "theoretical_loss": 3.391155274491807,
      "tokens_seen": 2555052032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022804413239719157,
      "loss": 1.2437,
      "theoretical_loss": 3.391148424597707,
      "tokens_seen": 2555117568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022802407221664993,
      "loss": 1.2758,
      "theoretical_loss": 3.3911415749284894,
      "tokens_seen": 2555183104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022800401203610832,
      "loss": 1.2632,
      "theoretical_loss": 3.3911347254841413,
      "tokens_seen": 2555248640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022798395185556669,
      "loss": 1.2471,
      "theoretical_loss": 3.3911278762646493,
      "tokens_seen": 2555314176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022796389167502508,
      "loss": 1.2772,
      "theoretical_loss": 3.3911210272700005,
      "tokens_seen": 2555379712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022794383149448344,
      "loss": 1.2512,
      "theoretical_loss": 3.3911141785001817,
      "tokens_seen": 2555445248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022792377131394183,
      "loss": 1.2253,
      "theoretical_loss": 3.3911073299551795,
      "tokens_seen": 2555510784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002279037111334002,
      "loss": 1.2573,
      "theoretical_loss": 3.391100481634981,
      "tokens_seen": 2555576320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022788365095285858,
      "loss": 1.2862,
      "theoretical_loss": 3.3910936335395734,
      "tokens_seen": 2555641856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022786359077231694,
      "loss": 1.2627,
      "theoretical_loss": 3.391086785668943,
      "tokens_seen": 2555707392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022784353059177536,
      "loss": 1.2816,
      "theoretical_loss": 3.3910799380230765,
      "tokens_seen": 2555772928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022782347041123372,
      "loss": 1.2736,
      "theoretical_loss": 3.391073090601961,
      "tokens_seen": 2555838464
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022780341023069208,
      "loss": 1.3276,
      "theoretical_loss": 3.391066243405584,
      "tokens_seen": 2555904000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022778335005015047,
      "loss": 1.3233,
      "theoretical_loss": 3.3910593964339313,
      "tokens_seen": 2555969536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022776328986960883,
      "loss": 1.2602,
      "theoretical_loss": 3.3910525496869908,
      "tokens_seen": 2556035072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022774322968906722,
      "loss": 1.244,
      "theoretical_loss": 3.3910457031647483,
      "tokens_seen": 2556100608
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022772316950852558,
      "loss": 1.3074,
      "theoretical_loss": 3.3910388568671914,
      "tokens_seen": 2556166144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022770310932798397,
      "loss": 1.1998,
      "theoretical_loss": 3.391032010794307,
      "tokens_seen": 2556231680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022768304914744234,
      "loss": 1.2459,
      "theoretical_loss": 3.3910251649460816,
      "tokens_seen": 2556297216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022766298896690073,
      "loss": 1.276,
      "theoretical_loss": 3.391018319322502,
      "tokens_seen": 2556362752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002276429287863591,
      "loss": 1.3326,
      "theoretical_loss": 3.3910114739235553,
      "tokens_seen": 2556428288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022762286860581745,
      "loss": 1.2504,
      "theoretical_loss": 3.391004628749229,
      "tokens_seen": 2556493824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022760280842527584,
      "loss": 1.3782,
      "theoretical_loss": 3.3909977837995084,
      "tokens_seen": 2556559360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002275827482447342,
      "loss": 1.2602,
      "theoretical_loss": 3.3909909390743818,
      "tokens_seen": 2556624896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002275626880641926,
      "loss": 1.258,
      "theoretical_loss": 3.3909840945738354,
      "tokens_seen": 2556690432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022754262788365095,
      "loss": 1.3264,
      "theoretical_loss": 3.390977250297856,
      "tokens_seen": 2556755968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022752256770310934,
      "loss": 1.2632,
      "theoretical_loss": 3.390970406246431,
      "tokens_seen": 2556821504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002275025075225677,
      "loss": 1.3073,
      "theoretical_loss": 3.390963562419547,
      "tokens_seen": 2556887040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002274824473420261,
      "loss": 1.2205,
      "theoretical_loss": 3.3909567188171907,
      "tokens_seen": 2556952576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022746238716148446,
      "loss": 1.2598,
      "theoretical_loss": 3.3909498754393494,
      "tokens_seen": 2557018112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022744232698094282,
      "loss": 1.2794,
      "theoretical_loss": 3.3909430322860095,
      "tokens_seen": 2557083648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002274222668004012,
      "loss": 1.2823,
      "theoretical_loss": 3.3909361893571583,
      "tokens_seen": 2557149184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022740220661985957,
      "loss": 1.3034,
      "theoretical_loss": 3.3909293466527823,
      "tokens_seen": 2557214720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022738214643931796,
      "loss": 1.2941,
      "theoretical_loss": 3.3909225041728686,
      "tokens_seen": 2557280256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022736208625877632,
      "loss": 1.2401,
      "theoretical_loss": 3.390915661917404,
      "tokens_seen": 2557345792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002273420260782347,
      "loss": 1.5283,
      "theoretical_loss": 3.3909088198863757,
      "tokens_seen": 2557411328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022732196589769307,
      "loss": 1.2034,
      "theoretical_loss": 3.3909019780797705,
      "tokens_seen": 2557476864
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022730190571715146,
      "loss": 1.3338,
      "theoretical_loss": 3.3908951364975746,
      "tokens_seen": 2557542400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022728184553660983,
      "loss": 1.288,
      "theoretical_loss": 3.3908882951397756,
      "tokens_seen": 2557607936
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 930481,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1277267932891846,
      "objective/train/theoretical_loss": 3.3908814540063603,
      "objective/train/tokens_used": 928132576,
      "theoretical_loss": 3.3908814540063603,
      "tokens_seen": 2557673472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022726178535606821,
      "loss": 1.1303,
      "theoretical_loss": 3.3908814540063603,
      "tokens_seen": 2557673472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022724172517552658,
      "loss": 1.3335,
      "theoretical_loss": 3.3908746130973153,
      "tokens_seen": 2557739008
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022722166499498494,
      "loss": 1.192,
      "theoretical_loss": 3.390867772412628,
      "tokens_seen": 2557804544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022720160481444333,
      "loss": 1.2036,
      "theoretical_loss": 3.3908609319522847,
      "tokens_seen": 2557870080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002271815446339017,
      "loss": 1.1932,
      "theoretical_loss": 3.390854091716273,
      "tokens_seen": 2557935616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022716148445336008,
      "loss": 1.2301,
      "theoretical_loss": 3.390847251704579,
      "tokens_seen": 2558001152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022714142427281844,
      "loss": 1.3051,
      "theoretical_loss": 3.3908404119171895,
      "tokens_seen": 2558066688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022712136409227683,
      "loss": 1.2399,
      "theoretical_loss": 3.3908335723540928,
      "tokens_seen": 2558132224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022710130391173522,
      "loss": 1.3998,
      "theoretical_loss": 3.3908267330152744,
      "tokens_seen": 2558197760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002270812437311936,
      "loss": 1.2031,
      "theoretical_loss": 3.3908198939007215,
      "tokens_seen": 2558263296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022706118355065197,
      "loss": 1.1693,
      "theoretical_loss": 3.390813055010421,
      "tokens_seen": 2558328832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022704112337011033,
      "loss": 1.3285,
      "theoretical_loss": 3.3908062163443606,
      "tokens_seen": 2558394368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022702106318956872,
      "loss": 1.2584,
      "theoretical_loss": 3.3907993779025265,
      "tokens_seen": 2558459904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022700100300902709,
      "loss": 1.2561,
      "theoretical_loss": 3.390792539684905,
      "tokens_seen": 2558525440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022698094282848548,
      "loss": 1.3101,
      "theoretical_loss": 3.3907857016914846,
      "tokens_seen": 2558590976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022696088264794384,
      "loss": 1.226,
      "theoretical_loss": 3.3907788639222507,
      "tokens_seen": 2558656512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022694082246740223,
      "loss": 1.2375,
      "theoretical_loss": 3.390772026377191,
      "tokens_seen": 2558722048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002269207622868606,
      "loss": 1.2791,
      "theoretical_loss": 3.390765189056292,
      "tokens_seen": 2558787584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022690070210631898,
      "loss": 1.253,
      "theoretical_loss": 3.390758351959541,
      "tokens_seen": 2558853120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022688064192577734,
      "loss": 1.3176,
      "theoretical_loss": 3.3907515150869245,
      "tokens_seen": 2558918656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002268605817452357,
      "loss": 1.2282,
      "theoretical_loss": 3.39074467843843,
      "tokens_seen": 2558984192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002268405215646941,
      "loss": 1.2043,
      "theoretical_loss": 3.390737842014044,
      "tokens_seen": 2559049728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022682046138415245,
      "loss": 1.2071,
      "theoretical_loss": 3.390731005813753,
      "tokens_seen": 2559115264
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022680040120361084,
      "loss": 1.2949,
      "theoretical_loss": 3.390724169837545,
      "tokens_seen": 2559180800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002267803410230692,
      "loss": 1.2831,
      "theoretical_loss": 3.390717334085406,
      "tokens_seen": 2559246336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002267602808425276,
      "loss": 1.2761,
      "theoretical_loss": 3.390710498557323,
      "tokens_seen": 2559311872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022674022066198596,
      "loss": 1.2964,
      "theoretical_loss": 3.3907036632532837,
      "tokens_seen": 2559377408
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022672016048144435,
      "loss": 1.2564,
      "theoretical_loss": 3.390696828173274,
      "tokens_seen": 2559442944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002267001003009027,
      "loss": 1.2203,
      "theoretical_loss": 3.3906899933172814,
      "tokens_seen": 2559508480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002266800401203611,
      "loss": 1.1786,
      "theoretical_loss": 3.3906831586852926,
      "tokens_seen": 2559574016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022665997993981946,
      "loss": 1.3037,
      "theoretical_loss": 3.390676324277295,
      "tokens_seen": 2559639552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022663991975927782,
      "loss": 1.2932,
      "theoretical_loss": 3.3906694900932743,
      "tokens_seen": 2559705088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002266198595787362,
      "loss": 1.2071,
      "theoretical_loss": 3.390662656133219,
      "tokens_seen": 2559770624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022659979939819458,
      "loss": 1.2184,
      "theoretical_loss": 3.390655822397115,
      "tokens_seen": 2559836160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022657973921765296,
      "loss": 1.2582,
      "theoretical_loss": 3.3906489888849496,
      "tokens_seen": 2559901696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022655967903711133,
      "loss": 1.31,
      "theoretical_loss": 3.39064215559671,
      "tokens_seen": 2559967232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022653961885656972,
      "loss": 1.2104,
      "theoretical_loss": 3.3906353225323826,
      "tokens_seen": 2560032768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022651955867602808,
      "loss": 1.3441,
      "theoretical_loss": 3.3906284896919545,
      "tokens_seen": 2560098304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022649949849548647,
      "loss": 1.2178,
      "theoretical_loss": 3.390621657075412,
      "tokens_seen": 2560163840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022647943831494483,
      "loss": 1.3156,
      "theoretical_loss": 3.3906148246827437,
      "tokens_seen": 2560229376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002264593781344032,
      "loss": 1.1912,
      "theoretical_loss": 3.390607992513935,
      "tokens_seen": 2560294912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022643931795386158,
      "loss": 1.312,
      "theoretical_loss": 3.3906011605689734,
      "tokens_seen": 2560360448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022641925777331994,
      "loss": 1.2481,
      "theoretical_loss": 3.390594328847846,
      "tokens_seen": 2560425984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022639919759277833,
      "loss": 1.3445,
      "theoretical_loss": 3.3905874973505394,
      "tokens_seen": 2560491520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002263791374122367,
      "loss": 1.19,
      "theoretical_loss": 3.3905806660770406,
      "tokens_seen": 2560557056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022635907723169508,
      "loss": 1.3432,
      "theoretical_loss": 3.3905738350273364,
      "tokens_seen": 2560622592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022633901705115347,
      "loss": 1.2233,
      "theoretical_loss": 3.390567004201414,
      "tokens_seen": 2560688128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022631895687061186,
      "loss": 1.2607,
      "theoretical_loss": 3.3905601735992605,
      "tokens_seen": 2560753664
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022629889669007023,
      "loss": 1.2545,
      "theoretical_loss": 3.390553343220863,
      "tokens_seen": 2560819200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022627883650952861,
      "loss": 1.268,
      "theoretical_loss": 3.3905465130662074,
      "tokens_seen": 2560884736
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 931622,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.0067193508148193,
      "objective/train/theoretical_loss": 3.3905396831352816,
      "objective/train/tokens_used": 931409376,
      "theoretical_loss": 3.3905396831352816,
      "tokens_seen": 2560950272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022625877632898698,
      "loss": 1.0906,
      "theoretical_loss": 3.3905396831352816,
      "tokens_seen": 2560950272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022623871614844534,
      "loss": 1.2708,
      "theoretical_loss": 3.3905328534280725,
      "tokens_seen": 2561015808
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022621865596790373,
      "loss": 1.2504,
      "theoretical_loss": 3.3905260239445663,
      "tokens_seen": 2561081344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002261985957873621,
      "loss": 1.1916,
      "theoretical_loss": 3.390519194684751,
      "tokens_seen": 2561146880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022617853560682048,
      "loss": 1.1895,
      "theoretical_loss": 3.3905123656486125,
      "tokens_seen": 2561212416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022615847542627884,
      "loss": 1.2994,
      "theoretical_loss": 3.3905055368361383,
      "tokens_seen": 2561277952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022613841524573723,
      "loss": 1.2573,
      "theoretical_loss": 3.390498708247316,
      "tokens_seen": 2561343488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002261183550651956,
      "loss": 1.4366,
      "theoretical_loss": 3.390491879882131,
      "tokens_seen": 2561409024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022609829488465398,
      "loss": 1.1878,
      "theoretical_loss": 3.390485051740572,
      "tokens_seen": 2561474560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022607823470411235,
      "loss": 1.2546,
      "theoretical_loss": 3.3904782238226243,
      "tokens_seen": 2561540096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002260581745235707,
      "loss": 1.3899,
      "theoretical_loss": 3.3904713961282757,
      "tokens_seen": 2561605632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002260381143430291,
      "loss": 1.1874,
      "theoretical_loss": 3.3904645686575137,
      "tokens_seen": 2561671168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022601805416248746,
      "loss": 1.1859,
      "theoretical_loss": 3.390457741410324,
      "tokens_seen": 2561736704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022599799398194585,
      "loss": 1.2505,
      "theoretical_loss": 3.3904509143866948,
      "tokens_seen": 2561802240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002259779338014042,
      "loss": 1.2346,
      "theoretical_loss": 3.390444087586612,
      "tokens_seen": 2561867776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002259578736208626,
      "loss": 1.2056,
      "theoretical_loss": 3.3904372610100633,
      "tokens_seen": 2561933312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022593781344032096,
      "loss": 1.3845,
      "theoretical_loss": 3.390430434657035,
      "tokens_seen": 2561998848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022591775325977935,
      "loss": 1.2934,
      "theoretical_loss": 3.390423608527515,
      "tokens_seen": 2562064384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022589769307923771,
      "loss": 1.2505,
      "theoretical_loss": 3.3904167826214895,
      "tokens_seen": 2562129920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022587763289869608,
      "loss": 1.267,
      "theoretical_loss": 3.3904099569389454,
      "tokens_seen": 2562195456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022585757271815447,
      "loss": 1.1387,
      "theoretical_loss": 3.3904031314798706,
      "tokens_seen": 2562260992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022583751253761283,
      "loss": 1.2045,
      "theoretical_loss": 3.390396306244251,
      "tokens_seen": 2562326528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022581745235707122,
      "loss": 1.3322,
      "theoretical_loss": 3.3903894812320736,
      "tokens_seen": 2562392064
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022579739217652958,
      "loss": 1.253,
      "theoretical_loss": 3.390382656443326,
      "tokens_seen": 2562457600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022577733199598797,
      "loss": 1.3081,
      "theoretical_loss": 3.3903758318779955,
      "tokens_seen": 2562523136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022575727181544633,
      "loss": 1.37,
      "theoretical_loss": 3.3903690075360684,
      "tokens_seen": 2562588672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022573721163490472,
      "loss": 1.3448,
      "theoretical_loss": 3.3903621834175315,
      "tokens_seen": 2562654208
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022571715145436308,
      "loss": 1.2017,
      "theoretical_loss": 3.390355359522372,
      "tokens_seen": 2562719744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022569709127382147,
      "loss": 1.2806,
      "theoretical_loss": 3.3903485358505767,
      "tokens_seen": 2562785280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022567703109327983,
      "loss": 1.2277,
      "theoretical_loss": 3.390341712402133,
      "tokens_seen": 2562850816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002256569709127382,
      "loss": 1.1669,
      "theoretical_loss": 3.390334889177028,
      "tokens_seen": 2562916352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022563691073219659,
      "loss": 1.3417,
      "theoretical_loss": 3.3903280661752486,
      "tokens_seen": 2562981888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022561685055165495,
      "loss": 1.2812,
      "theoretical_loss": 3.390321243396781,
      "tokens_seen": 2563047424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022559679037111336,
      "loss": 1.4431,
      "theoretical_loss": 3.3903144208416127,
      "tokens_seen": 2563112960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022557673019057173,
      "loss": 1.301,
      "theoretical_loss": 3.390307598509731,
      "tokens_seen": 2563178496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022555667001003012,
      "loss": 1.3115,
      "theoretical_loss": 3.390300776401123,
      "tokens_seen": 2563244032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022553660982948848,
      "loss": 1.3287,
      "theoretical_loss": 3.390293954515774,
      "tokens_seen": 2563309568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022551654964894687,
      "loss": 1.2775,
      "theoretical_loss": 3.3902871328536732,
      "tokens_seen": 2563375104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022549648946840523,
      "loss": 1.2908,
      "theoretical_loss": 3.3902803114148066,
      "tokens_seen": 2563440640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002254764292878636,
      "loss": 1.278,
      "theoretical_loss": 3.390273490199161,
      "tokens_seen": 2563506176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022545636910732198,
      "loss": 1.3345,
      "theoretical_loss": 3.390266669206724,
      "tokens_seen": 2563571712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022543630892678034,
      "loss": 1.2854,
      "theoretical_loss": 3.390259848437482,
      "tokens_seen": 2563637248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022541624874623873,
      "loss": 1.3501,
      "theoretical_loss": 3.390253027891422,
      "tokens_seen": 2563702784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002253961885656971,
      "loss": 1.2909,
      "theoretical_loss": 3.3902462075685316,
      "tokens_seen": 2563768320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022537612838515548,
      "loss": 1.2359,
      "theoretical_loss": 3.390239387468797,
      "tokens_seen": 2563833856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022535606820461385,
      "loss": 1.1691,
      "theoretical_loss": 3.390232567592206,
      "tokens_seen": 2563899392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022533600802407224,
      "loss": 1.3987,
      "theoretical_loss": 3.3902257479387448,
      "tokens_seen": 2563964928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002253159478435306,
      "loss": 1.2887,
      "theoretical_loss": 3.390218928508401,
      "tokens_seen": 2564030464
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022529588766298896,
      "loss": 1.1763,
      "theoretical_loss": 3.3902121093011615,
      "tokens_seen": 2564096000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022527582748244735,
      "loss": 1.2393,
      "theoretical_loss": 3.390205290317013,
      "tokens_seen": 2564161536
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 932851,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3499561548233032,
      "objective/train/theoretical_loss": 3.3901984715559426,
      "objective/train/tokens_used": 934686176,
      "theoretical_loss": 3.3901984715559426,
      "tokens_seen": 2564227072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002252557673019057,
      "loss": 1.3266,
      "theoretical_loss": 3.3901984715559426,
      "tokens_seen": 2564227072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002252357071213641,
      "loss": 1.2986,
      "theoretical_loss": 3.3901916530179372,
      "tokens_seen": 2564292608
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022521564694082246,
      "loss": 1.2773,
      "theoretical_loss": 3.3901848347029846,
      "tokens_seen": 2564358144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022519558676028085,
      "loss": 1.3602,
      "theoretical_loss": 3.390178016611071,
      "tokens_seen": 2564423680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022517552657973922,
      "loss": 1.4323,
      "theoretical_loss": 3.390171198742183,
      "tokens_seen": 2564489216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002251554663991976,
      "loss": 1.2046,
      "theoretical_loss": 3.390164381096309,
      "tokens_seen": 2564554752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022513540621865597,
      "loss": 1.3094,
      "theoretical_loss": 3.390157563673435,
      "tokens_seen": 2564620288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022511534603811436,
      "loss": 1.2662,
      "theoretical_loss": 3.390150746473548,
      "tokens_seen": 2564685824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022509528585757272,
      "loss": 1.3273,
      "theoretical_loss": 3.3901439294966353,
      "tokens_seen": 2564751360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022507522567703108,
      "loss": 1.2659,
      "theoretical_loss": 3.390137112742684,
      "tokens_seen": 2564816896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022505516549648947,
      "loss": 1.1546,
      "theoretical_loss": 3.3901302962116806,
      "tokens_seen": 2564882432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022503510531594783,
      "loss": 1.3612,
      "theoretical_loss": 3.3901234799036124,
      "tokens_seen": 2564947968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022501504513540622,
      "loss": 1.2759,
      "theoretical_loss": 3.3901166638184668,
      "tokens_seen": 2565013504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022499498495486458,
      "loss": 1.3883,
      "theoretical_loss": 3.3901098479562304,
      "tokens_seen": 2565079040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022497492477432297,
      "loss": 1.3977,
      "theoretical_loss": 3.3901030323168904,
      "tokens_seen": 2565144576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022495486459378134,
      "loss": 1.287,
      "theoretical_loss": 3.3900962169004334,
      "tokens_seen": 2565210112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022493480441323973,
      "loss": 1.3046,
      "theoretical_loss": 3.390089401706847,
      "tokens_seen": 2565275648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002249147442326981,
      "loss": 1.2899,
      "theoretical_loss": 3.390082586736118,
      "tokens_seen": 2565341184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022489468405215645,
      "loss": 1.3483,
      "theoretical_loss": 3.390075771988233,
      "tokens_seen": 2565406720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022487462387161484,
      "loss": 1.311,
      "theoretical_loss": 3.3900689574631797,
      "tokens_seen": 2565472256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002248545636910732,
      "loss": 1.3224,
      "theoretical_loss": 3.390062143160945,
      "tokens_seen": 2565537792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022483450351053162,
      "loss": 1.3757,
      "theoretical_loss": 3.390055329081515,
      "tokens_seen": 2565603328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022481444332998998,
      "loss": 1.2855,
      "theoretical_loss": 3.390048515224878,
      "tokens_seen": 2565668864
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022479438314944837,
      "loss": 1.3388,
      "theoretical_loss": 3.3900417015910205,
      "tokens_seen": 2565734400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022477432296890673,
      "loss": 1.2142,
      "theoretical_loss": 3.390034888179929,
      "tokens_seen": 2565799936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022475426278836512,
      "loss": 1.2702,
      "theoretical_loss": 3.3900280749915916,
      "tokens_seen": 2565865472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022473420260782348,
      "loss": 1.1059,
      "theoretical_loss": 3.3900212620259946,
      "tokens_seen": 2565931008
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022471414242728187,
      "loss": 1.2501,
      "theoretical_loss": 3.3900144492831252,
      "tokens_seen": 2565996544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022469408224674023,
      "loss": 1.3234,
      "theoretical_loss": 3.3900076367629706,
      "tokens_seen": 2566062080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002246740220661986,
      "loss": 1.2152,
      "theoretical_loss": 3.3900008244655173,
      "tokens_seen": 2566127616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022465396188565699,
      "loss": 1.2538,
      "theoretical_loss": 3.389994012390753,
      "tokens_seen": 2566193152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022463390170511535,
      "loss": 1.4163,
      "theoretical_loss": 3.3899872005386644,
      "tokens_seen": 2566258688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022461384152457374,
      "loss": 1.2178,
      "theoretical_loss": 3.389980388909238,
      "tokens_seen": 2566324224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002245937813440321,
      "loss": 1.282,
      "theoretical_loss": 3.389973577502462,
      "tokens_seen": 2566389760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002245737211634905,
      "loss": 1.3096,
      "theoretical_loss": 3.389966766318323,
      "tokens_seen": 2566455296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022455366098294885,
      "loss": 1.2495,
      "theoretical_loss": 3.389959955356807,
      "tokens_seen": 2566520832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022453360080240724,
      "loss": 1.2295,
      "theoretical_loss": 3.3899531446179028,
      "tokens_seen": 2566586368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002245135406218656,
      "loss": 1.3284,
      "theoretical_loss": 3.389946334101596,
      "tokens_seen": 2566651904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022449348044132397,
      "loss": 1.2621,
      "theoretical_loss": 3.3899395238078744,
      "tokens_seen": 2566717440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022447342026078236,
      "loss": 1.3465,
      "theoretical_loss": 3.389932713736725,
      "tokens_seen": 2566782976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022445336008024072,
      "loss": 1.3138,
      "theoretical_loss": 3.389925903888134,
      "tokens_seen": 2566848512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002244332998996991,
      "loss": 1.3155,
      "theoretical_loss": 3.3899190942620896,
      "tokens_seen": 2566914048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022441323971915747,
      "loss": 1.4059,
      "theoretical_loss": 3.3899122848585783,
      "tokens_seen": 2566979584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022439317953861586,
      "loss": 1.4369,
      "theoretical_loss": 3.3899054756775873,
      "tokens_seen": 2567045120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022437311935807422,
      "loss": 1.1655,
      "theoretical_loss": 3.3898986667191036,
      "tokens_seen": 2567110656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002243530591775326,
      "loss": 1.347,
      "theoretical_loss": 3.389891857983114,
      "tokens_seen": 2567176192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022433299899699097,
      "loss": 1.3014,
      "theoretical_loss": 3.389885049469606,
      "tokens_seen": 2567241728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022431293881644933,
      "loss": 1.2648,
      "theoretical_loss": 3.3898782411785664,
      "tokens_seen": 2567307264
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022429287863590772,
      "loss": 1.2783,
      "theoretical_loss": 3.389871433109982,
      "tokens_seen": 2567372800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022427281845536609,
      "loss": 1.2455,
      "theoretical_loss": 3.38986462526384,
      "tokens_seen": 2567438336
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 933971,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1539313793182373,
      "objective/train/theoretical_loss": 3.389857817640128,
      "objective/train/tokens_used": 937962976,
      "theoretical_loss": 3.389857817640128,
      "tokens_seen": 2567503872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022425275827482448,
      "loss": 1.2347,
      "theoretical_loss": 3.389857817640128,
      "tokens_seen": 2567503872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022423269809428284,
      "loss": 1.3264,
      "theoretical_loss": 3.389851010238832,
      "tokens_seen": 2567569408
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022421263791374123,
      "loss": 1.4469,
      "theoretical_loss": 3.3898442030599405,
      "tokens_seen": 2567634944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002241925777331996,
      "loss": 1.3354,
      "theoretical_loss": 3.3898373961034394,
      "tokens_seen": 2567700480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022417251755265798,
      "loss": 1.323,
      "theoretical_loss": 3.389830589369316,
      "tokens_seen": 2567766016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022415245737211634,
      "loss": 1.3827,
      "theoretical_loss": 3.3898237828575573,
      "tokens_seen": 2567831552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002241323971915747,
      "loss": 1.3502,
      "theoretical_loss": 3.389816976568151,
      "tokens_seen": 2567897088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002241123370110331,
      "loss": 1.2593,
      "theoretical_loss": 3.3898101705010832,
      "tokens_seen": 2567962624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022409227683049148,
      "loss": 1.224,
      "theoretical_loss": 3.3898033646563417,
      "tokens_seen": 2568028160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022407221664994987,
      "loss": 1.308,
      "theoretical_loss": 3.389796559033913,
      "tokens_seen": 2568093696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022405215646940823,
      "loss": 1.2603,
      "theoretical_loss": 3.389789753633785,
      "tokens_seen": 2568159232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022403209628886662,
      "loss": 1.3496,
      "theoretical_loss": 3.389782948455944,
      "tokens_seen": 2568224768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022401203610832498,
      "loss": 1.2856,
      "theoretical_loss": 3.3897761435003773,
      "tokens_seen": 2568290304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022399197592778337,
      "loss": 1.3056,
      "theoretical_loss": 3.3897693387670715,
      "tokens_seen": 2568355840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022397191574724174,
      "loss": 1.4114,
      "theoretical_loss": 3.3897625342560147,
      "tokens_seen": 2568421376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022395185556670013,
      "loss": 1.2533,
      "theoretical_loss": 3.3897557299671934,
      "tokens_seen": 2568486912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002239317953861585,
      "loss": 1.3902,
      "theoretical_loss": 3.3897489259005944,
      "tokens_seen": 2568552448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022391173520561685,
      "loss": 1.3716,
      "theoretical_loss": 3.389742122056205,
      "tokens_seen": 2568617984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022389167502507524,
      "loss": 1.3404,
      "theoretical_loss": 3.389735318434013,
      "tokens_seen": 2568683520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002238716148445336,
      "loss": 1.2121,
      "theoretical_loss": 3.389728515034004,
      "tokens_seen": 2568749056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000223851554663992,
      "loss": 1.3619,
      "theoretical_loss": 3.389721711856166,
      "tokens_seen": 2568814592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022383149448345035,
      "loss": 1.3159,
      "theoretical_loss": 3.3897149089004865,
      "tokens_seen": 2568880128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022381143430290874,
      "loss": 1.3131,
      "theoretical_loss": 3.3897081061669514,
      "tokens_seen": 2568945664
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002237913741223671,
      "loss": 1.2708,
      "theoretical_loss": 3.389701303655549,
      "tokens_seen": 2569011200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002237713139418255,
      "loss": 1.3967,
      "theoretical_loss": 3.3896945013662654,
      "tokens_seen": 2569076736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022375125376128386,
      "loss": 1.2901,
      "theoretical_loss": 3.3896876992990883,
      "tokens_seen": 2569142272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022373119358074222,
      "loss": 1.3518,
      "theoretical_loss": 3.3896808974540042,
      "tokens_seen": 2569207808
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002237111334002006,
      "loss": 1.2018,
      "theoretical_loss": 3.389674095831001,
      "tokens_seen": 2569273344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022369107321965897,
      "loss": 1.3063,
      "theoretical_loss": 3.389667294430065,
      "tokens_seen": 2569338880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022367101303911736,
      "loss": 1.2333,
      "theoretical_loss": 3.389660493251184,
      "tokens_seen": 2569404416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022365095285857572,
      "loss": 1.2359,
      "theoretical_loss": 3.389653692294344,
      "tokens_seen": 2569469952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002236308926780341,
      "loss": 1.4002,
      "theoretical_loss": 3.3896468915595337,
      "tokens_seen": 2569535488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022361083249749247,
      "loss": 1.2422,
      "theoretical_loss": 3.3896400910467386,
      "tokens_seen": 2569601024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022359077231695086,
      "loss": 1.3985,
      "theoretical_loss": 3.3896332907559468,
      "tokens_seen": 2569666560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022357071213640923,
      "loss": 1.2322,
      "theoretical_loss": 3.389626490687145,
      "tokens_seen": 2569732096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022355065195586761,
      "loss": 1.4224,
      "theoretical_loss": 3.3896196908403198,
      "tokens_seen": 2569797632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022353059177532598,
      "loss": 1.399,
      "theoretical_loss": 3.3896128912154593,
      "tokens_seen": 2569863168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022351053159478434,
      "loss": 1.3241,
      "theoretical_loss": 3.38960609181255,
      "tokens_seen": 2569928704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022349047141424273,
      "loss": 1.3168,
      "theoretical_loss": 3.38959929263158,
      "tokens_seen": 2569994240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002234704112337011,
      "loss": 1.3395,
      "theoretical_loss": 3.3895924936725343,
      "tokens_seen": 2570059776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022345035105315948,
      "loss": 1.3385,
      "theoretical_loss": 3.389585694935402,
      "tokens_seen": 2570125312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022343029087261784,
      "loss": 1.2579,
      "theoretical_loss": 3.389578896420169,
      "tokens_seen": 2570190848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022341023069207623,
      "loss": 1.3115,
      "theoretical_loss": 3.389572098126823,
      "tokens_seen": 2570256384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002233901705115346,
      "loss": 1.3606,
      "theoretical_loss": 3.389565300055351,
      "tokens_seen": 2570321920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022337011033099298,
      "loss": 1.3374,
      "theoretical_loss": 3.3895585022057397,
      "tokens_seen": 2570387456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022335005015045135,
      "loss": 1.4263,
      "theoretical_loss": 3.389551704577977,
      "tokens_seen": 2570452992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022332998996990973,
      "loss": 1.2502,
      "theoretical_loss": 3.3895449071720494,
      "tokens_seen": 2570518528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022330992978936812,
      "loss": 1.299,
      "theoretical_loss": 3.389538109987944,
      "tokens_seen": 2570584064
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022328986960882649,
      "loss": 1.2784,
      "theoretical_loss": 3.389531313025648,
      "tokens_seen": 2570649600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022326980942828488,
      "loss": 1.2771,
      "theoretical_loss": 3.3895245162851486,
      "tokens_seen": 2570715136
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 935107,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1545178890228271,
      "objective/train/theoretical_loss": 3.389517719766433,
      "objective/train/tokens_used": 941239776,
      "theoretical_loss": 3.389517719766433,
      "tokens_seen": 2570780672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022324974924774324,
      "loss": 1.2435,
      "theoretical_loss": 3.389517719766433,
      "tokens_seen": 2570780672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022322968906720163,
      "loss": 1.4146,
      "theoretical_loss": 3.389510923469488,
      "tokens_seen": 2570846208
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022320962888666,
      "loss": 1.4585,
      "theoretical_loss": 3.389504127394301,
      "tokens_seen": 2570911744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022318956870611838,
      "loss": 1.2648,
      "theoretical_loss": 3.3894973315408587,
      "tokens_seen": 2570977280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022316950852557674,
      "loss": 1.2951,
      "theoretical_loss": 3.389490535909149,
      "tokens_seen": 2571042816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002231494483450351,
      "loss": 1.26,
      "theoretical_loss": 3.3894837404991582,
      "tokens_seen": 2571108352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002231293881644935,
      "loss": 1.3512,
      "theoretical_loss": 3.3894769453108737,
      "tokens_seen": 2571173888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022310932798395185,
      "loss": 1.234,
      "theoretical_loss": 3.3894701503442826,
      "tokens_seen": 2571239424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022308926780341024,
      "loss": 1.2663,
      "theoretical_loss": 3.3894633555993723,
      "tokens_seen": 2571304960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002230692076228686,
      "loss": 1.3575,
      "theoretical_loss": 3.3894565610761296,
      "tokens_seen": 2571370496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000223049147442327,
      "loss": 1.3399,
      "theoretical_loss": 3.3894497667745416,
      "tokens_seen": 2571436032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022302908726178536,
      "loss": 1.3957,
      "theoretical_loss": 3.389442972694596,
      "tokens_seen": 2571501568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022300902708124375,
      "loss": 1.2917,
      "theoretical_loss": 3.389436178836279,
      "tokens_seen": 2571567104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002229889669007021,
      "loss": 1.4657,
      "theoretical_loss": 3.3894293851995783,
      "tokens_seen": 2571632640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002229689067201605,
      "loss": 1.2593,
      "theoretical_loss": 3.3894225917844807,
      "tokens_seen": 2571698176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022294884653961886,
      "loss": 1.3117,
      "theoretical_loss": 3.3894157985909734,
      "tokens_seen": 2571763712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022292878635907722,
      "loss": 1.2708,
      "theoretical_loss": 3.389409005619044,
      "tokens_seen": 2571829248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002229087261785356,
      "loss": 1.2876,
      "theoretical_loss": 3.389402212868679,
      "tokens_seen": 2571894784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022288866599799397,
      "loss": 1.4493,
      "theoretical_loss": 3.3893954203398664,
      "tokens_seen": 2571960320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022286860581745236,
      "loss": 1.2858,
      "theoretical_loss": 3.3893886280325924,
      "tokens_seen": 2572025856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022284854563691073,
      "loss": 1.4089,
      "theoretical_loss": 3.3893818359468444,
      "tokens_seen": 2572091392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022282848545636912,
      "loss": 1.3053,
      "theoretical_loss": 3.389375044082609,
      "tokens_seen": 2572156928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022280842527582748,
      "loss": 1.288,
      "theoretical_loss": 3.3893682524398745,
      "tokens_seen": 2572222464
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022278836509528587,
      "loss": 1.3242,
      "theoretical_loss": 3.3893614610186273,
      "tokens_seen": 2572288000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022276830491474423,
      "loss": 1.2921,
      "theoretical_loss": 3.389354669818855,
      "tokens_seen": 2572353536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002227482447342026,
      "loss": 1.3193,
      "theoretical_loss": 3.3893478788405442,
      "tokens_seen": 2572419072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022272818455366098,
      "loss": 1.3499,
      "theoretical_loss": 3.3893410880836825,
      "tokens_seen": 2572484608
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022270812437311934,
      "loss": 1.3765,
      "theoretical_loss": 3.389334297548256,
      "tokens_seen": 2572550144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022268806419257773,
      "loss": 1.292,
      "theoretical_loss": 3.3893275072342535,
      "tokens_seen": 2572615680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002226680040120361,
      "loss": 1.2545,
      "theoretical_loss": 3.389320717141661,
      "tokens_seen": 2572681216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022264794383149448,
      "loss": 1.3835,
      "theoretical_loss": 3.3893139272704658,
      "tokens_seen": 2572746752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022262788365095285,
      "loss": 1.2129,
      "theoretical_loss": 3.389307137620655,
      "tokens_seen": 2572812288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022260782347041124,
      "loss": 1.3797,
      "theoretical_loss": 3.3893003481922164,
      "tokens_seen": 2572877824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022258776328986963,
      "loss": 1.2878,
      "theoretical_loss": 3.3892935589851363,
      "tokens_seen": 2572943360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022256770310932801,
      "loss": 1.1948,
      "theoretical_loss": 3.3892867699994023,
      "tokens_seen": 2573008896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022254764292878638,
      "loss": 1.3066,
      "theoretical_loss": 3.389279981235001,
      "tokens_seen": 2573074432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022252758274824474,
      "loss": 1.3384,
      "theoretical_loss": 3.3892731926919204,
      "tokens_seen": 2573139968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022250752256770313,
      "loss": 1.3614,
      "theoretical_loss": 3.3892664043701473,
      "tokens_seen": 2573205504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002224874623871615,
      "loss": 1.3016,
      "theoretical_loss": 3.3892596162696687,
      "tokens_seen": 2573271040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022246740220661988,
      "loss": 1.2924,
      "theoretical_loss": 3.3892528283904717,
      "tokens_seen": 2573336576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022244734202607824,
      "loss": 1.3467,
      "theoretical_loss": 3.3892460407325435,
      "tokens_seen": 2573402112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022242728184553663,
      "loss": 1.1972,
      "theoretical_loss": 3.3892392532958717,
      "tokens_seen": 2573467648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000222407221664995,
      "loss": 1.3852,
      "theoretical_loss": 3.389232466080443,
      "tokens_seen": 2573533184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022238716148445338,
      "loss": 1.2848,
      "theoretical_loss": 3.3892256790862447,
      "tokens_seen": 2573598720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022236710130391175,
      "loss": 1.3283,
      "theoretical_loss": 3.3892188923132633,
      "tokens_seen": 2573664256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002223470411233701,
      "loss": 1.2935,
      "theoretical_loss": 3.3892121057614872,
      "tokens_seen": 2573729792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002223269809428285,
      "loss": 1.3555,
      "theoretical_loss": 3.3892053194309026,
      "tokens_seen": 2573795328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022230692076228686,
      "loss": 1.3757,
      "theoretical_loss": 3.389198533321497,
      "tokens_seen": 2573860864
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022228686058174525,
      "loss": 1.3584,
      "theoretical_loss": 3.389191747433258,
      "tokens_seen": 2573926400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002222668004012036,
      "loss": 1.2843,
      "theoretical_loss": 3.3891849617661713,
      "tokens_seen": 2573991936
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 936262,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2910391092300415,
      "objective/train/theoretical_loss": 3.3891781763202258,
      "objective/train/tokens_used": 944516576,
      "theoretical_loss": 3.3891781763202258,
      "tokens_seen": 2574057472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000222246740220662,
      "loss": 1.2646,
      "theoretical_loss": 3.3891781763202258,
      "tokens_seen": 2574057472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022222668004012036,
      "loss": 1.3274,
      "theoretical_loss": 3.3891713910954078,
      "tokens_seen": 2574123008
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022220661985957875,
      "loss": 1.1979,
      "theoretical_loss": 3.3891646060917044,
      "tokens_seen": 2574188544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022218655967903711,
      "loss": 1.4173,
      "theoretical_loss": 3.389157821309103,
      "tokens_seen": 2574254080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022216649949849548,
      "loss": 1.3303,
      "theoretical_loss": 3.3891510367475908,
      "tokens_seen": 2574319616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022214643931795387,
      "loss": 1.3349,
      "theoretical_loss": 3.3891442524071547,
      "tokens_seen": 2574385152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022212637913741223,
      "loss": 1.2275,
      "theoretical_loss": 3.389137468287782,
      "tokens_seen": 2574450688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022210631895687062,
      "loss": 1.24,
      "theoretical_loss": 3.3891306843894604,
      "tokens_seen": 2574516224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022208625877632898,
      "loss": 1.2977,
      "theoretical_loss": 3.389123900712176,
      "tokens_seen": 2574581760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022206619859578737,
      "loss": 1.2669,
      "theoretical_loss": 3.3891171172559167,
      "tokens_seen": 2574647296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022204613841524573,
      "loss": 1.2546,
      "theoretical_loss": 3.38911033402067,
      "tokens_seen": 2574712832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022202607823470412,
      "loss": 1.254,
      "theoretical_loss": 3.3891035510064222,
      "tokens_seen": 2574778368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022200601805416248,
      "loss": 1.4988,
      "theoretical_loss": 3.389096768213161,
      "tokens_seen": 2574843904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022198595787362084,
      "loss": 1.2963,
      "theoretical_loss": 3.389089985640873,
      "tokens_seen": 2574909440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022196589769307923,
      "loss": 1.3037,
      "theoretical_loss": 3.3890832032895464,
      "tokens_seen": 2574974976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002219458375125376,
      "loss": 1.3525,
      "theoretical_loss": 3.3890764211591677,
      "tokens_seen": 2575040512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022192577733199599,
      "loss": 1.3097,
      "theoretical_loss": 3.3890696392497244,
      "tokens_seen": 2575106048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022190571715145435,
      "loss": 1.2034,
      "theoretical_loss": 3.389062857561203,
      "tokens_seen": 2575171584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022188565697091274,
      "loss": 1.3004,
      "theoretical_loss": 3.3890560760935915,
      "tokens_seen": 2575237120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002218655967903711,
      "loss": 1.3135,
      "theoretical_loss": 3.3890492948468767,
      "tokens_seen": 2575302656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022184553660982952,
      "loss": 1.2574,
      "theoretical_loss": 3.3890425138210456,
      "tokens_seen": 2575368192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022182547642928788,
      "loss": 1.3913,
      "theoretical_loss": 3.3890357330160854,
      "tokens_seen": 2575433728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022180541624874627,
      "loss": 1.3158,
      "theoretical_loss": 3.3890289524319837,
      "tokens_seen": 2575499264
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022178535606820463,
      "loss": 1.4259,
      "theoretical_loss": 3.389022172068728,
      "tokens_seen": 2575564800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000221765295887663,
      "loss": 1.302,
      "theoretical_loss": 3.3890153919263044,
      "tokens_seen": 2575630336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022174523570712138,
      "loss": 1.2542,
      "theoretical_loss": 3.3890086120047007,
      "tokens_seen": 2575695872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022172517552657974,
      "loss": 1.3244,
      "theoretical_loss": 3.389001832303904,
      "tokens_seen": 2575761408
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022170511534603813,
      "loss": 1.2714,
      "theoretical_loss": 3.3889950528239017,
      "tokens_seen": 2575826944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002216850551654965,
      "loss": 1.3623,
      "theoretical_loss": 3.3889882735646806,
      "tokens_seen": 2575892480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022166499498495488,
      "loss": 1.2513,
      "theoretical_loss": 3.3889814945262287,
      "tokens_seen": 2575958016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022164493480441325,
      "loss": 1.2617,
      "theoretical_loss": 3.388974715708532,
      "tokens_seen": 2576023552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022162487462387164,
      "loss": 1.3275,
      "theoretical_loss": 3.3889679371115786,
      "tokens_seen": 2576089088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022160481444333,
      "loss": 1.2478,
      "theoretical_loss": 3.3889611587353556,
      "tokens_seen": 2576154624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022158475426278836,
      "loss": 1.3191,
      "theoretical_loss": 3.3889543805798494,
      "tokens_seen": 2576220160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022156469408224675,
      "loss": 1.2841,
      "theoretical_loss": 3.388947602645048,
      "tokens_seen": 2576285696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002215446339017051,
      "loss": 1.2245,
      "theoretical_loss": 3.3889408249309385,
      "tokens_seen": 2576351232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002215245737211635,
      "loss": 1.3988,
      "theoretical_loss": 3.388934047437508,
      "tokens_seen": 2576416768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022150451354062186,
      "loss": 1.4035,
      "theoretical_loss": 3.3889272701647437,
      "tokens_seen": 2576482304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022148445336008025,
      "loss": 1.2229,
      "theoretical_loss": 3.388920493112633,
      "tokens_seen": 2576547840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022146439317953862,
      "loss": 1.2585,
      "theoretical_loss": 3.388913716281163,
      "tokens_seen": 2576613376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000221444332998997,
      "loss": 1.3534,
      "theoretical_loss": 3.3889069396703206,
      "tokens_seen": 2576678912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022142427281845537,
      "loss": 1.408,
      "theoretical_loss": 3.388900163280093,
      "tokens_seen": 2576744448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022140421263791376,
      "loss": 1.2418,
      "theoretical_loss": 3.3888933871104676,
      "tokens_seen": 2576809984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022138415245737212,
      "loss": 1.3517,
      "theoretical_loss": 3.388886611161432,
      "tokens_seen": 2576875520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022136409227683048,
      "loss": 1.3343,
      "theoretical_loss": 3.3888798354329728,
      "tokens_seen": 2576941056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022134403209628887,
      "loss": 1.2922,
      "theoretical_loss": 3.3888730599250776,
      "tokens_seen": 2577006592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022132397191574723,
      "loss": 1.365,
      "theoretical_loss": 3.388866284637733,
      "tokens_seen": 2577072128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022130391173520562,
      "loss": 1.213,
      "theoretical_loss": 3.3888595095709273,
      "tokens_seen": 2577137664
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022128385155466398,
      "loss": 1.3841,
      "theoretical_loss": 3.388852734724647,
      "tokens_seen": 2577203200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022126379137412237,
      "loss": 1.3248,
      "theoretical_loss": 3.388845960098879,
      "tokens_seen": 2577268736
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 937411,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2590935230255127,
      "objective/train/theoretical_loss": 3.3888391856936115,
      "objective/train/tokens_used": 947793376,
      "theoretical_loss": 3.3888391856936115,
      "tokens_seen": 2577334272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022124373119358074,
      "loss": 1.3489,
      "theoretical_loss": 3.3888391856936115,
      "tokens_seen": 2577334272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022122367101303913,
      "loss": 1.1843,
      "theoretical_loss": 3.3888324115088304,
      "tokens_seen": 2577399808
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002212036108324975,
      "loss": 1.1495,
      "theoretical_loss": 3.3888256375445245,
      "tokens_seen": 2577465344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022118355065195585,
      "loss": 1.3111,
      "theoretical_loss": 3.3888188638006795,
      "tokens_seen": 2577530880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022116349047141424,
      "loss": 1.1546,
      "theoretical_loss": 3.3888120902772836,
      "tokens_seen": 2577596416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002211434302908726,
      "loss": 1.2159,
      "theoretical_loss": 3.3888053169743237,
      "tokens_seen": 2577661952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000221123370110331,
      "loss": 1.3333,
      "theoretical_loss": 3.388798543891787,
      "tokens_seen": 2577727488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022110330992978935,
      "loss": 1.4019,
      "theoretical_loss": 3.3887917710296604,
      "tokens_seen": 2577793024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022108324974924777,
      "loss": 1.3249,
      "theoretical_loss": 3.388784998387932,
      "tokens_seen": 2577858560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022106318956870613,
      "loss": 1.3419,
      "theoretical_loss": 3.3887782259665884,
      "tokens_seen": 2577924096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022104312938816452,
      "loss": 1.2986,
      "theoretical_loss": 3.388771453765617,
      "tokens_seen": 2577989632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022102306920762288,
      "loss": 1.3086,
      "theoretical_loss": 3.3887646817850046,
      "tokens_seen": 2578055168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022100300902708125,
      "loss": 1.3599,
      "theoretical_loss": 3.388757910024739,
      "tokens_seen": 2578120704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022098294884653963,
      "loss": 1.3771,
      "theoretical_loss": 3.3887511384848072,
      "tokens_seen": 2578186240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220962888665998,
      "loss": 1.3966,
      "theoretical_loss": 3.388744367165197,
      "tokens_seen": 2578251776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022094282848545639,
      "loss": 1.3356,
      "theoretical_loss": 3.3887375960658943,
      "tokens_seen": 2578317312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022092276830491475,
      "loss": 1.3992,
      "theoretical_loss": 3.388730825186887,
      "tokens_seen": 2578382848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022090270812437314,
      "loss": 1.3421,
      "theoretical_loss": 3.388724054528163,
      "tokens_seen": 2578448384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002208826479438315,
      "loss": 1.3526,
      "theoretical_loss": 3.388717284089709,
      "tokens_seen": 2578513920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002208625877632899,
      "loss": 1.3457,
      "theoretical_loss": 3.388710513871512,
      "tokens_seen": 2578579456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022084252758274825,
      "loss": 1.2566,
      "theoretical_loss": 3.3887037438735597,
      "tokens_seen": 2578644992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022082246740220664,
      "loss": 1.2289,
      "theoretical_loss": 3.3886969740958386,
      "tokens_seen": 2578710528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220802407221665,
      "loss": 1.4146,
      "theoretical_loss": 3.388690204538337,
      "tokens_seen": 2578776064
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022078234704112337,
      "loss": 1.2102,
      "theoretical_loss": 3.388683435201041,
      "tokens_seen": 2578841600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022076228686058175,
      "loss": 1.3504,
      "theoretical_loss": 3.3886766660839385,
      "tokens_seen": 2578907136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022074222668004012,
      "loss": 1.1813,
      "theoretical_loss": 3.3886698971870173,
      "tokens_seen": 2578972672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002207221664994985,
      "loss": 1.323,
      "theoretical_loss": 3.388663128510263,
      "tokens_seen": 2579038208
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022070210631895687,
      "loss": 1.2812,
      "theoretical_loss": 3.3886563600536648,
      "tokens_seen": 2579103744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022068204613841526,
      "loss": 1.2407,
      "theoretical_loss": 3.3886495918172086,
      "tokens_seen": 2579169280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022066198595787362,
      "loss": 1.3465,
      "theoretical_loss": 3.3886428238008817,
      "tokens_seen": 2579234816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220641925777332,
      "loss": 1.2961,
      "theoretical_loss": 3.3886360560046724,
      "tokens_seen": 2579300352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022062186559679037,
      "loss": 1.2614,
      "theoretical_loss": 3.3886292884285667,
      "tokens_seen": 2579365888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022060180541624873,
      "loss": 1.3342,
      "theoretical_loss": 3.3886225210725525,
      "tokens_seen": 2579431424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022058174523570712,
      "loss": 1.3126,
      "theoretical_loss": 3.388615753936617,
      "tokens_seen": 2579496960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022056168505516549,
      "loss": 1.3641,
      "theoretical_loss": 3.3886089870207474,
      "tokens_seen": 2579562496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022054162487462387,
      "loss": 1.2825,
      "theoretical_loss": 3.3886022203249304,
      "tokens_seen": 2579628032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022052156469408224,
      "loss": 1.2858,
      "theoretical_loss": 3.3885954538491543,
      "tokens_seen": 2579693568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022050150451354063,
      "loss": 1.026,
      "theoretical_loss": 3.388588687593406,
      "tokens_seen": 2579759104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220481444332999,
      "loss": 1.475,
      "theoretical_loss": 3.3885819215576722,
      "tokens_seen": 2579824640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022046138415245738,
      "loss": 1.2341,
      "theoretical_loss": 3.388575155741941,
      "tokens_seen": 2579890176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022044132397191574,
      "loss": 1.225,
      "theoretical_loss": 3.388568390146199,
      "tokens_seen": 2579955712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002204212637913741,
      "loss": 1.2431,
      "theoretical_loss": 3.3885616247704338,
      "tokens_seen": 2580021248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002204012036108325,
      "loss": 1.1438,
      "theoretical_loss": 3.388554859614632,
      "tokens_seen": 2580086784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022038114343029085,
      "loss": 1.323,
      "theoretical_loss": 3.3885480946787823,
      "tokens_seen": 2580152320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022036108324974924,
      "loss": 1.3085,
      "theoretical_loss": 3.3885413299628704,
      "tokens_seen": 2580217856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022034102306920763,
      "loss": 1.2261,
      "theoretical_loss": 3.3885345654668844,
      "tokens_seen": 2580283392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022032096288866602,
      "loss": 1.2305,
      "theoretical_loss": 3.3885278011908113,
      "tokens_seen": 2580348928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022030090270812438,
      "loss": 1.4236,
      "theoretical_loss": 3.3885210371346384,
      "tokens_seen": 2580414464
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022028084252758277,
      "loss": 1.371,
      "theoretical_loss": 3.388514273298353,
      "tokens_seen": 2580480000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022026078234704114,
      "loss": 1.238,
      "theoretical_loss": 3.3885075096819426,
      "tokens_seen": 2580545536
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 938550,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.151209831237793,
      "objective/train/theoretical_loss": 3.388500746285394,
      "objective/train/tokens_used": 951070176,
      "theoretical_loss": 3.388500746285394,
      "tokens_seen": 2580611072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022024072216649953,
      "loss": 1.2161,
      "theoretical_loss": 3.388500746285394,
      "tokens_seen": 2580611072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002202206619859579,
      "loss": 1.3549,
      "theoretical_loss": 3.388493983108695,
      "tokens_seen": 2580676608
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022020060180541625,
      "loss": 1.2652,
      "theoretical_loss": 3.3884872201518323,
      "tokens_seen": 2580742144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022018054162487464,
      "loss": 1.3827,
      "theoretical_loss": 3.388480457414794,
      "tokens_seen": 2580807680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220160481444333,
      "loss": 1.2145,
      "theoretical_loss": 3.3884736948975664,
      "tokens_seen": 2580873216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002201404212637914,
      "loss": 1.1846,
      "theoretical_loss": 3.3884669326001373,
      "tokens_seen": 2580938752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022012036108324975,
      "loss": 1.4169,
      "theoretical_loss": 3.3884601705224937,
      "tokens_seen": 2581004288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022010030090270814,
      "loss": 1.3475,
      "theoretical_loss": 3.388453408664623,
      "tokens_seen": 2581069824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002200802407221665,
      "loss": 1.2245,
      "theoretical_loss": 3.3884466470265133,
      "tokens_seen": 2581135360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002200601805416249,
      "loss": 1.2955,
      "theoretical_loss": 3.3884398856081503,
      "tokens_seen": 2581200896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022004012036108326,
      "loss": 1.2953,
      "theoretical_loss": 3.3884331244095227,
      "tokens_seen": 2581266432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022002006018054162,
      "loss": 1.3739,
      "theoretical_loss": 3.388426363430617,
      "tokens_seen": 2581331968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022,
      "loss": 1.3167,
      "theoretical_loss": 3.38841960267142,
      "tokens_seen": 2581397504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021997993981945837,
      "loss": 1.0896,
      "theoretical_loss": 3.388412842131921,
      "tokens_seen": 2581463040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021995987963891676,
      "loss": 1.3279,
      "theoretical_loss": 3.388406081812105,
      "tokens_seen": 2581528576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021993981945837512,
      "loss": 1.3452,
      "theoretical_loss": 3.3883993217119603,
      "tokens_seen": 2581594112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002199197592778335,
      "loss": 1.3415,
      "theoretical_loss": 3.3883925618314743,
      "tokens_seen": 2581659648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021989969909729187,
      "loss": 1.498,
      "theoretical_loss": 3.3883858021706335,
      "tokens_seen": 2581725184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021987963891675026,
      "loss": 1.3083,
      "theoretical_loss": 3.3883790427294267,
      "tokens_seen": 2581790720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021985957873620862,
      "loss": 1.3676,
      "theoretical_loss": 3.38837228350784,
      "tokens_seen": 2581856256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021983951855566701,
      "loss": 1.3168,
      "theoretical_loss": 3.3883655245058604,
      "tokens_seen": 2581921792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021981945837512538,
      "loss": 1.4092,
      "theoretical_loss": 3.388358765723476,
      "tokens_seen": 2581987328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021979939819458374,
      "loss": 1.2599,
      "theoretical_loss": 3.388352007160674,
      "tokens_seen": 2582052864
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021977933801404213,
      "loss": 1.3216,
      "theoretical_loss": 3.3883452488174415,
      "tokens_seen": 2582118400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002197592778335005,
      "loss": 1.2602,
      "theoretical_loss": 3.3883384906937657,
      "tokens_seen": 2582183936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021973921765295888,
      "loss": 1.3284,
      "theoretical_loss": 3.388331732789634,
      "tokens_seen": 2582249472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021971915747241724,
      "loss": 1.3153,
      "theoretical_loss": 3.388324975105034,
      "tokens_seen": 2582315008
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021969909729187563,
      "loss": 1.2298,
      "theoretical_loss": 3.3883182176399522,
      "tokens_seen": 2582380544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000219679037111334,
      "loss": 1.2799,
      "theoretical_loss": 3.3883114603943767,
      "tokens_seen": 2582446080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021965897693079238,
      "loss": 1.4192,
      "theoretical_loss": 3.3883047033682945,
      "tokens_seen": 2582511616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021963891675025075,
      "loss": 1.2732,
      "theoretical_loss": 3.388297946561693,
      "tokens_seen": 2582577152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002196188565697091,
      "loss": 1.3011,
      "theoretical_loss": 3.3882911899745594,
      "tokens_seen": 2582642688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002195987963891675,
      "loss": 1.2516,
      "theoretical_loss": 3.388284433606881,
      "tokens_seen": 2582708224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021957873620862589,
      "loss": 1.3317,
      "theoretical_loss": 3.388277677458645,
      "tokens_seen": 2582773760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021955867602808428,
      "loss": 1.2456,
      "theoretical_loss": 3.388270921529839,
      "tokens_seen": 2582839296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021953861584754264,
      "loss": 1.278,
      "theoretical_loss": 3.38826416582045,
      "tokens_seen": 2582904832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021951855566700103,
      "loss": 1.1486,
      "theoretical_loss": 3.3882574103304655,
      "tokens_seen": 2582970368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002194984954864594,
      "loss": 1.2977,
      "theoretical_loss": 3.3882506550598728,
      "tokens_seen": 2583035904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021947843530591778,
      "loss": 1.1829,
      "theoretical_loss": 3.388243900008659,
      "tokens_seen": 2583101440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021945837512537614,
      "loss": 1.3225,
      "theoretical_loss": 3.3882371451768116,
      "tokens_seen": 2583166976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002194383149448345,
      "loss": 1.2473,
      "theoretical_loss": 3.388230390564318,
      "tokens_seen": 2583232512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002194182547642929,
      "loss": 1.2673,
      "theoretical_loss": 3.388223636171165,
      "tokens_seen": 2583298048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021939819458375125,
      "loss": 1.2482,
      "theoretical_loss": 3.388216881997341,
      "tokens_seen": 2583363584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021937813440320964,
      "loss": 1.2795,
      "theoretical_loss": 3.388210128042832,
      "tokens_seen": 2583429120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000219358074222668,
      "loss": 1.2556,
      "theoretical_loss": 3.388203374307626,
      "tokens_seen": 2583494656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002193380140421264,
      "loss": 1.2754,
      "theoretical_loss": 3.3881966207917102,
      "tokens_seen": 2583560192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021931795386158476,
      "loss": 1.2542,
      "theoretical_loss": 3.3881898674950723,
      "tokens_seen": 2583625728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021929789368104315,
      "loss": 1.2379,
      "theoretical_loss": 3.388183114417699,
      "tokens_seen": 2583691264
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002192778335005015,
      "loss": 1.335,
      "theoretical_loss": 3.3881763615595784,
      "tokens_seen": 2583756800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002192577733199599,
      "loss": 1.2097,
      "theoretical_loss": 3.3881696089206965,
      "tokens_seen": 2583822336
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 939673,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.261291742324829,
      "objective/train/theoretical_loss": 3.388162856501042,
      "objective/train/tokens_used": 954346976,
      "theoretical_loss": 3.388162856501042,
      "tokens_seen": 2583887872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021923771313941826,
      "loss": 1.2778,
      "theoretical_loss": 3.388162856501042,
      "tokens_seen": 2583887872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021921765295887662,
      "loss": 1.4465,
      "theoretical_loss": 3.3881561043006014,
      "tokens_seen": 2583953408
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000219197592778335,
      "loss": 1.2542,
      "theoretical_loss": 3.3881493523193624,
      "tokens_seen": 2584018944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021917753259779337,
      "loss": 1.3014,
      "theoretical_loss": 3.3881426005573125,
      "tokens_seen": 2584084480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021915747241725176,
      "loss": 1.2977,
      "theoretical_loss": 3.3881358490144384,
      "tokens_seen": 2584150016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021913741223671013,
      "loss": 1.3086,
      "theoretical_loss": 3.3881290976907277,
      "tokens_seen": 2584215552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021911735205616852,
      "loss": 1.2667,
      "theoretical_loss": 3.3881223465861683,
      "tokens_seen": 2584281088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021909729187562688,
      "loss": 1.3341,
      "theoretical_loss": 3.3881155957007465,
      "tokens_seen": 2584346624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021907723169508527,
      "loss": 1.2774,
      "theoretical_loss": 3.3881088450344503,
      "tokens_seen": 2584412160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021905717151454363,
      "loss": 1.3352,
      "theoretical_loss": 3.388102094587267,
      "tokens_seen": 2584477696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000219037111334002,
      "loss": 1.2903,
      "theoretical_loss": 3.3880953443591837,
      "tokens_seen": 2584543232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021901705115346038,
      "loss": 1.2157,
      "theoretical_loss": 3.3880885943501875,
      "tokens_seen": 2584608768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021899699097291874,
      "loss": 1.2359,
      "theoretical_loss": 3.3880818445602667,
      "tokens_seen": 2584674304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021897693079237713,
      "loss": 1.3053,
      "theoretical_loss": 3.388075094989408,
      "tokens_seen": 2584739840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002189568706118355,
      "loss": 1.142,
      "theoretical_loss": 3.388068345637598,
      "tokens_seen": 2584805376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021893681043129388,
      "loss": 1.3428,
      "theoretical_loss": 3.3880615965048255,
      "tokens_seen": 2584870912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021891675025075225,
      "loss": 1.3832,
      "theoretical_loss": 3.388054847591077,
      "tokens_seen": 2584936448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021889669007021064,
      "loss": 1.2131,
      "theoretical_loss": 3.38804809889634,
      "tokens_seen": 2585001984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000218876629889669,
      "loss": 1.2205,
      "theoretical_loss": 3.3880413504206017,
      "tokens_seen": 2585067520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021885656970912736,
      "loss": 1.2751,
      "theoretical_loss": 3.3880346021638497,
      "tokens_seen": 2585133056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021883650952858578,
      "loss": 1.2977,
      "theoretical_loss": 3.388027854126071,
      "tokens_seen": 2585198592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021881644934804414,
      "loss": 1.3599,
      "theoretical_loss": 3.388021106307253,
      "tokens_seen": 2585264128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021879638916750253,
      "loss": 1.254,
      "theoretical_loss": 3.3880143587073834,
      "tokens_seen": 2585329664
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002187763289869609,
      "loss": 1.423,
      "theoretical_loss": 3.3880076113264495,
      "tokens_seen": 2585395200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021875626880641928,
      "loss": 1.3074,
      "theoretical_loss": 3.3880008641644386,
      "tokens_seen": 2585460736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021873620862587764,
      "loss": 1.4021,
      "theoretical_loss": 3.3879941172213375,
      "tokens_seen": 2585526272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021871614844533603,
      "loss": 1.2713,
      "theoretical_loss": 3.3879873704971337,
      "tokens_seen": 2585591808
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002186960882647944,
      "loss": 1.2147,
      "theoretical_loss": 3.3879806239918153,
      "tokens_seen": 2585657344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021867602808425278,
      "loss": 1.3216,
      "theoretical_loss": 3.3879738777053694,
      "tokens_seen": 2585722880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021865596790371115,
      "loss": 1.3081,
      "theoretical_loss": 3.3879671316377826,
      "tokens_seen": 2585788416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002186359077231695,
      "loss": 1.1993,
      "theoretical_loss": 3.387960385789043,
      "tokens_seen": 2585853952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002186158475426279,
      "loss": 1.2671,
      "theoretical_loss": 3.3879536401591377,
      "tokens_seen": 2585919488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021859578736208626,
      "loss": 1.224,
      "theoretical_loss": 3.387946894748054,
      "tokens_seen": 2585985024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021857572718154465,
      "loss": 1.3833,
      "theoretical_loss": 3.3879401495557797,
      "tokens_seen": 2586050560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000218555667001003,
      "loss": 1.2145,
      "theoretical_loss": 3.3879334045823017,
      "tokens_seen": 2586116096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002185356068204614,
      "loss": 1.3177,
      "theoretical_loss": 3.387926659827607,
      "tokens_seen": 2586181632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021851554663991976,
      "loss": 1.2118,
      "theoretical_loss": 3.387919915291684,
      "tokens_seen": 2586247168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021849548645937815,
      "loss": 1.3505,
      "theoretical_loss": 3.387913170974519,
      "tokens_seen": 2586312704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021847542627883651,
      "loss": 1.2221,
      "theoretical_loss": 3.3879064268761003,
      "tokens_seen": 2586378240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021845536609829488,
      "loss": 1.2083,
      "theoretical_loss": 3.3878996829964145,
      "tokens_seen": 2586443776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021843530591775327,
      "loss": 1.2276,
      "theoretical_loss": 3.3878929393354493,
      "tokens_seen": 2586509312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021841524573721163,
      "loss": 1.3386,
      "theoretical_loss": 3.3878861958931923,
      "tokens_seen": 2586574848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021839518555667002,
      "loss": 1.2346,
      "theoretical_loss": 3.3878794526696305,
      "tokens_seen": 2586640384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021837512537612838,
      "loss": 1.401,
      "theoretical_loss": 3.387872709664751,
      "tokens_seen": 2586705920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021835506519558677,
      "loss": 1.4459,
      "theoretical_loss": 3.387865966878542,
      "tokens_seen": 2586771456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021833500501504513,
      "loss": 1.3222,
      "theoretical_loss": 3.38785922431099,
      "tokens_seen": 2586836992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021831494483450352,
      "loss": 1.2416,
      "theoretical_loss": 3.387852481962083,
      "tokens_seen": 2586902528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021829488465396188,
      "loss": 1.3538,
      "theoretical_loss": 3.3878457398318083,
      "tokens_seen": 2586968064
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021827482447342024,
      "loss": 1.3868,
      "theoretical_loss": 3.387838997920153,
      "tokens_seen": 2587033600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021825476429287863,
      "loss": 1.2339,
      "theoretical_loss": 3.3878322562271044,
      "tokens_seen": 2587099136
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 940829,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2509891986846924,
      "objective/train/theoretical_loss": 3.38782551475265,
      "objective/train/tokens_used": 957623776,
      "theoretical_loss": 3.38782551475265,
      "tokens_seen": 2587164672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000218234704112337,
      "loss": 1.3083,
      "theoretical_loss": 3.38782551475265,
      "tokens_seen": 2587164672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021821464393179539,
      "loss": 1.3246,
      "theoretical_loss": 3.3878187734967775,
      "tokens_seen": 2587230208
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021819458375125375,
      "loss": 1.3046,
      "theoretical_loss": 3.387812032459474,
      "tokens_seen": 2587295744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021817452357071214,
      "loss": 1.2092,
      "theoretical_loss": 3.387805291640727,
      "tokens_seen": 2587361280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002181544633901705,
      "loss": 1.3138,
      "theoretical_loss": 3.3877985510405235,
      "tokens_seen": 2587426816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002181344032096289,
      "loss": 1.2732,
      "theoretical_loss": 3.387791810658851,
      "tokens_seen": 2587492352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021811434302908725,
      "loss": 1.239,
      "theoretical_loss": 3.3877850704956978,
      "tokens_seen": 2587557888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021809428284854564,
      "loss": 1.3111,
      "theoretical_loss": 3.3877783305510496,
      "tokens_seen": 2587623424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021807422266800403,
      "loss": 1.2669,
      "theoretical_loss": 3.3877715908248955,
      "tokens_seen": 2587688960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002180541624874624,
      "loss": 1.234,
      "theoretical_loss": 3.387764851317222,
      "tokens_seen": 2587754496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021803410230692078,
      "loss": 1.3953,
      "theoretical_loss": 3.387758112028016,
      "tokens_seen": 2587820032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021801404212637914,
      "loss": 1.2365,
      "theoretical_loss": 3.3877513729572657,
      "tokens_seen": 2587885568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021799398194583753,
      "loss": 1.2418,
      "theoretical_loss": 3.3877446341049584,
      "tokens_seen": 2587951104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002179739217652959,
      "loss": 1.3406,
      "theoretical_loss": 3.387737895471081,
      "tokens_seen": 2588016640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021795386158475428,
      "loss": 1.2822,
      "theoretical_loss": 3.3877311570556214,
      "tokens_seen": 2588082176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021793380140421265,
      "loss": 1.2864,
      "theoretical_loss": 3.387724418858567,
      "tokens_seen": 2588147712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021791374122367104,
      "loss": 1.2476,
      "theoretical_loss": 3.3877176808799043,
      "tokens_seen": 2588213248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002178936810431294,
      "loss": 1.2869,
      "theoretical_loss": 3.3877109431196217,
      "tokens_seen": 2588278784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021787362086258776,
      "loss": 1.3003,
      "theoretical_loss": 3.3877042055777067,
      "tokens_seen": 2588344320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021785356068204615,
      "loss": 1.2826,
      "theoretical_loss": 3.387697468254146,
      "tokens_seen": 2588409856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002178335005015045,
      "loss": 1.3873,
      "theoretical_loss": 3.3876907311489273,
      "tokens_seen": 2588475392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002178134403209629,
      "loss": 1.2663,
      "theoretical_loss": 3.387683994262038,
      "tokens_seen": 2588540928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021779338014042126,
      "loss": 1.3139,
      "theoretical_loss": 3.3876772575934657,
      "tokens_seen": 2588606464
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021777331995987965,
      "loss": 1.236,
      "theoretical_loss": 3.387670521143197,
      "tokens_seen": 2588672000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021775325977933802,
      "loss": 1.1984,
      "theoretical_loss": 3.3876637849112203,
      "tokens_seen": 2588737536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002177331995987964,
      "loss": 1.3467,
      "theoretical_loss": 3.3876570488975224,
      "tokens_seen": 2588803072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021771313941825477,
      "loss": 1.4377,
      "theoretical_loss": 3.3876503131020907,
      "tokens_seen": 2588868608
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021769307923771316,
      "loss": 1.2978,
      "theoretical_loss": 3.387643577524913,
      "tokens_seen": 2588934144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021767301905717152,
      "loss": 1.3231,
      "theoretical_loss": 3.3876368421659766,
      "tokens_seen": 2588999680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021765295887662988,
      "loss": 1.3015,
      "theoretical_loss": 3.3876301070252683,
      "tokens_seen": 2589065216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021763289869608827,
      "loss": 1.3072,
      "theoretical_loss": 3.3876233721027766,
      "tokens_seen": 2589130752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021761283851554663,
      "loss": 1.31,
      "theoretical_loss": 3.3876166373984877,
      "tokens_seen": 2589196288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021759277833500502,
      "loss": 1.2507,
      "theoretical_loss": 3.3876099029123896,
      "tokens_seen": 2589261824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021757271815446338,
      "loss": 1.4041,
      "theoretical_loss": 3.3876031686444703,
      "tokens_seen": 2589327360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021755265797392177,
      "loss": 1.2979,
      "theoretical_loss": 3.387596434594716,
      "tokens_seen": 2589392896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021753259779338014,
      "loss": 1.273,
      "theoretical_loss": 3.3875897007631153,
      "tokens_seen": 2589458432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021751253761283852,
      "loss": 1.3271,
      "theoretical_loss": 3.3875829671496547,
      "tokens_seen": 2589523968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174924774322969,
      "loss": 1.4244,
      "theoretical_loss": 3.387576233754322,
      "tokens_seen": 2589589504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021747241725175525,
      "loss": 1.2764,
      "theoretical_loss": 3.3875695005771043,
      "tokens_seen": 2589655040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021745235707121364,
      "loss": 1.3825,
      "theoretical_loss": 3.3875627676179896,
      "tokens_seen": 2589720576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000217432296890672,
      "loss": 1.3443,
      "theoretical_loss": 3.387556034876965,
      "tokens_seen": 2589786112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174122367101304,
      "loss": 1.3369,
      "theoretical_loss": 3.3875493023540173,
      "tokens_seen": 2589851648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021739217652958875,
      "loss": 1.2365,
      "theoretical_loss": 3.3875425700491353,
      "tokens_seen": 2589917184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021737211634904714,
      "loss": 1.2822,
      "theoretical_loss": 3.3875358379623055,
      "tokens_seen": 2589982720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002173520561685055,
      "loss": 1.3167,
      "theoretical_loss": 3.387529106093515,
      "tokens_seen": 2590048256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021733199598796392,
      "loss": 1.3383,
      "theoretical_loss": 3.387522374442752,
      "tokens_seen": 2590113792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021731193580742228,
      "loss": 1.3321,
      "theoretical_loss": 3.387515643010004,
      "tokens_seen": 2590179328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021729187562688065,
      "loss": 1.2522,
      "theoretical_loss": 3.3875089117952575,
      "tokens_seen": 2590244864
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021727181544633903,
      "loss": 1.3387,
      "theoretical_loss": 3.3875021807985006,
      "tokens_seen": 2590310400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002172517552657974,
      "loss": 1.2404,
      "theoretical_loss": 3.3874954500197205,
      "tokens_seen": 2590375936
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 942074,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3940274715423584,
      "objective/train/theoretical_loss": 3.387488719458905,
      "objective/train/tokens_used": 960900576,
      "theoretical_loss": 3.387488719458905,
      "tokens_seen": 2590441472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021723169508525579,
      "loss": 1.3043,
      "theoretical_loss": 3.387488719458905,
      "tokens_seen": 2590441472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021721163490471415,
      "loss": 1.1847,
      "theoretical_loss": 3.387481989116041,
      "tokens_seen": 2590507008
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021719157472417254,
      "loss": 1.3461,
      "theoretical_loss": 3.387475258991116,
      "tokens_seen": 2590572544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002171715145436309,
      "loss": 1.2616,
      "theoretical_loss": 3.387468529084118,
      "tokens_seen": 2590638080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002171514543630893,
      "loss": 1.2405,
      "theoretical_loss": 3.387461799395034,
      "tokens_seen": 2590703616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021713139418254765,
      "loss": 1.2641,
      "theoretical_loss": 3.3874550699238513,
      "tokens_seen": 2590769152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021711133400200604,
      "loss": 1.2447,
      "theoretical_loss": 3.3874483406705576,
      "tokens_seen": 2590834688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002170912738214644,
      "loss": 1.27,
      "theoretical_loss": 3.38744161163514,
      "tokens_seen": 2590900224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021707121364092277,
      "loss": 1.3115,
      "theoretical_loss": 3.387434882817586,
      "tokens_seen": 2590965760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021705115346038115,
      "loss": 1.1958,
      "theoretical_loss": 3.3874281542178837,
      "tokens_seen": 2591031296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021703109327983952,
      "loss": 1.3434,
      "theoretical_loss": 3.3874214258360196,
      "tokens_seen": 2591096832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002170110330992979,
      "loss": 1.2896,
      "theoretical_loss": 3.387414697671982,
      "tokens_seen": 2591162368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021699097291875627,
      "loss": 1.2482,
      "theoretical_loss": 3.3874079697257575,
      "tokens_seen": 2591227904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021697091273821466,
      "loss": 1.2597,
      "theoretical_loss": 3.3874012419973343,
      "tokens_seen": 2591293440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021695085255767302,
      "loss": 1.2256,
      "theoretical_loss": 3.387394514486699,
      "tokens_seen": 2591358976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002169307923771314,
      "loss": 1.4062,
      "theoretical_loss": 3.3873877871938403,
      "tokens_seen": 2591424512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021691073219658977,
      "loss": 1.2218,
      "theoretical_loss": 3.387381060118744,
      "tokens_seen": 2591490048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021689067201604813,
      "loss": 1.3944,
      "theoretical_loss": 3.3873743332613993,
      "tokens_seen": 2591555584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021687061183550652,
      "loss": 1.3162,
      "theoretical_loss": 3.387367606621792,
      "tokens_seen": 2591621120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021685055165496489,
      "loss": 1.2974,
      "theoretical_loss": 3.387360880199911,
      "tokens_seen": 2591686656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021683049147442327,
      "loss": 1.2311,
      "theoretical_loss": 3.3873541539957426,
      "tokens_seen": 2591752192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021681043129388164,
      "loss": 1.2581,
      "theoretical_loss": 3.387347428009275,
      "tokens_seen": 2591817728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021679037111334003,
      "loss": 1.2783,
      "theoretical_loss": 3.387340702240495,
      "tokens_seen": 2591883264
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002167703109327984,
      "loss": 1.3354,
      "theoretical_loss": 3.3873339766893906,
      "tokens_seen": 2591948800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021675025075225678,
      "loss": 1.3031,
      "theoretical_loss": 3.387327251355949,
      "tokens_seen": 2592014336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021673019057171514,
      "loss": 1.2351,
      "theoretical_loss": 3.387320526240158,
      "tokens_seen": 2592079872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002167101303911735,
      "loss": 1.3426,
      "theoretical_loss": 3.387313801342004,
      "tokens_seen": 2592145408
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002166900702106319,
      "loss": 1.3081,
      "theoretical_loss": 3.387307076661476,
      "tokens_seen": 2592210944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021667001003009025,
      "loss": 1.3367,
      "theoretical_loss": 3.3873003521985603,
      "tokens_seen": 2592276480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021664994984954864,
      "loss": 1.3211,
      "theoretical_loss": 3.387293627953245,
      "tokens_seen": 2592342016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000216629889669007,
      "loss": 1.282,
      "theoretical_loss": 3.387286903925517,
      "tokens_seen": 2592407552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002166098294884654,
      "loss": 1.2634,
      "theoretical_loss": 3.3872801801153645,
      "tokens_seen": 2592473088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021658976930792378,
      "loss": 1.1624,
      "theoretical_loss": 3.387273456522774,
      "tokens_seen": 2592538624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021656970912738217,
      "loss": 1.2401,
      "theoretical_loss": 3.3872667331477335,
      "tokens_seen": 2592604160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021654964894684054,
      "loss": 1.3269,
      "theoretical_loss": 3.387260009990231,
      "tokens_seen": 2592669696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021652958876629893,
      "loss": 1.3005,
      "theoretical_loss": 3.3872532870502527,
      "tokens_seen": 2592735232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002165095285857573,
      "loss": 1.2359,
      "theoretical_loss": 3.387246564327787,
      "tokens_seen": 2592800768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021648946840521565,
      "loss": 1.3499,
      "theoretical_loss": 3.3872398418228213,
      "tokens_seen": 2592866304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021646940822467404,
      "loss": 1.219,
      "theoretical_loss": 3.3872331195353427,
      "tokens_seen": 2592931840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002164493480441324,
      "loss": 1.3621,
      "theoretical_loss": 3.387226397465339,
      "tokens_seen": 2592997376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002164292878635908,
      "loss": 1.3074,
      "theoretical_loss": 3.3872196756127972,
      "tokens_seen": 2593062912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021640922768304915,
      "loss": 1.3631,
      "theoretical_loss": 3.3872129539777056,
      "tokens_seen": 2593128448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021638916750250754,
      "loss": 1.2258,
      "theoretical_loss": 3.387206232560051,
      "tokens_seen": 2593193984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002163691073219659,
      "loss": 1.2506,
      "theoretical_loss": 3.387199511359821,
      "tokens_seen": 2593259520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002163490471414243,
      "loss": 1.309,
      "theoretical_loss": 3.3871927903770027,
      "tokens_seen": 2593325056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021632898696088266,
      "loss": 1.2013,
      "theoretical_loss": 3.3871860696115847,
      "tokens_seen": 2593390592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021630892678034102,
      "loss": 1.2941,
      "theoretical_loss": 3.3871793490635533,
      "tokens_seen": 2593456128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002162888665997994,
      "loss": 1.2226,
      "theoretical_loss": 3.3871726287328965,
      "tokens_seen": 2593521664
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021626880641925777,
      "loss": 1.3222,
      "theoretical_loss": 3.3871659086196018,
      "tokens_seen": 2593587200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021624874623871616,
      "loss": 1.2064,
      "theoretical_loss": 3.3871591887236567,
      "tokens_seen": 2593652736
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 943233,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2437530755996704,
      "objective/train/theoretical_loss": 3.387152469045048,
      "objective/train/tokens_used": 964177376,
      "theoretical_loss": 3.387152469045048,
      "tokens_seen": 2593718272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021622868605817452,
      "loss": 1.271,
      "theoretical_loss": 3.387152469045048,
      "tokens_seen": 2593718272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002162086258776329,
      "loss": 1.2822,
      "theoretical_loss": 3.3871457495837642,
      "tokens_seen": 2593783808
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021618856569709127,
      "loss": 1.3069,
      "theoretical_loss": 3.3871390303397924,
      "tokens_seen": 2593849344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021616850551654966,
      "loss": 1.211,
      "theoretical_loss": 3.38713231131312,
      "tokens_seen": 2593914880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021614844533600802,
      "loss": 1.3634,
      "theoretical_loss": 3.3871255925037342,
      "tokens_seen": 2593980416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021612838515546641,
      "loss": 1.349,
      "theoretical_loss": 3.387118873911623,
      "tokens_seen": 2594045952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021610832497492478,
      "loss": 1.3906,
      "theoretical_loss": 3.3871121555367734,
      "tokens_seen": 2594111488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021608826479438314,
      "loss": 1.2689,
      "theoretical_loss": 3.3871054373791734,
      "tokens_seen": 2594177024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021606820461384153,
      "loss": 1.2012,
      "theoretical_loss": 3.3870987194388102,
      "tokens_seen": 2594242560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002160481444332999,
      "loss": 1.2212,
      "theoretical_loss": 3.387092001715671,
      "tokens_seen": 2594308096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021602808425275828,
      "loss": 1.4563,
      "theoretical_loss": 3.387085284209744,
      "tokens_seen": 2594373632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021600802407221664,
      "loss": 1.3245,
      "theoretical_loss": 3.3870785669210166,
      "tokens_seen": 2594439168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021598796389167503,
      "loss": 1.3423,
      "theoretical_loss": 3.3870718498494754,
      "tokens_seen": 2594504704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002159679037111334,
      "loss": 1.3317,
      "theoretical_loss": 3.387065132995109,
      "tokens_seen": 2594570240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021594784353059178,
      "loss": 1.2311,
      "theoretical_loss": 3.3870584163579034,
      "tokens_seen": 2594635776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021592778335005014,
      "loss": 1.2042,
      "theoretical_loss": 3.3870516999378477,
      "tokens_seen": 2594701312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002159077231695085,
      "loss": 1.1755,
      "theoretical_loss": 3.387044983734929,
      "tokens_seen": 2594766848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002158876629889669,
      "loss": 1.204,
      "theoretical_loss": 3.3870382677491344,
      "tokens_seen": 2594832384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021586760280842526,
      "loss": 1.2663,
      "theoretical_loss": 3.3870315519804515,
      "tokens_seen": 2594897920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021584754262788365,
      "loss": 1.2467,
      "theoretical_loss": 3.3870248364288678,
      "tokens_seen": 2594963456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021582748244734204,
      "loss": 1.2131,
      "theoretical_loss": 3.387018121094371,
      "tokens_seen": 2595028992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021580742226680043,
      "loss": 1.4148,
      "theoretical_loss": 3.3870114059769483,
      "tokens_seen": 2595094528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157873620862588,
      "loss": 1.2572,
      "theoretical_loss": 3.387004691076587,
      "tokens_seen": 2595160064
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021576730190571718,
      "loss": 1.2429,
      "theoretical_loss": 3.3869979763932756,
      "tokens_seen": 2595225600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021574724172517554,
      "loss": 1.3457,
      "theoretical_loss": 3.3869912619270006,
      "tokens_seen": 2595291136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157271815446339,
      "loss": 1.4355,
      "theoretical_loss": 3.3869845476777503,
      "tokens_seen": 2595356672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157071213640923,
      "loss": 1.296,
      "theoretical_loss": 3.386977833645511,
      "tokens_seen": 2595422208
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021568706118355065,
      "loss": 1.2851,
      "theoretical_loss": 3.386971119830272,
      "tokens_seen": 2595487744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021566700100300904,
      "loss": 1.3013,
      "theoretical_loss": 3.386964406232019,
      "tokens_seen": 2595553280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002156469408224674,
      "loss": 1.1833,
      "theoretical_loss": 3.38695769285074,
      "tokens_seen": 2595618816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002156268806419258,
      "loss": 1.2655,
      "theoretical_loss": 3.3869509796864237,
      "tokens_seen": 2595684352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021560682046138416,
      "loss": 1.1923,
      "theoretical_loss": 3.3869442667390564,
      "tokens_seen": 2595749888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021558676028084255,
      "loss": 1.3299,
      "theoretical_loss": 3.3869375540086257,
      "tokens_seen": 2595815424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002155667001003009,
      "loss": 1.3791,
      "theoretical_loss": 3.3869308414951194,
      "tokens_seen": 2595880960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002155466399197593,
      "loss": 1.2975,
      "theoretical_loss": 3.3869241291985253,
      "tokens_seen": 2595946496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021552657973921766,
      "loss": 1.4468,
      "theoretical_loss": 3.38691741711883,
      "tokens_seen": 2596012032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021550651955867602,
      "loss": 1.3449,
      "theoretical_loss": 3.386910705256022,
      "tokens_seen": 2596077568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002154864593781344,
      "loss": 1.2333,
      "theoretical_loss": 3.386903993610088,
      "tokens_seen": 2596143104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021546639919759277,
      "loss": 1.2764,
      "theoretical_loss": 3.3868972821810166,
      "tokens_seen": 2596208640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021544633901705116,
      "loss": 1.3744,
      "theoretical_loss": 3.386890570968794,
      "tokens_seen": 2596274176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021542627883650953,
      "loss": 1.2479,
      "theoretical_loss": 3.3868838599734086,
      "tokens_seen": 2596339712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021540621865596792,
      "loss": 1.1589,
      "theoretical_loss": 3.3868771491948477,
      "tokens_seen": 2596405248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021538615847542628,
      "loss": 1.2463,
      "theoretical_loss": 3.386870438633099,
      "tokens_seen": 2596470784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021536609829488467,
      "loss": 1.2987,
      "theoretical_loss": 3.3868637282881493,
      "tokens_seen": 2596536320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021534603811434303,
      "loss": 1.2384,
      "theoretical_loss": 3.3868570181599873,
      "tokens_seen": 2596601856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002153259779338014,
      "loss": 1.4199,
      "theoretical_loss": 3.3868503082485995,
      "tokens_seen": 2596667392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021530591775325978,
      "loss": 1.3321,
      "theoretical_loss": 3.3868435985539738,
      "tokens_seen": 2596732928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021528585757271814,
      "loss": 1.3804,
      "theoretical_loss": 3.386836889076098,
      "tokens_seen": 2596798464
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021526579739217653,
      "loss": 1.3499,
      "theoretical_loss": 3.3868301798149587,
      "tokens_seen": 2596864000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002152457372116349,
      "loss": 1.3372,
      "theoretical_loss": 3.3868234707705445,
      "tokens_seen": 2596929536
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 944531,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2842936515808105,
      "objective/train/theoretical_loss": 3.3868167619428426,
      "objective/train/tokens_used": 967454176,
      "theoretical_loss": 3.3868167619428426,
      "tokens_seen": 2596995072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021522567703109328,
      "loss": 1.3485,
      "theoretical_loss": 3.3868167619428426,
      "tokens_seen": 2596995072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021520561685055165,
      "loss": 1.2758,
      "theoretical_loss": 3.38681005333184,
      "tokens_seen": 2597060608
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021518555667001004,
      "loss": 1.2756,
      "theoretical_loss": 3.386803344937525,
      "tokens_seen": 2597126144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002151654964894684,
      "loss": 1.3194,
      "theoretical_loss": 3.386796636759885,
      "tokens_seen": 2597191680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021514543630892676,
      "loss": 1.3259,
      "theoretical_loss": 3.3867899287989074,
      "tokens_seen": 2597257216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021512537612838515,
      "loss": 1.2657,
      "theoretical_loss": 3.386783221054579,
      "tokens_seen": 2597322752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002151053159478435,
      "loss": 1.2777,
      "theoretical_loss": 3.386776513526889,
      "tokens_seen": 2597388288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021508525576730193,
      "loss": 1.2804,
      "theoretical_loss": 3.386769806215823,
      "tokens_seen": 2597453824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002150651955867603,
      "loss": 1.5169,
      "theoretical_loss": 3.38676309912137,
      "tokens_seen": 2597519360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021504513540621868,
      "loss": 1.3961,
      "theoretical_loss": 3.386756392243517,
      "tokens_seen": 2597584896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021502507522567704,
      "loss": 1.3238,
      "theoretical_loss": 3.3867496855822514,
      "tokens_seen": 2597650432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021500501504513543,
      "loss": 1.4555,
      "theoretical_loss": 3.386742979137561,
      "tokens_seen": 2597715968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002149849548645938,
      "loss": 1.2589,
      "theoretical_loss": 3.3867362729094332,
      "tokens_seen": 2597781504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021496489468405218,
      "loss": 1.3459,
      "theoretical_loss": 3.3867295668978556,
      "tokens_seen": 2597847040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021494483450351055,
      "loss": 1.2854,
      "theoretical_loss": 3.386722861102816,
      "tokens_seen": 2597912576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002149247743229689,
      "loss": 1.3886,
      "theoretical_loss": 3.3867161555243017,
      "tokens_seen": 2597978112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002149047141424273,
      "loss": 1.3515,
      "theoretical_loss": 3.3867094501623,
      "tokens_seen": 2598043648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021488465396188566,
      "loss": 1.3191,
      "theoretical_loss": 3.3867027450167986,
      "tokens_seen": 2598109184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021486459378134405,
      "loss": 1.3681,
      "theoretical_loss": 3.3866960400877852,
      "tokens_seen": 2598174720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002148445336008024,
      "loss": 1.329,
      "theoretical_loss": 3.3866893353752476,
      "tokens_seen": 2598240256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002148244734202608,
      "loss": 1.4251,
      "theoretical_loss": 3.3866826308791724,
      "tokens_seen": 2598305792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021480441323971916,
      "loss": 1.2626,
      "theoretical_loss": 3.3866759265995485,
      "tokens_seen": 2598371328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021478435305917755,
      "loss": 1.2658,
      "theoretical_loss": 3.386669222536362,
      "tokens_seen": 2598436864
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021476429287863591,
      "loss": 1.4198,
      "theoretical_loss": 3.3866625186896018,
      "tokens_seen": 2598502400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021474423269809428,
      "loss": 1.3032,
      "theoretical_loss": 3.386655815059255,
      "tokens_seen": 2598567936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021472417251755267,
      "loss": 1.326,
      "theoretical_loss": 3.3866491116453084,
      "tokens_seen": 2598633472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021470411233701103,
      "loss": 1.2853,
      "theoretical_loss": 3.3866424084477504,
      "tokens_seen": 2598699008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021468405215646942,
      "loss": 1.2904,
      "theoretical_loss": 3.386635705466569,
      "tokens_seen": 2598764544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021466399197592778,
      "loss": 1.244,
      "theoretical_loss": 3.38662900270175,
      "tokens_seen": 2598830080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021464393179538617,
      "loss": 1.2915,
      "theoretical_loss": 3.3866223001532827,
      "tokens_seen": 2598895616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021462387161484453,
      "loss": 1.2506,
      "theoretical_loss": 3.3866155978211534,
      "tokens_seen": 2598961152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021460381143430292,
      "loss": 1.2992,
      "theoretical_loss": 3.3866088957053506,
      "tokens_seen": 2599026688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021458375125376128,
      "loss": 1.326,
      "theoretical_loss": 3.386602193805862,
      "tokens_seen": 2599092224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021456369107321964,
      "loss": 1.2076,
      "theoretical_loss": 3.386595492122674,
      "tokens_seen": 2599157760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021454363089267803,
      "loss": 1.2627,
      "theoretical_loss": 3.3865887906557752,
      "tokens_seen": 2599223296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002145235707121364,
      "loss": 1.3047,
      "theoretical_loss": 3.386582089405153,
      "tokens_seen": 2599288832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021450351053159479,
      "loss": 1.2666,
      "theoretical_loss": 3.3865753883707943,
      "tokens_seen": 2599354368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021448345035105315,
      "loss": 1.247,
      "theoretical_loss": 3.3865686875526873,
      "tokens_seen": 2599419904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021446339017051154,
      "loss": 1.3109,
      "theoretical_loss": 3.3865619869508197,
      "tokens_seen": 2599485440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002144433299899699,
      "loss": 1.2982,
      "theoretical_loss": 3.3865552865651782,
      "tokens_seen": 2599550976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002144232698094283,
      "loss": 1.225,
      "theoretical_loss": 3.386548586395752,
      "tokens_seen": 2599616512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021440320962888665,
      "loss": 1.3705,
      "theoretical_loss": 3.3865418864425267,
      "tokens_seen": 2599682048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021438314944834504,
      "loss": 1.2968,
      "theoretical_loss": 3.3865351867054914,
      "tokens_seen": 2599747584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002143630892678034,
      "loss": 1.2971,
      "theoretical_loss": 3.3865284871846324,
      "tokens_seen": 2599813120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021434302908726176,
      "loss": 1.2057,
      "theoretical_loss": 3.3865217878799387,
      "tokens_seen": 2599878656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021432296890672018,
      "loss": 1.3999,
      "theoretical_loss": 3.3865150887913966,
      "tokens_seen": 2599944192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021430290872617854,
      "loss": 1.2995,
      "theoretical_loss": 3.386508389918995,
      "tokens_seen": 2600009728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021428284854563693,
      "loss": 1.3812,
      "theoretical_loss": 3.38650169126272,
      "tokens_seen": 2600075264
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002142627883650953,
      "loss": 1.348,
      "theoretical_loss": 3.38649499282256,
      "tokens_seen": 2600140800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021424272818455368,
      "loss": 1.2649,
      "theoretical_loss": 3.3864882945985024,
      "tokens_seen": 2600206336
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 945762,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.239673137664795,
      "objective/train/theoretical_loss": 3.386481596590535,
      "objective/train/tokens_used": 970730976,
      "theoretical_loss": 3.386481596590535,
      "tokens_seen": 2600271872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021422266800401205,
      "loss": 1.2688,
      "theoretical_loss": 3.386481596590535,
      "tokens_seen": 2600271872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021420260782347044,
      "loss": 1.3786,
      "theoretical_loss": 3.386474898798645,
      "tokens_seen": 2600337408
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002141825476429288,
      "loss": 1.2528,
      "theoretical_loss": 3.3864682012228204,
      "tokens_seen": 2600402944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021416248746238716,
      "loss": 1.2738,
      "theoretical_loss": 3.386461503863049,
      "tokens_seen": 2600468480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021414242728184555,
      "loss": 1.4186,
      "theoretical_loss": 3.386454806719317,
      "tokens_seen": 2600534016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002141223671013039,
      "loss": 1.3249,
      "theoretical_loss": 3.3864481097916137,
      "tokens_seen": 2600599552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002141023069207623,
      "loss": 1.4794,
      "theoretical_loss": 3.386441413079926,
      "tokens_seen": 2600665088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021408224674022066,
      "loss": 1.3045,
      "theoretical_loss": 3.3864347165842412,
      "tokens_seen": 2600730624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021406218655967905,
      "loss": 1.2861,
      "theoretical_loss": 3.386428020304547,
      "tokens_seen": 2600796160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021404212637913742,
      "loss": 1.2858,
      "theoretical_loss": 3.386421324240831,
      "tokens_seen": 2600861696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002140220661985958,
      "loss": 1.2522,
      "theoretical_loss": 3.3864146283930814,
      "tokens_seen": 2600927232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021400200601805417,
      "loss": 1.3613,
      "theoretical_loss": 3.3864079327612853,
      "tokens_seen": 2600992768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021398194583751256,
      "loss": 1.315,
      "theoretical_loss": 3.38640123734543,
      "tokens_seen": 2601058304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021396188565697092,
      "loss": 1.3365,
      "theoretical_loss": 3.3863945421455037,
      "tokens_seen": 2601123840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021394182547642928,
      "loss": 1.2409,
      "theoretical_loss": 3.3863878471614934,
      "tokens_seen": 2601189376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021392176529588767,
      "loss": 1.4198,
      "theoretical_loss": 3.386381152393387,
      "tokens_seen": 2601254912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021390170511534603,
      "loss": 1.3346,
      "theoretical_loss": 3.386374457841172,
      "tokens_seen": 2601320448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021388164493480442,
      "loss": 1.3286,
      "theoretical_loss": 3.3863677635048366,
      "tokens_seen": 2601385984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021386158475426278,
      "loss": 1.3035,
      "theoretical_loss": 3.3863610693843675,
      "tokens_seen": 2601451520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021384152457372117,
      "loss": 1.2868,
      "theoretical_loss": 3.3863543754797525,
      "tokens_seen": 2601517056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021382146439317954,
      "loss": 1.178,
      "theoretical_loss": 3.38634768179098,
      "tokens_seen": 2601582592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021380140421263792,
      "loss": 1.4094,
      "theoretical_loss": 3.3863409883180364,
      "tokens_seen": 2601648128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137813440320963,
      "loss": 1.367,
      "theoretical_loss": 3.38633429506091,
      "tokens_seen": 2601713664
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021376128385155465,
      "loss": 1.248,
      "theoretical_loss": 3.3863276020195885,
      "tokens_seen": 2601779200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021374122367101304,
      "loss": 1.3276,
      "theoretical_loss": 3.386320909194059,
      "tokens_seen": 2601844736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137211634904714,
      "loss": 1.2629,
      "theoretical_loss": 3.38631421658431,
      "tokens_seen": 2601910272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137011033099298,
      "loss": 1.3289,
      "theoretical_loss": 3.386307524190328,
      "tokens_seen": 2601975808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021368104312938815,
      "loss": 1.2458,
      "theoretical_loss": 3.3863008320121013,
      "tokens_seen": 2602041344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021366098294884654,
      "loss": 1.3434,
      "theoretical_loss": 3.3862941400496176,
      "tokens_seen": 2602106880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002136409227683049,
      "loss": 1.281,
      "theoretical_loss": 3.386287448302864,
      "tokens_seen": 2602172416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002136208625877633,
      "loss": 1.3291,
      "theoretical_loss": 3.3862807567718285,
      "tokens_seen": 2602237952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021360080240722166,
      "loss": 1.259,
      "theoretical_loss": 3.3862740654564987,
      "tokens_seen": 2602303488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021358074222668004,
      "loss": 1.1932,
      "theoretical_loss": 3.3862673743568616,
      "tokens_seen": 2602369024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021356068204613843,
      "loss": 1.2586,
      "theoretical_loss": 3.386260683472906,
      "tokens_seen": 2602434560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002135406218655968,
      "loss": 1.3246,
      "theoretical_loss": 3.386253992804618,
      "tokens_seen": 2602500096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021352056168505519,
      "loss": 1.4207,
      "theoretical_loss": 3.3862473023519866,
      "tokens_seen": 2602565632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021350050150451355,
      "loss": 1.2571,
      "theoretical_loss": 3.3862406121149986,
      "tokens_seen": 2602631168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021348044132397194,
      "loss": 1.4282,
      "theoretical_loss": 3.386233922093642,
      "tokens_seen": 2602696704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002134603811434303,
      "loss": 1.2704,
      "theoretical_loss": 3.386227232287905,
      "tokens_seen": 2602762240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002134403209628887,
      "loss": 1.2008,
      "theoretical_loss": 3.3862205426977736,
      "tokens_seen": 2602827776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021342026078234705,
      "loss": 1.3949,
      "theoretical_loss": 3.3862138533232367,
      "tokens_seen": 2602893312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021340020060180544,
      "loss": 1.2713,
      "theoretical_loss": 3.3862071641642815,
      "tokens_seen": 2602958848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002133801404212638,
      "loss": 1.3066,
      "theoretical_loss": 3.3862004752208956,
      "tokens_seen": 2603024384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021336008024072216,
      "loss": 1.2632,
      "theoretical_loss": 3.3861937864930667,
      "tokens_seen": 2603089920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021334002006018055,
      "loss": 1.3674,
      "theoretical_loss": 3.386187097980783,
      "tokens_seen": 2603155456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021331995987963892,
      "loss": 1.4546,
      "theoretical_loss": 3.386180409684031,
      "tokens_seen": 2603220992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002132998996990973,
      "loss": 1.2932,
      "theoretical_loss": 3.386173721602799,
      "tokens_seen": 2603286528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021327983951855567,
      "loss": 1.3861,
      "theoretical_loss": 3.386167033737075,
      "tokens_seen": 2603352064
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021325977933801406,
      "loss": 1.2028,
      "theoretical_loss": 3.3861603460868457,
      "tokens_seen": 2603417600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021323971915747242,
      "loss": 1.3224,
      "theoretical_loss": 3.3861536586520993,
      "tokens_seen": 2603483136
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 946919,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1864168643951416,
      "objective/train/theoretical_loss": 3.386146971432823,
      "objective/train/tokens_used": 974007776,
      "theoretical_loss": 3.386146971432823,
      "tokens_seen": 2603548672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002132196589769308,
      "loss": 1.3414,
      "theoretical_loss": 3.386146971432823,
      "tokens_seen": 2603548672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021319959879638917,
      "loss": 1.4402,
      "theoretical_loss": 3.3861402844290054,
      "tokens_seen": 2603614208
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021317953861584753,
      "loss": 1.3805,
      "theoretical_loss": 3.386133597640633,
      "tokens_seen": 2603679744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021315947843530592,
      "loss": 1.3066,
      "theoretical_loss": 3.386126911067694,
      "tokens_seen": 2603745280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021313941825476429,
      "loss": 1.4219,
      "theoretical_loss": 3.3861202247101767,
      "tokens_seen": 2603810816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021311935807422267,
      "loss": 1.2963,
      "theoretical_loss": 3.3861135385680674,
      "tokens_seen": 2603876352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021309929789368104,
      "loss": 1.1878,
      "theoretical_loss": 3.3861068526413542,
      "tokens_seen": 2603941888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021307923771313943,
      "loss": 1.4805,
      "theoretical_loss": 3.3861001669300252,
      "tokens_seen": 2604007424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002130591775325978,
      "loss": 1.5368,
      "theoretical_loss": 3.3860934814340675,
      "tokens_seen": 2604072960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021303911735205618,
      "loss": 1.2183,
      "theoretical_loss": 3.3860867961534695,
      "tokens_seen": 2604138496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021301905717151454,
      "loss": 1.3409,
      "theoretical_loss": 3.386080111088218,
      "tokens_seen": 2604204032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002129989969909729,
      "loss": 1.3847,
      "theoretical_loss": 3.3860734262383008,
      "tokens_seen": 2604269568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002129789368104313,
      "loss": 1.2489,
      "theoretical_loss": 3.386066741603706,
      "tokens_seen": 2604335104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021295887662988965,
      "loss": 1.3172,
      "theoretical_loss": 3.3860600571844204,
      "tokens_seen": 2604400640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021293881644934804,
      "loss": 1.1577,
      "theoretical_loss": 3.3860533729804327,
      "tokens_seen": 2604466176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002129187562688064,
      "loss": 1.3699,
      "theoretical_loss": 3.38604668899173,
      "tokens_seen": 2604531712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002128986960882648,
      "loss": 1.3584,
      "theoretical_loss": 3.3860400052183,
      "tokens_seen": 2604597248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021287863590772316,
      "loss": 1.323,
      "theoretical_loss": 3.3860333216601304,
      "tokens_seen": 2604662784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021285857572718155,
      "loss": 1.3588,
      "theoretical_loss": 3.3860266383172086,
      "tokens_seen": 2604728320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002128385155466399,
      "loss": 1.3598,
      "theoretical_loss": 3.3860199551895227,
      "tokens_seen": 2604793856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021281845536609832,
      "loss": 1.216,
      "theoretical_loss": 3.3860132722770597,
      "tokens_seen": 2604859392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002127983951855567,
      "loss": 1.2507,
      "theoretical_loss": 3.386006589579808,
      "tokens_seen": 2604924928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021277833500501505,
      "loss": 1.2601,
      "theoretical_loss": 3.385999907097755,
      "tokens_seen": 2604990464
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021275827482447344,
      "loss": 1.3663,
      "theoretical_loss": 3.385993224830888,
      "tokens_seen": 2605056000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002127382146439318,
      "loss": 1.4498,
      "theoretical_loss": 3.3859865427791953,
      "tokens_seen": 2605121536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002127181544633902,
      "loss": 1.184,
      "theoretical_loss": 3.385979860942664,
      "tokens_seen": 2605187072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021269809428284855,
      "loss": 1.38,
      "theoretical_loss": 3.385973179321282,
      "tokens_seen": 2605252608
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021267803410230694,
      "loss": 1.3054,
      "theoretical_loss": 3.3859664979150366,
      "tokens_seen": 2605318144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002126579739217653,
      "loss": 1.2271,
      "theoretical_loss": 3.385959816723916,
      "tokens_seen": 2605383680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002126379137412237,
      "loss": 1.2764,
      "theoretical_loss": 3.385953135747908,
      "tokens_seen": 2605449216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021261785356068206,
      "loss": 1.3442,
      "theoretical_loss": 3.3859464549869993,
      "tokens_seen": 2605514752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021259779338014042,
      "loss": 1.3302,
      "theoretical_loss": 3.3859397744411783,
      "tokens_seen": 2605580288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002125777331995988,
      "loss": 1.3114,
      "theoretical_loss": 3.385933094110433,
      "tokens_seen": 2605645824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021255767301905717,
      "loss": 1.2694,
      "theoretical_loss": 3.38592641399475,
      "tokens_seen": 2605711360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021253761283851556,
      "loss": 1.2228,
      "theoretical_loss": 3.385919734094118,
      "tokens_seen": 2605776896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021251755265797392,
      "loss": 1.2205,
      "theoretical_loss": 3.385913054408524,
      "tokens_seen": 2605842432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002124974924774323,
      "loss": 1.234,
      "theoretical_loss": 3.3859063749379557,
      "tokens_seen": 2605907968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021247743229689067,
      "loss": 1.4278,
      "theoretical_loss": 3.3858996956824012,
      "tokens_seen": 2605973504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021245737211634906,
      "loss": 1.2267,
      "theoretical_loss": 3.385893016641848,
      "tokens_seen": 2606039040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021243731193580742,
      "loss": 1.3179,
      "theoretical_loss": 3.3858863378162836,
      "tokens_seen": 2606104576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002124172517552658,
      "loss": 1.3058,
      "theoretical_loss": 3.385879659205696,
      "tokens_seen": 2606170112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021239719157472418,
      "loss": 1.394,
      "theoretical_loss": 3.3858729808100727,
      "tokens_seen": 2606235648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021237713139418254,
      "loss": 1.361,
      "theoretical_loss": 3.385866302629401,
      "tokens_seen": 2606301184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021235707121364093,
      "loss": 1.3683,
      "theoretical_loss": 3.385859624663669,
      "tokens_seen": 2606366720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002123370110330993,
      "loss": 1.3794,
      "theoretical_loss": 3.3858529469128644,
      "tokens_seen": 2606432256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021231695085255768,
      "loss": 1.3559,
      "theoretical_loss": 3.385846269376975,
      "tokens_seen": 2606497792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021229689067201604,
      "loss": 1.3473,
      "theoretical_loss": 3.385839592055988,
      "tokens_seen": 2606563328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021227683049147443,
      "loss": 1.4184,
      "theoretical_loss": 3.385832914949891,
      "tokens_seen": 2606628864
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002122567703109328,
      "loss": 1.2923,
      "theoretical_loss": 3.3858262380586726,
      "tokens_seen": 2606694400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021223671013039118,
      "loss": 1.3233,
      "theoretical_loss": 3.38581956138232,
      "tokens_seen": 2606759936
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 948007,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.389406442642212,
      "objective/train/theoretical_loss": 3.38581288492082,
      "objective/train/tokens_used": 977284576,
      "theoretical_loss": 3.38581288492082,
      "tokens_seen": 2606825472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021221664994984954,
      "loss": 1.3685,
      "theoretical_loss": 3.38581288492082,
      "tokens_seen": 2606825472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121965897693079,
      "loss": 1.2768,
      "theoretical_loss": 3.3858062086741616,
      "tokens_seen": 2606891008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121765295887663,
      "loss": 1.2539,
      "theoretical_loss": 3.385799532642332,
      "tokens_seen": 2606956544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021215646940822466,
      "loss": 1.2671,
      "theoretical_loss": 3.3857928568253186,
      "tokens_seen": 2607022080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021213640922768305,
      "loss": 1.2328,
      "theoretical_loss": 3.38578618122311,
      "tokens_seen": 2607087616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121163490471414,
      "loss": 1.3718,
      "theoretical_loss": 3.3857795058356923,
      "tokens_seen": 2607153152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002120962888665998,
      "loss": 1.3025,
      "theoretical_loss": 3.3857728306630546,
      "tokens_seen": 2607218688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002120762286860582,
      "loss": 1.2705,
      "theoretical_loss": 3.385766155705184,
      "tokens_seen": 2607284224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021205616850551658,
      "loss": 1.4094,
      "theoretical_loss": 3.385759480962068,
      "tokens_seen": 2607349760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021203610832497494,
      "loss": 1.2648,
      "theoretical_loss": 3.385752806433695,
      "tokens_seen": 2607415296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002120160481444333,
      "loss": 1.1654,
      "theoretical_loss": 3.385746132120052,
      "tokens_seen": 2607480832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002119959879638917,
      "loss": 1.3795,
      "theoretical_loss": 3.3857394580211273,
      "tokens_seen": 2607546368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021197592778335005,
      "loss": 1.3531,
      "theoretical_loss": 3.3857327841369083,
      "tokens_seen": 2607611904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021195586760280844,
      "loss": 1.1575,
      "theoretical_loss": 3.3857261104673824,
      "tokens_seen": 2607677440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002119358074222668,
      "loss": 1.3679,
      "theoretical_loss": 3.3857194370125376,
      "tokens_seen": 2607742976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002119157472417252,
      "loss": 1.4051,
      "theoretical_loss": 3.3857127637723616,
      "tokens_seen": 2607808512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021189568706118356,
      "loss": 1.3877,
      "theoretical_loss": 3.385706090746842,
      "tokens_seen": 2607874048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021187562688064195,
      "loss": 1.2889,
      "theoretical_loss": 3.3856994179359665,
      "tokens_seen": 2607939584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002118555667001003,
      "loss": 1.3065,
      "theoretical_loss": 3.385692745339723,
      "tokens_seen": 2608005120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002118355065195587,
      "loss": 1.2586,
      "theoretical_loss": 3.3856860729580993,
      "tokens_seen": 2608070656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021181544633901706,
      "loss": 1.2676,
      "theoretical_loss": 3.3856794007910826,
      "tokens_seen": 2608136192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021179538615847542,
      "loss": 1.4326,
      "theoretical_loss": 3.385672728838661,
      "tokens_seen": 2608201728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002117753259779338,
      "loss": 1.2722,
      "theoretical_loss": 3.385666057100822,
      "tokens_seen": 2608267264
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021175526579739217,
      "loss": 1.2817,
      "theoretical_loss": 3.3856593855775534,
      "tokens_seen": 2608332800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021173520561685056,
      "loss": 1.2932,
      "theoretical_loss": 3.3856527142688426,
      "tokens_seen": 2608398336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021171514543630893,
      "loss": 1.2759,
      "theoretical_loss": 3.385646043174678,
      "tokens_seen": 2608463872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021169508525576732,
      "loss": 1.3388,
      "theoretical_loss": 3.3856393722950466,
      "tokens_seen": 2608529408
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021167502507522568,
      "loss": 1.2351,
      "theoretical_loss": 3.385632701629937,
      "tokens_seen": 2608594944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021165496489468407,
      "loss": 1.2411,
      "theoretical_loss": 3.385626031179336,
      "tokens_seen": 2608660480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021163490471414243,
      "loss": 1.277,
      "theoretical_loss": 3.3856193609432315,
      "tokens_seen": 2608726016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002116148445336008,
      "loss": 1.339,
      "theoretical_loss": 3.3856126909216115,
      "tokens_seen": 2608791552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021159478435305918,
      "loss": 1.2675,
      "theoretical_loss": 3.3856060211144636,
      "tokens_seen": 2608857088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021157472417251754,
      "loss": 1.3877,
      "theoretical_loss": 3.3855993515217757,
      "tokens_seen": 2608922624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021155466399197593,
      "loss": 1.2376,
      "theoretical_loss": 3.385592682143535,
      "tokens_seen": 2608988160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002115346038114343,
      "loss": 1.2086,
      "theoretical_loss": 3.3855860129797297,
      "tokens_seen": 2609053696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021151454363089268,
      "loss": 1.2228,
      "theoretical_loss": 3.385579344030347,
      "tokens_seen": 2609119232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021149448345035105,
      "loss": 1.3658,
      "theoretical_loss": 3.385572675295376,
      "tokens_seen": 2609184768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021147442326980944,
      "loss": 1.3255,
      "theoretical_loss": 3.385566006774803,
      "tokens_seen": 2609250304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002114543630892678,
      "loss": 1.2569,
      "theoretical_loss": 3.3855593384686156,
      "tokens_seen": 2609315840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021143430290872616,
      "loss": 1.3271,
      "theoretical_loss": 3.385552670376802,
      "tokens_seen": 2609381376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021141424272818455,
      "loss": 1.276,
      "theoretical_loss": 3.3855460024993502,
      "tokens_seen": 2609446912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002113941825476429,
      "loss": 1.3329,
      "theoretical_loss": 3.385539334836248,
      "tokens_seen": 2609512448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002113741223671013,
      "loss": 1.3814,
      "theoretical_loss": 3.3855326673874826,
      "tokens_seen": 2609577984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021135406218655966,
      "loss": 1.1943,
      "theoretical_loss": 3.385526000153042,
      "tokens_seen": 2609643520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021133400200601808,
      "loss": 1.3316,
      "theoretical_loss": 3.3855193331329136,
      "tokens_seen": 2609709056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021131394182547644,
      "loss": 1.307,
      "theoretical_loss": 3.385512666327086,
      "tokens_seen": 2609774592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021129388164493483,
      "loss": 1.3388,
      "theoretical_loss": 3.3855059997355457,
      "tokens_seen": 2609840128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002112738214643932,
      "loss": 1.3299,
      "theoretical_loss": 3.3854993333582817,
      "tokens_seen": 2609905664
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021125376128385158,
      "loss": 1.3542,
      "theoretical_loss": 3.3854926671952805,
      "tokens_seen": 2609971200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021123370110330994,
      "loss": 1.4304,
      "theoretical_loss": 3.3854860012465307,
      "tokens_seen": 2610036736
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 949200,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3941702842712402,
      "objective/train/theoretical_loss": 3.3854793355120196,
      "objective/train/tokens_used": 980561376,
      "theoretical_loss": 3.3854793355120196,
      "tokens_seen": 2610102272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002112136409227683,
      "loss": 1.3241,
      "theoretical_loss": 3.3854793355120196,
      "tokens_seen": 2610102272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002111935807422267,
      "loss": 1.2624,
      "theoretical_loss": 3.3854726699917355,
      "tokens_seen": 2610167808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021117352056168506,
      "loss": 1.3665,
      "theoretical_loss": 3.3854660046856653,
      "tokens_seen": 2610233344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021115346038114345,
      "loss": 1.1832,
      "theoretical_loss": 3.3854593395937975,
      "tokens_seen": 2610298880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002111334002006018,
      "loss": 1.3294,
      "theoretical_loss": 3.3854526747161193,
      "tokens_seen": 2610364416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002111133400200602,
      "loss": 1.2467,
      "theoretical_loss": 3.3854460100526187,
      "tokens_seen": 2610429952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021109327983951856,
      "loss": 1.3404,
      "theoretical_loss": 3.3854393456032836,
      "tokens_seen": 2610495488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021107321965897695,
      "loss": 1.255,
      "theoretical_loss": 3.3854326813681013,
      "tokens_seen": 2610561024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002110531594784353,
      "loss": 1.2404,
      "theoretical_loss": 3.38542601734706,
      "tokens_seen": 2610626560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021103309929789368,
      "loss": 1.3178,
      "theoretical_loss": 3.385419353540147,
      "tokens_seen": 2610692096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021101303911735207,
      "loss": 1.3363,
      "theoretical_loss": 3.3854126899473505,
      "tokens_seen": 2610757632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021099297893681043,
      "loss": 1.4226,
      "theoretical_loss": 3.385406026568658,
      "tokens_seen": 2610823168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021097291875626882,
      "loss": 1.1809,
      "theoretical_loss": 3.385399363404057,
      "tokens_seen": 2610888704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021095285857572718,
      "loss": 1.1952,
      "theoretical_loss": 3.3853927004535356,
      "tokens_seen": 2610954240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021093279839518557,
      "loss": 1.3801,
      "theoretical_loss": 3.3853860377170815,
      "tokens_seen": 2611019776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021091273821464393,
      "loss": 1.2187,
      "theoretical_loss": 3.385379375194683,
      "tokens_seen": 2611085312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021089267803410232,
      "loss": 1.2584,
      "theoretical_loss": 3.3853727128863262,
      "tokens_seen": 2611150848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021087261785356068,
      "loss": 1.4076,
      "theoretical_loss": 3.3853660507920007,
      "tokens_seen": 2611216384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021085255767301904,
      "loss": 1.2887,
      "theoretical_loss": 3.3853593889116933,
      "tokens_seen": 2611281920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021083249749247743,
      "loss": 1.3198,
      "theoretical_loss": 3.3853527272453916,
      "tokens_seen": 2611347456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002108124373119358,
      "loss": 1.3302,
      "theoretical_loss": 3.385346065793084,
      "tokens_seen": 2611412992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021079237713139419,
      "loss": 1.2385,
      "theoretical_loss": 3.385339404554758,
      "tokens_seen": 2611478528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021077231695085255,
      "loss": 1.2978,
      "theoretical_loss": 3.385332743530401,
      "tokens_seen": 2611544064
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021075225677031094,
      "loss": 1.2065,
      "theoretical_loss": 3.385326082720001,
      "tokens_seen": 2611609600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002107321965897693,
      "loss": 1.2002,
      "theoretical_loss": 3.385319422123546,
      "tokens_seen": 2611675136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002107121364092277,
      "loss": 1.3562,
      "theoretical_loss": 3.3853127617410235,
      "tokens_seen": 2611740672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021069207622868605,
      "loss": 1.3017,
      "theoretical_loss": 3.385306101572421,
      "tokens_seen": 2611806208
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021067201604814444,
      "loss": 1.3705,
      "theoretical_loss": 3.385299441617727,
      "tokens_seen": 2611871744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002106519558676028,
      "loss": 1.2761,
      "theoretical_loss": 3.385292781876929,
      "tokens_seen": 2611937280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021063189568706116,
      "loss": 1.3827,
      "theoretical_loss": 3.3852861223500144,
      "tokens_seen": 2612002816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021061183550651955,
      "loss": 1.2891,
      "theoretical_loss": 3.385279463036971,
      "tokens_seen": 2612068352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021059177532597792,
      "loss": 1.3492,
      "theoretical_loss": 3.385272803937787,
      "tokens_seen": 2612133888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021057171514543633,
      "loss": 1.2769,
      "theoretical_loss": 3.3852661450524497,
      "tokens_seen": 2612199424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002105516549648947,
      "loss": 1.2566,
      "theoretical_loss": 3.3852594863809475,
      "tokens_seen": 2612264960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021053159478435308,
      "loss": 1.2464,
      "theoretical_loss": 3.385252827923267,
      "tokens_seen": 2612330496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021051153460381145,
      "loss": 1.1797,
      "theoretical_loss": 3.385246169679397,
      "tokens_seen": 2612396032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021049147442326984,
      "loss": 1.2945,
      "theoretical_loss": 3.3852395116493255,
      "tokens_seen": 2612461568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002104714142427282,
      "loss": 1.273,
      "theoretical_loss": 3.385232853833039,
      "tokens_seen": 2612527104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021045135406218656,
      "loss": 1.3013,
      "theoretical_loss": 3.385226196230527,
      "tokens_seen": 2612592640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021043129388164495,
      "loss": 1.3807,
      "theoretical_loss": 3.385219538841776,
      "tokens_seen": 2612658176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002104112337011033,
      "loss": 1.3754,
      "theoretical_loss": 3.3852128816667735,
      "tokens_seen": 2612723712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002103911735205617,
      "loss": 1.2452,
      "theoretical_loss": 3.385206224705508,
      "tokens_seen": 2612789248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021037111334002006,
      "loss": 1.3381,
      "theoretical_loss": 3.3851995679579674,
      "tokens_seen": 2612854784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021035105315947845,
      "loss": 1.3398,
      "theoretical_loss": 3.385192911424139,
      "tokens_seen": 2612920320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021033099297893681,
      "loss": 1.3751,
      "theoretical_loss": 3.385186255104011,
      "tokens_seen": 2612985856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002103109327983952,
      "loss": 1.3993,
      "theoretical_loss": 3.3851795989975706,
      "tokens_seen": 2613051392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021029087261785357,
      "loss": 1.3228,
      "theoretical_loss": 3.3851729431048065,
      "tokens_seen": 2613116928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021027081243731196,
      "loss": 1.2678,
      "theoretical_loss": 3.3851662874257054,
      "tokens_seen": 2613182464
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021025075225677032,
      "loss": 1.2178,
      "theoretical_loss": 3.385159631960256,
      "tokens_seen": 2613248000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021023069207622868,
      "loss": 1.2701,
      "theoretical_loss": 3.3851529767084454,
      "tokens_seen": 2613313536
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 950332,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1345751285552979,
      "objective/train/theoretical_loss": 3.3851463216702617,
      "objective/train/tokens_used": 983838176,
      "theoretical_loss": 3.3851463216702617,
      "tokens_seen": 2613379072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021021063189568707,
      "loss": 1.2405,
      "theoretical_loss": 3.3851463216702617,
      "tokens_seen": 2613379072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021019057171514543,
      "loss": 1.2675,
      "theoretical_loss": 3.385139666845693,
      "tokens_seen": 2613444608
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021017051153460382,
      "loss": 1.3132,
      "theoretical_loss": 3.3851330122347267,
      "tokens_seen": 2613510144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021015045135406218,
      "loss": 1.2989,
      "theoretical_loss": 3.3851263578373505,
      "tokens_seen": 2613575680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021013039117352057,
      "loss": 1.352,
      "theoretical_loss": 3.3851197036535523,
      "tokens_seen": 2613641216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021011033099297894,
      "loss": 1.2469,
      "theoretical_loss": 3.38511304968332,
      "tokens_seen": 2613706752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021009027081243732,
      "loss": 1.3733,
      "theoretical_loss": 3.385106395926641,
      "tokens_seen": 2613772288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002100702106318957,
      "loss": 1.3916,
      "theoretical_loss": 3.385099742383504,
      "tokens_seen": 2613837824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021005015045135405,
      "loss": 1.2952,
      "theoretical_loss": 3.3850930890538957,
      "tokens_seen": 2613903360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021003009027081244,
      "loss": 1.3203,
      "theoretical_loss": 3.385086435937805,
      "tokens_seen": 2613968896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002100100300902708,
      "loss": 1.3856,
      "theoretical_loss": 3.385079783035218,
      "tokens_seen": 2614034432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002099899699097292,
      "loss": 1.4235,
      "theoretical_loss": 3.3850731303461243,
      "tokens_seen": 2614099968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020996990972918755,
      "loss": 1.3422,
      "theoretical_loss": 3.3850664778705113,
      "tokens_seen": 2614165504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020994984954864594,
      "loss": 1.2666,
      "theoretical_loss": 3.3850598256083657,
      "tokens_seen": 2614231040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002099297893681043,
      "loss": 1.2404,
      "theoretical_loss": 3.385053173559677,
      "tokens_seen": 2614296576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002099097291875627,
      "loss": 1.3187,
      "theoretical_loss": 3.385046521724431,
      "tokens_seen": 2614362112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020988966900702106,
      "loss": 1.2882,
      "theoretical_loss": 3.3850398701026174,
      "tokens_seen": 2614427648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020986960882647942,
      "loss": 1.2034,
      "theoretical_loss": 3.385033218694223,
      "tokens_seen": 2614493184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002098495486459378,
      "loss": 1.3275,
      "theoretical_loss": 3.385026567499236,
      "tokens_seen": 2614558720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002098294884653962,
      "loss": 1.3267,
      "theoretical_loss": 3.3850199165176433,
      "tokens_seen": 2614624256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020980942828485459,
      "loss": 1.2717,
      "theoretical_loss": 3.385013265749434,
      "tokens_seen": 2614689792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020978936810431295,
      "loss": 1.3211,
      "theoretical_loss": 3.385006615194595,
      "tokens_seen": 2614755328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020976930792377134,
      "loss": 1.2694,
      "theoretical_loss": 3.3849999648531144,
      "tokens_seen": 2614820864
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002097492477432297,
      "loss": 1.4034,
      "theoretical_loss": 3.38499331472498,
      "tokens_seen": 2614886400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002097291875626881,
      "loss": 1.2451,
      "theoretical_loss": 3.38498666481018,
      "tokens_seen": 2614951936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020970912738214645,
      "loss": 1.3335,
      "theoretical_loss": 3.3849800151087015,
      "tokens_seen": 2615017472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020968906720160484,
      "loss": 1.3094,
      "theoretical_loss": 3.3849733656205325,
      "tokens_seen": 2615083008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002096690070210632,
      "loss": 1.3608,
      "theoretical_loss": 3.3849667163456614,
      "tokens_seen": 2615148544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020964894684052156,
      "loss": 1.2232,
      "theoretical_loss": 3.384960067284075,
      "tokens_seen": 2615214080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020962888665997995,
      "loss": 1.3217,
      "theoretical_loss": 3.384953418435762,
      "tokens_seen": 2615279616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020960882647943832,
      "loss": 1.2801,
      "theoretical_loss": 3.38494676980071,
      "tokens_seen": 2615345152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002095887662988967,
      "loss": 1.1749,
      "theoretical_loss": 3.3849401213789068,
      "tokens_seen": 2615410688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020956870611835507,
      "loss": 1.2152,
      "theoretical_loss": 3.3849334731703395,
      "tokens_seen": 2615476224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020954864593781346,
      "loss": 1.3174,
      "theoretical_loss": 3.3849268251749973,
      "tokens_seen": 2615541760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020952858575727182,
      "loss": 1.2834,
      "theoretical_loss": 3.3849201773928668,
      "tokens_seen": 2615607296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002095085255767302,
      "loss": 1.3839,
      "theoretical_loss": 3.384913529823937,
      "tokens_seen": 2615672832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020948846539618857,
      "loss": 1.2381,
      "theoretical_loss": 3.384906882468194,
      "tokens_seen": 2615738368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020946840521564693,
      "loss": 1.2972,
      "theoretical_loss": 3.384900235325627,
      "tokens_seen": 2615803904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020944834503510532,
      "loss": 1.3334,
      "theoretical_loss": 3.3848935883962237,
      "tokens_seen": 2615869440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020942828485456368,
      "loss": 1.2936,
      "theoretical_loss": 3.384886941679971,
      "tokens_seen": 2615934976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020940822467402207,
      "loss": 1.2695,
      "theoretical_loss": 3.3848802951768584,
      "tokens_seen": 2616000512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020938816449348044,
      "loss": 1.2069,
      "theoretical_loss": 3.384873648886872,
      "tokens_seen": 2616066048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020936810431293883,
      "loss": 1.3178,
      "theoretical_loss": 3.384867002810001,
      "tokens_seen": 2616131584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002093480441323972,
      "loss": 1.2583,
      "theoretical_loss": 3.384860356946232,
      "tokens_seen": 2616197120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020932798395185558,
      "loss": 1.3277,
      "theoretical_loss": 3.3848537112955537,
      "tokens_seen": 2616262656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020930792377131394,
      "loss": 1.3963,
      "theoretical_loss": 3.3848470658579535,
      "tokens_seen": 2616328192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002092878635907723,
      "loss": 1.2654,
      "theoretical_loss": 3.3848404206334193,
      "tokens_seen": 2616393728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002092678034102307,
      "loss": 1.2415,
      "theoretical_loss": 3.384833775621939,
      "tokens_seen": 2616459264
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020924774322968905,
      "loss": 1.4234,
      "theoretical_loss": 3.3848271308235005,
      "tokens_seen": 2616524800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020922768304914744,
      "loss": 1.2251,
      "theoretical_loss": 3.3848204862380915,
      "tokens_seen": 2616590336
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 951607,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1770057678222656,
      "objective/train/theoretical_loss": 3.3848138418657,
      "objective/train/tokens_used": 987114976,
      "theoretical_loss": 3.3848138418657,
      "tokens_seen": 2616655872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002092076228686058,
      "loss": 1.2223,
      "theoretical_loss": 3.3848138418657,
      "tokens_seen": 2616655872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091875626880642,
      "loss": 1.2837,
      "theoretical_loss": 3.384807197706314,
      "tokens_seen": 2616721408
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020916750250752256,
      "loss": 1.2457,
      "theoretical_loss": 3.384800553759921,
      "tokens_seen": 2616786944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020914744232698095,
      "loss": 1.2019,
      "theoretical_loss": 3.3847939100265085,
      "tokens_seen": 2616852480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091273821464393,
      "loss": 1.376,
      "theoretical_loss": 3.384787266506065,
      "tokens_seen": 2616918016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091073219658977,
      "loss": 1.3133,
      "theoretical_loss": 3.384780623198578,
      "tokens_seen": 2616983552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020908726178535606,
      "loss": 1.284,
      "theoretical_loss": 3.3847739801040357,
      "tokens_seen": 2617049088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020906720160481445,
      "loss": 1.3725,
      "theoretical_loss": 3.3847673372224256,
      "tokens_seen": 2617114624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020904714142427284,
      "loss": 1.2273,
      "theoretical_loss": 3.3847606945537354,
      "tokens_seen": 2617180160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002090270812437312,
      "loss": 1.2256,
      "theoretical_loss": 3.3847540520979535,
      "tokens_seen": 2617245696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002090070210631896,
      "loss": 1.176,
      "theoretical_loss": 3.384747409855067,
      "tokens_seen": 2617311232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020898696088264795,
      "loss": 1.3427,
      "theoretical_loss": 3.3847407678250647,
      "tokens_seen": 2617376768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020896690070210634,
      "loss": 1.3989,
      "theoretical_loss": 3.3847341260079338,
      "tokens_seen": 2617442304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002089468405215647,
      "loss": 1.3529,
      "theoretical_loss": 3.384727484403662,
      "tokens_seen": 2617507840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002089267803410231,
      "loss": 1.2614,
      "theoretical_loss": 3.3847208430122375,
      "tokens_seen": 2617573376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020890672016048146,
      "loss": 1.2949,
      "theoretical_loss": 3.384714201833648,
      "tokens_seen": 2617638912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020888665997993982,
      "loss": 1.1856,
      "theoretical_loss": 3.384707560867881,
      "tokens_seen": 2617704448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002088665997993982,
      "loss": 1.2664,
      "theoretical_loss": 3.384700920114925,
      "tokens_seen": 2617769984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020884653961885657,
      "loss": 1.2858,
      "theoretical_loss": 3.3846942795747683,
      "tokens_seen": 2617835520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020882647943831496,
      "loss": 1.3635,
      "theoretical_loss": 3.3846876392473972,
      "tokens_seen": 2617901056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020880641925777332,
      "loss": 1.2692,
      "theoretical_loss": 3.384680999132801,
      "tokens_seen": 2617966592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002087863590772317,
      "loss": 1.3224,
      "theoretical_loss": 3.3846743592309667,
      "tokens_seen": 2618032128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020876629889669007,
      "loss": 1.2806,
      "theoretical_loss": 3.3846677195418824,
      "tokens_seen": 2618097664
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020874623871614846,
      "loss": 1.2066,
      "theoretical_loss": 3.384661080065536,
      "tokens_seen": 2618163200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020872617853560682,
      "loss": 1.2574,
      "theoretical_loss": 3.3846544408019152,
      "tokens_seen": 2618228736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020870611835506519,
      "loss": 1.2838,
      "theoretical_loss": 3.3846478017510084,
      "tokens_seen": 2618294272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020868605817452358,
      "loss": 1.4147,
      "theoretical_loss": 3.384641162912803,
      "tokens_seen": 2618359808
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020866599799398194,
      "loss": 1.2744,
      "theoretical_loss": 3.3846345242872866,
      "tokens_seen": 2618425344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020864593781344033,
      "loss": 1.175,
      "theoretical_loss": 3.3846278858744476,
      "tokens_seen": 2618490880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002086258776328987,
      "loss": 1.462,
      "theoretical_loss": 3.3846212476742736,
      "tokens_seen": 2618556416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020860581745235708,
      "loss": 1.2659,
      "theoretical_loss": 3.3846146096867527,
      "tokens_seen": 2618621952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020858575727181544,
      "loss": 1.3922,
      "theoretical_loss": 3.3846079719118727,
      "tokens_seen": 2618687488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020856569709127383,
      "loss": 1.2167,
      "theoretical_loss": 3.384601334349621,
      "tokens_seen": 2618753024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002085456369107322,
      "loss": 1.2984,
      "theoretical_loss": 3.3845946969999856,
      "tokens_seen": 2618818560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020852557673019058,
      "loss": 1.3219,
      "theoretical_loss": 3.384588059862955,
      "tokens_seen": 2618884096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020850551654964894,
      "loss": 1.3638,
      "theoretical_loss": 3.3845814229385165,
      "tokens_seen": 2618949632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002084854563691073,
      "loss": 1.3519,
      "theoretical_loss": 3.384574786226658,
      "tokens_seen": 2619015168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002084653961885657,
      "loss": 1.3442,
      "theoretical_loss": 3.384568149727368,
      "tokens_seen": 2619080704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020844533600802406,
      "loss": 1.3298,
      "theoretical_loss": 3.3845615134406337,
      "tokens_seen": 2619146240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020842527582748245,
      "loss": 1.3213,
      "theoretical_loss": 3.384554877366443,
      "tokens_seen": 2619211776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002084052156469408,
      "loss": 1.3295,
      "theoretical_loss": 3.384548241504784,
      "tokens_seen": 2619277312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002083851554663992,
      "loss": 1.3272,
      "theoretical_loss": 3.3845416058556443,
      "tokens_seen": 2619342848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020836509528585756,
      "loss": 1.394,
      "theoretical_loss": 3.384534970419012,
      "tokens_seen": 2619408384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020834503510531595,
      "loss": 1.2891,
      "theoretical_loss": 3.3845283351948754,
      "tokens_seen": 2619473920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020832497492477434,
      "loss": 1.3231,
      "theoretical_loss": 3.3845217001832215,
      "tokens_seen": 2619539456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002083049147442327,
      "loss": 1.3214,
      "theoretical_loss": 3.384515065384039,
      "tokens_seen": 2619604992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082848545636911,
      "loss": 1.2981,
      "theoretical_loss": 3.384508430797315,
      "tokens_seen": 2619670528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020826479438314945,
      "loss": 1.3241,
      "theoretical_loss": 3.384501796423038,
      "tokens_seen": 2619736064
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020824473420260784,
      "loss": 1.3119,
      "theoretical_loss": 3.3844951622611954,
      "tokens_seen": 2619801600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082246740220662,
      "loss": 1.3339,
      "theoretical_loss": 3.384488528311776,
      "tokens_seen": 2619867136
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 952663,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2578439712524414,
      "objective/train/theoretical_loss": 3.3844818945747663,
      "objective/train/tokens_used": 990391776,
      "theoretical_loss": 3.3844818945747663,
      "tokens_seen": 2619932672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082046138415246,
      "loss": 1.2398,
      "theoretical_loss": 3.3844818945747663,
      "tokens_seen": 2619932672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020818455366098296,
      "loss": 1.2191,
      "theoretical_loss": 3.384475261050155,
      "tokens_seen": 2619998208
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020816449348044135,
      "loss": 1.2984,
      "theoretical_loss": 3.38446862773793,
      "tokens_seen": 2620063744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002081444332998997,
      "loss": 1.287,
      "theoretical_loss": 3.3844619946380794,
      "tokens_seen": 2620129280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002081243731193581,
      "loss": 1.355,
      "theoretical_loss": 3.3844553617505904,
      "tokens_seen": 2620194816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020810431293881646,
      "loss": 1.4031,
      "theoretical_loss": 3.3844487290754515,
      "tokens_seen": 2620260352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020808425275827482,
      "loss": 1.3084,
      "theoretical_loss": 3.3844420966126503,
      "tokens_seen": 2620325888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002080641925777332,
      "loss": 1.1663,
      "theoretical_loss": 3.384435464362175,
      "tokens_seen": 2620391424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020804413239719157,
      "loss": 1.2969,
      "theoretical_loss": 3.3844288323240126,
      "tokens_seen": 2620456960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020802407221664996,
      "loss": 1.2917,
      "theoretical_loss": 3.384422200498152,
      "tokens_seen": 2620522496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020800401203610833,
      "loss": 1.2512,
      "theoretical_loss": 3.384415568884581,
      "tokens_seen": 2620588032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020798395185556671,
      "loss": 1.3212,
      "theoretical_loss": 3.384408937483287,
      "tokens_seen": 2620653568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020796389167502508,
      "loss": 1.4728,
      "theoretical_loss": 3.384402306294258,
      "tokens_seen": 2620719104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020794383149448347,
      "loss": 1.3355,
      "theoretical_loss": 3.384395675317482,
      "tokens_seen": 2620784640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020792377131394183,
      "loss": 1.2694,
      "theoretical_loss": 3.384389044552947,
      "tokens_seen": 2620850176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002079037111334002,
      "loss": 1.3449,
      "theoretical_loss": 3.384382414000641,
      "tokens_seen": 2620915712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020788365095285858,
      "loss": 1.3319,
      "theoretical_loss": 3.3843757836605515,
      "tokens_seen": 2620981248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020786359077231694,
      "loss": 1.2803,
      "theoretical_loss": 3.3843691535326665,
      "tokens_seen": 2621046784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020784353059177533,
      "loss": 1.2715,
      "theoretical_loss": 3.384362523616974,
      "tokens_seen": 2621112320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002078234704112337,
      "loss": 1.383,
      "theoretical_loss": 3.3843558939134626,
      "tokens_seen": 2621177856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020780341023069208,
      "loss": 1.1835,
      "theoretical_loss": 3.384349264422119,
      "tokens_seen": 2621243392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020778335005015045,
      "loss": 1.2959,
      "theoretical_loss": 3.3843426351429318,
      "tokens_seen": 2621308928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020776328986960884,
      "loss": 1.2128,
      "theoretical_loss": 3.3843360060758885,
      "tokens_seen": 2621374464
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002077432296890672,
      "loss": 1.266,
      "theoretical_loss": 3.3843293772209777,
      "tokens_seen": 2621440000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020772316950852556,
      "loss": 1.3024,
      "theoretical_loss": 3.3843227485781866,
      "tokens_seen": 2621505536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020770310932798395,
      "loss": 1.2488,
      "theoretical_loss": 3.384316120147503,
      "tokens_seen": 2621571072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002076830491474423,
      "loss": 1.3177,
      "theoretical_loss": 3.384309491928916,
      "tokens_seen": 2621636608
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002076629889669007,
      "loss": 1.3799,
      "theoretical_loss": 3.3843028639224118,
      "tokens_seen": 2621702144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020764292878635906,
      "loss": 1.3233,
      "theoretical_loss": 3.38429623612798,
      "tokens_seen": 2621767680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020762286860581745,
      "loss": 1.2874,
      "theoretical_loss": 3.384289608545607,
      "tokens_seen": 2621833216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020760280842527581,
      "loss": 1.283,
      "theoretical_loss": 3.384282981175282,
      "tokens_seen": 2621898752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002075827482447342,
      "loss": 1.3398,
      "theoretical_loss": 3.3842763540169924,
      "tokens_seen": 2621964288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002075626880641926,
      "loss": 1.164,
      "theoretical_loss": 3.3842697270707256,
      "tokens_seen": 2622029824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020754262788365098,
      "loss": 1.3305,
      "theoretical_loss": 3.3842631003364705,
      "tokens_seen": 2622095360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020752256770310934,
      "loss": 1.3078,
      "theoretical_loss": 3.3842564738142142,
      "tokens_seen": 2622160896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002075025075225677,
      "loss": 1.3596,
      "theoretical_loss": 3.384249847503945,
      "tokens_seen": 2622226432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002074824473420261,
      "loss": 1.3373,
      "theoretical_loss": 3.3842432214056504,
      "tokens_seen": 2622291968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020746238716148446,
      "loss": 1.3418,
      "theoretical_loss": 3.3842365955193188,
      "tokens_seen": 2622357504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020744232698094285,
      "loss": 1.321,
      "theoretical_loss": 3.3842299698449385,
      "tokens_seen": 2622423040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002074222668004012,
      "loss": 1.2428,
      "theoretical_loss": 3.3842233443824963,
      "tokens_seen": 2622488576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002074022066198596,
      "loss": 1.2696,
      "theoretical_loss": 3.384216719131981,
      "tokens_seen": 2622554112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020738214643931796,
      "loss": 1.268,
      "theoretical_loss": 3.3842100940933806,
      "tokens_seen": 2622619648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020736208625877635,
      "loss": 1.24,
      "theoretical_loss": 3.3842034692666823,
      "tokens_seen": 2622685184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002073420260782347,
      "loss": 1.3166,
      "theoretical_loss": 3.384196844651874,
      "tokens_seen": 2622750720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020732196589769308,
      "loss": 1.3988,
      "theoretical_loss": 3.3841902202489447,
      "tokens_seen": 2622816256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020730190571715146,
      "loss": 1.2268,
      "theoretical_loss": 3.3841835960578814,
      "tokens_seen": 2622881792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020728184553660983,
      "loss": 1.3014,
      "theoretical_loss": 3.3841769720786727,
      "tokens_seen": 2622947328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020726178535606822,
      "loss": 1.2907,
      "theoretical_loss": 3.3841703483113057,
      "tokens_seen": 2623012864
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020724172517552658,
      "loss": 1.3183,
      "theoretical_loss": 3.3841637247557688,
      "tokens_seen": 2623078400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020722166499498497,
      "loss": 1.1986,
      "theoretical_loss": 3.38415710141205,
      "tokens_seen": 2623143936
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 953489,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.287718653678894,
      "objective/train/theoretical_loss": 3.3841504782801373,
      "objective/train/tokens_used": 993668576,
      "theoretical_loss": 3.3841504782801373,
      "tokens_seen": 2623209472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020720160481444333,
      "loss": 1.1881,
      "theoretical_loss": 3.3841504782801373,
      "tokens_seen": 2623209472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020718154463390172,
      "loss": 1.3365,
      "theoretical_loss": 3.3841438553600183,
      "tokens_seen": 2623275008
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020716148445336008,
      "loss": 1.258,
      "theoretical_loss": 3.384137232651681,
      "tokens_seen": 2623340544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020714142427281844,
      "loss": 1.3091,
      "theoretical_loss": 3.3841306101551134,
      "tokens_seen": 2623406080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020712136409227683,
      "loss": 1.3221,
      "theoretical_loss": 3.3841239878703036,
      "tokens_seen": 2623471616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002071013039117352,
      "loss": 1.1964,
      "theoretical_loss": 3.3841173657972394,
      "tokens_seen": 2623537152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020708124373119358,
      "loss": 1.3042,
      "theoretical_loss": 3.384110743935909,
      "tokens_seen": 2623602688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020706118355065195,
      "loss": 1.3658,
      "theoretical_loss": 3.3841041222863,
      "tokens_seen": 2623668224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020704112337011034,
      "loss": 1.3491,
      "theoretical_loss": 3.3840975008484,
      "tokens_seen": 2623733760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002070210631895687,
      "loss": 1.287,
      "theoretical_loss": 3.3840908796221982,
      "tokens_seen": 2623799296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002070010030090271,
      "loss": 1.4193,
      "theoretical_loss": 3.384084258607681,
      "tokens_seen": 2623864832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020698094282848545,
      "loss": 1.3772,
      "theoretical_loss": 3.384077637804838,
      "tokens_seen": 2623930368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020696088264794384,
      "loss": 1.4291,
      "theoretical_loss": 3.384071017213655,
      "tokens_seen": 2623995904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002069408224674022,
      "loss": 1.2768,
      "theoretical_loss": 3.384064396834122,
      "tokens_seen": 2624061440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020692076228686056,
      "loss": 1.3338,
      "theoretical_loss": 3.384057776666226,
      "tokens_seen": 2624126976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020690070210631895,
      "loss": 1.2407,
      "theoretical_loss": 3.3840511567099547,
      "tokens_seen": 2624192512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020688064192577732,
      "loss": 1.3007,
      "theoretical_loss": 3.384044536965297,
      "tokens_seen": 2624258048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002068605817452357,
      "loss": 1.298,
      "theoretical_loss": 3.3840379174322397,
      "tokens_seen": 2624323584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020684052156469407,
      "loss": 1.2544,
      "theoretical_loss": 3.384031298110772,
      "tokens_seen": 2624389120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020682046138415248,
      "loss": 1.3135,
      "theoretical_loss": 3.3840246790008806,
      "tokens_seen": 2624454656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020680040120361085,
      "loss": 1.2356,
      "theoretical_loss": 3.3840180601025542,
      "tokens_seen": 2624520192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020678034102306924,
      "loss": 1.2767,
      "theoretical_loss": 3.3840114414157805,
      "tokens_seen": 2624585728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002067602808425276,
      "loss": 1.2168,
      "theoretical_loss": 3.3840048229405477,
      "tokens_seen": 2624651264
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020674022066198596,
      "loss": 1.3496,
      "theoretical_loss": 3.3839982046768435,
      "tokens_seen": 2624716800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020672016048144435,
      "loss": 1.2074,
      "theoretical_loss": 3.383991586624656,
      "tokens_seen": 2624782336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002067001003009027,
      "loss": 1.4021,
      "theoretical_loss": 3.383984968783973,
      "tokens_seen": 2624847872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002066800401203611,
      "loss": 1.3027,
      "theoretical_loss": 3.3839783511547825,
      "tokens_seen": 2624913408
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020665997993981946,
      "loss": 1.2421,
      "theoretical_loss": 3.3839717337370727,
      "tokens_seen": 2624978944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020663991975927785,
      "loss": 1.3328,
      "theoretical_loss": 3.3839651165308315,
      "tokens_seen": 2625044480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020661985957873621,
      "loss": 1.3721,
      "theoretical_loss": 3.383958499536047,
      "tokens_seen": 2625110016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002065997993981946,
      "loss": 1.3472,
      "theoretical_loss": 3.383951882752706,
      "tokens_seen": 2625175552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020657973921765297,
      "loss": 1.3602,
      "theoretical_loss": 3.3839452661807985,
      "tokens_seen": 2625241088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020655967903711136,
      "loss": 1.4061,
      "theoretical_loss": 3.383938649820311,
      "tokens_seen": 2625306624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020653961885656972,
      "loss": 1.2976,
      "theoretical_loss": 3.3839320336712317,
      "tokens_seen": 2625372160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020651955867602808,
      "loss": 1.3779,
      "theoretical_loss": 3.3839254177335487,
      "tokens_seen": 2625437696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020649949849548647,
      "loss": 1.417,
      "theoretical_loss": 3.3839188020072495,
      "tokens_seen": 2625503232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020647943831494483,
      "loss": 1.2843,
      "theoretical_loss": 3.3839121864923234,
      "tokens_seen": 2625568768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020645937813440322,
      "loss": 1.304,
      "theoretical_loss": 3.383905571188757,
      "tokens_seen": 2625634304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020643931795386158,
      "loss": 1.2843,
      "theoretical_loss": 3.383898956096539,
      "tokens_seen": 2625699840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020641925777331997,
      "loss": 1.237,
      "theoretical_loss": 3.383892341215657,
      "tokens_seen": 2625765376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020639919759277833,
      "loss": 1.3371,
      "theoretical_loss": 3.383885726546099,
      "tokens_seen": 2625830912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020637913741223672,
      "loss": 1.3473,
      "theoretical_loss": 3.383879112087853,
      "tokens_seen": 2625896448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020635907723169509,
      "loss": 1.4216,
      "theoretical_loss": 3.3838724978409074,
      "tokens_seen": 2625961984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020633901705115345,
      "loss": 1.4312,
      "theoretical_loss": 3.38386588380525,
      "tokens_seen": 2626027520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020631895687061184,
      "loss": 1.2335,
      "theoretical_loss": 3.383859269980868,
      "tokens_seen": 2626093056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062988966900702,
      "loss": 1.2849,
      "theoretical_loss": 3.3838526563677505,
      "tokens_seen": 2626158592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062788365095286,
      "loss": 1.389,
      "theoretical_loss": 3.383846042965885,
      "tokens_seen": 2626224128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020625877632898695,
      "loss": 1.2948,
      "theoretical_loss": 3.3838394297752594,
      "tokens_seen": 2626289664
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020623871614844534,
      "loss": 1.3282,
      "theoretical_loss": 3.3838328167958616,
      "tokens_seen": 2626355200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062186559679037,
      "loss": 1.3534,
      "theoretical_loss": 3.38382620402768,
      "tokens_seen": 2626420736
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 954702,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.0849006175994873,
      "objective/train/theoretical_loss": 3.383819591470702,
      "objective/train/tokens_used": 996945376,
      "theoretical_loss": 3.383819591470702,
      "tokens_seen": 2626486272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002061985957873621,
      "loss": 1.1477,
      "theoretical_loss": 3.383819591470702,
      "tokens_seen": 2626486272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020617853560682046,
      "loss": 1.2556,
      "theoretical_loss": 3.383812979124916,
      "tokens_seen": 2626551808
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020615847542627882,
      "loss": 1.3105,
      "theoretical_loss": 3.38380636699031,
      "tokens_seen": 2626617344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002061384152457372,
      "loss": 1.2859,
      "theoretical_loss": 3.383799755066872,
      "tokens_seen": 2626682880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020611835506519557,
      "loss": 1.3241,
      "theoretical_loss": 3.3837931433545894,
      "tokens_seen": 2626748416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020609829488465396,
      "loss": 1.3895,
      "theoretical_loss": 3.383786531853451,
      "tokens_seen": 2626813952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020607823470411235,
      "loss": 1.2616,
      "theoretical_loss": 3.3837799205634442,
      "tokens_seen": 2626879488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020605817452357074,
      "loss": 1.2748,
      "theoretical_loss": 3.3837733094845572,
      "tokens_seen": 2626945024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002060381143430291,
      "loss": 1.2992,
      "theoretical_loss": 3.3837666986167783,
      "tokens_seen": 2627010560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002060180541624875,
      "loss": 1.2374,
      "theoretical_loss": 3.3837600879600953,
      "tokens_seen": 2627076096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020599799398194585,
      "loss": 1.353,
      "theoretical_loss": 3.383753477514496,
      "tokens_seen": 2627141632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020597793380140424,
      "loss": 1.3047,
      "theoretical_loss": 3.383746867279968,
      "tokens_seen": 2627207168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002059578736208626,
      "loss": 1.3592,
      "theoretical_loss": 3.3837402572565005,
      "tokens_seen": 2627272704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020593781344032096,
      "loss": 1.2548,
      "theoretical_loss": 3.3837336474440805,
      "tokens_seen": 2627338240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020591775325977935,
      "loss": 1.329,
      "theoretical_loss": 3.383727037842696,
      "tokens_seen": 2627403776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020589769307923772,
      "loss": 1.2599,
      "theoretical_loss": 3.3837204284523357,
      "tokens_seen": 2627469312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002058776328986961,
      "loss": 1.1678,
      "theoretical_loss": 3.383713819272987,
      "tokens_seen": 2627534848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020585757271815447,
      "loss": 1.3363,
      "theoretical_loss": 3.3837072103046384,
      "tokens_seen": 2627600384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020583751253761286,
      "loss": 1.3702,
      "theoretical_loss": 3.383700601547277,
      "tokens_seen": 2627665920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020581745235707122,
      "loss": 1.2711,
      "theoretical_loss": 3.383693993000892,
      "tokens_seen": 2627731456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002057973921765296,
      "loss": 1.2179,
      "theoretical_loss": 3.3836873846654703,
      "tokens_seen": 2627796992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020577733199598797,
      "loss": 1.2412,
      "theoretical_loss": 3.383680776541001,
      "tokens_seen": 2627862528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020575727181544633,
      "loss": 1.2966,
      "theoretical_loss": 3.383674168627471,
      "tokens_seen": 2627928064
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020573721163490472,
      "loss": 1.4539,
      "theoretical_loss": 3.3836675609248688,
      "tokens_seen": 2627993600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020571715145436308,
      "loss": 1.3414,
      "theoretical_loss": 3.3836609534331825,
      "tokens_seen": 2628059136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020569709127382147,
      "loss": 1.189,
      "theoretical_loss": 3.3836543461524,
      "tokens_seen": 2628124672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020567703109327984,
      "loss": 1.3724,
      "theoretical_loss": 3.3836477390825097,
      "tokens_seen": 2628190208
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020565697091273823,
      "loss": 1.3481,
      "theoretical_loss": 3.3836411322234987,
      "tokens_seen": 2628255744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002056369107321966,
      "loss": 1.3553,
      "theoretical_loss": 3.383634525575356,
      "tokens_seen": 2628321280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020561685055165498,
      "loss": 1.292,
      "theoretical_loss": 3.3836279191380694,
      "tokens_seen": 2628386816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020559679037111334,
      "loss": 1.3067,
      "theoretical_loss": 3.383621312911626,
      "tokens_seen": 2628452352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002055767301905717,
      "loss": 1.2241,
      "theoretical_loss": 3.3836147068960147,
      "tokens_seen": 2628517888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002055566700100301,
      "loss": 1.2936,
      "theoretical_loss": 3.3836081010912236,
      "tokens_seen": 2628583424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020553660982948845,
      "loss": 1.2211,
      "theoretical_loss": 3.38360149549724,
      "tokens_seen": 2628648960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020551654964894684,
      "loss": 1.3436,
      "theoretical_loss": 3.3835948901140527,
      "tokens_seen": 2628714496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054964894684052,
      "loss": 1.2761,
      "theoretical_loss": 3.3835882849416494,
      "tokens_seen": 2628780032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054764292878636,
      "loss": 1.4263,
      "theoretical_loss": 3.3835816799800176,
      "tokens_seen": 2628845568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020545636910732196,
      "loss": 1.3702,
      "theoretical_loss": 3.383575075229146,
      "tokens_seen": 2628911104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020543630892678035,
      "loss": 1.3149,
      "theoretical_loss": 3.383568470689023,
      "tokens_seen": 2628976640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054162487462387,
      "loss": 1.3477,
      "theoretical_loss": 3.3835618663596354,
      "tokens_seen": 2629042176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053961885656971,
      "loss": 1.2733,
      "theoretical_loss": 3.383555262240972,
      "tokens_seen": 2629107712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020537612838515546,
      "loss": 1.3245,
      "theoretical_loss": 3.3835486583330208,
      "tokens_seen": 2629173248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020535606820461382,
      "loss": 1.2971,
      "theoretical_loss": 3.38354205463577,
      "tokens_seen": 2629238784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053360080240722,
      "loss": 1.39,
      "theoretical_loss": 3.3835354511492066,
      "tokens_seen": 2629304320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053159478435306,
      "loss": 1.3068,
      "theoretical_loss": 3.38352884787332,
      "tokens_seen": 2629369856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000205295887662989,
      "loss": 1.3838,
      "theoretical_loss": 3.3835222448080975,
      "tokens_seen": 2629435392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020527582748244735,
      "loss": 1.29,
      "theoretical_loss": 3.383515641953527,
      "tokens_seen": 2629500928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020525576730190574,
      "loss": 1.2904,
      "theoretical_loss": 3.3835090393095966,
      "tokens_seen": 2629566464
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002052357071213641,
      "loss": 1.3154,
      "theoretical_loss": 3.383502436876295,
      "tokens_seen": 2629632000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002052156469408225,
      "loss": 1.1901,
      "theoretical_loss": 3.3834958346536093,
      "tokens_seen": 2629697536
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 955785,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4545732736587524,
      "objective/train/theoretical_loss": 3.3834892326415282,
      "objective/train/tokens_used": 1000222176,
      "theoretical_loss": 3.3834892326415282,
      "tokens_seen": 2629763072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020519558676028086,
      "loss": 1.3051,
      "theoretical_loss": 3.3834892326415282,
      "tokens_seen": 2629763072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020517552657973922,
      "loss": 1.1722,
      "theoretical_loss": 3.3834826308400396,
      "tokens_seen": 2629828608
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002051554663991976,
      "loss": 1.2645,
      "theoretical_loss": 3.383476029249131,
      "tokens_seen": 2629894144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020513540621865597,
      "loss": 1.2748,
      "theoretical_loss": 3.3834694278687913,
      "tokens_seen": 2629959680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020511534603811436,
      "loss": 1.3362,
      "theoretical_loss": 3.3834628266990077,
      "tokens_seen": 2630025216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020509528585757272,
      "loss": 1.2211,
      "theoretical_loss": 3.383456225739769,
      "tokens_seen": 2630090752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002050752256770311,
      "loss": 1.3308,
      "theoretical_loss": 3.3834496249910626,
      "tokens_seen": 2630156288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020505516549648947,
      "loss": 1.2035,
      "theoretical_loss": 3.383443024452877,
      "tokens_seen": 2630221824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020503510531594786,
      "loss": 1.3142,
      "theoretical_loss": 3.3834364241252,
      "tokens_seen": 2630287360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020501504513540622,
      "loss": 1.1652,
      "theoretical_loss": 3.3834298240080196,
      "tokens_seen": 2630352896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020499498495486459,
      "loss": 1.2588,
      "theoretical_loss": 3.383423224101324,
      "tokens_seen": 2630418432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020497492477432298,
      "loss": 1.3933,
      "theoretical_loss": 3.383416624405101,
      "tokens_seen": 2630483968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020495486459378134,
      "loss": 1.3266,
      "theoretical_loss": 3.383410024919339,
      "tokens_seen": 2630549504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020493480441323973,
      "loss": 1.3421,
      "theoretical_loss": 3.3834034256440257,
      "tokens_seen": 2630615040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002049147442326981,
      "loss": 1.2816,
      "theoretical_loss": 3.3833968265791494,
      "tokens_seen": 2630680576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020489468405215648,
      "loss": 1.3787,
      "theoretical_loss": 3.383390227724698,
      "tokens_seen": 2630746112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020487462387161484,
      "loss": 1.2024,
      "theoretical_loss": 3.38338362908066,
      "tokens_seen": 2630811648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020485456369107323,
      "loss": 1.359,
      "theoretical_loss": 3.383377030647023,
      "tokens_seen": 2630877184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002048345035105316,
      "loss": 1.2657,
      "theoretical_loss": 3.3833704324237748,
      "tokens_seen": 2630942720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020481444332998998,
      "loss": 1.335,
      "theoretical_loss": 3.3833638344109036,
      "tokens_seen": 2631008256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020479438314944834,
      "loss": 1.3214,
      "theoretical_loss": 3.383357236608398,
      "tokens_seen": 2631073792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002047743229689067,
      "loss": 1.3942,
      "theoretical_loss": 3.3833506390162453,
      "tokens_seen": 2631139328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002047542627883651,
      "loss": 1.2536,
      "theoretical_loss": 3.3833440416344347,
      "tokens_seen": 2631204864
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020473420260782346,
      "loss": 1.2553,
      "theoretical_loss": 3.3833374444629527,
      "tokens_seen": 2631270400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020471414242728185,
      "loss": 1.1939,
      "theoretical_loss": 3.3833308475017883,
      "tokens_seen": 2631335936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002046940822467402,
      "loss": 1.1219,
      "theoretical_loss": 3.3833242507509294,
      "tokens_seen": 2631401472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002046740220661986,
      "loss": 1.2789,
      "theoretical_loss": 3.383317654210364,
      "tokens_seen": 2631467008
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020465396188565696,
      "loss": 1.259,
      "theoretical_loss": 3.3833110578800807,
      "tokens_seen": 2631532544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020463390170511535,
      "loss": 1.2403,
      "theoretical_loss": 3.3833044617600665,
      "tokens_seen": 2631598080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002046138415245737,
      "loss": 1.3504,
      "theoretical_loss": 3.3832978658503103,
      "tokens_seen": 2631663616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020459378134403207,
      "loss": 1.1628,
      "theoretical_loss": 3.3832912701508,
      "tokens_seen": 2631729152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002045737211634905,
      "loss": 1.2988,
      "theoretical_loss": 3.383284674661523,
      "tokens_seen": 2631794688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020455366098294885,
      "loss": 1.2068,
      "theoretical_loss": 3.3832780793824684,
      "tokens_seen": 2631860224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020453360080240724,
      "loss": 1.2834,
      "theoretical_loss": 3.3832714843136236,
      "tokens_seen": 2631925760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002045135406218656,
      "loss": 1.2207,
      "theoretical_loss": 3.383264889454977,
      "tokens_seen": 2631991296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000204493480441324,
      "loss": 1.1474,
      "theoretical_loss": 3.383258294806516,
      "tokens_seen": 2632056832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020447342026078236,
      "loss": 1.2967,
      "theoretical_loss": 3.3832517003682296,
      "tokens_seen": 2632122368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020445336008024075,
      "loss": 1.2123,
      "theoretical_loss": 3.3832451061401057,
      "tokens_seen": 2632187904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002044332998996991,
      "loss": 1.3014,
      "theoretical_loss": 3.383238512122132,
      "tokens_seen": 2632253440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002044132397191575,
      "loss": 1.3797,
      "theoretical_loss": 3.3832319183142965,
      "tokens_seen": 2632318976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020439317953861586,
      "loss": 1.3559,
      "theoretical_loss": 3.3832253247165873,
      "tokens_seen": 2632384512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020437311935807422,
      "loss": 1.2383,
      "theoretical_loss": 3.383218731328993,
      "tokens_seen": 2632450048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002043530591775326,
      "loss": 1.316,
      "theoretical_loss": 3.3832121381515012,
      "tokens_seen": 2632515584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020433299899699097,
      "loss": 1.257,
      "theoretical_loss": 3.3832055451841,
      "tokens_seen": 2632581120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020431293881644936,
      "loss": 1.2435,
      "theoretical_loss": 3.3831989524267776,
      "tokens_seen": 2632646656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020429287863590773,
      "loss": 1.2417,
      "theoretical_loss": 3.383192359879522,
      "tokens_seen": 2632712192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020427281845536611,
      "loss": 1.1876,
      "theoretical_loss": 3.3831857675423214,
      "tokens_seen": 2632777728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020425275827482448,
      "loss": 1.342,
      "theoretical_loss": 3.383179175415164,
      "tokens_seen": 2632843264
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020423269809428287,
      "loss": 1.1676,
      "theoretical_loss": 3.383172583498037,
      "tokens_seen": 2632908800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020421263791374123,
      "loss": 1.2602,
      "theoretical_loss": 3.38316599179093,
      "tokens_seen": 2632974336
    },
    {
      "debugging/Compilability": 1.0,
      "debugging/distinct-1-grams": 0.7520842035644667,
      "debugging/entropy-1-grams": 4.765424186883236,
      "debugging/length": 508.6666666666667,
      "debugging/num_segments": 6,
      "debugging/score": 0.0044137430059683815,
      "debugging/score_std": 0.0029234356048069474,
      "epoch": 0.6,
      "objective/train/docs_used": 956895,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2060573101043701,
      "objective/train/theoretical_loss": 3.3831594002938297,
      "objective/train/tokens_used": 1003498976,
      "theoretical_loss": 3.3831594002938297,
      "tokens_seen": 2633039872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002041925777331996,
      "loss": 1.281,
      "theoretical_loss": 3.3831594002938297,
      "tokens_seen": 2633039872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020417251755265798,
      "loss": 1.3121,
      "theoretical_loss": 3.3831528090067247,
      "tokens_seen": 2633105408
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020415245737211634,
      "loss": 1.2675,
      "theoretical_loss": 3.3831462179296032,
      "tokens_seen": 2633170944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020413239719157473,
      "loss": 1.2554,
      "theoretical_loss": 3.3831396270624534,
      "tokens_seen": 2633236480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002041123370110331,
      "loss": 1.2279,
      "theoretical_loss": 3.3831330364052628,
      "tokens_seen": 2633302016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020409227683049148,
      "loss": 1.2542,
      "theoretical_loss": 3.38312644595802,
      "tokens_seen": 2633367552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020407221664994985,
      "loss": 1.3532,
      "theoretical_loss": 3.383119855720713,
      "tokens_seen": 2633433088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020405215646940823,
      "loss": 1.3485,
      "theoretical_loss": 3.38311326569333,
      "tokens_seen": 2633498624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002040320962888666,
      "loss": 1.2305,
      "theoretical_loss": 3.3831066758758586,
      "tokens_seen": 2633564160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020401203610832496,
      "loss": 1.2319,
      "theoretical_loss": 3.3831000862682874,
      "tokens_seen": 2633629696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020399197592778335,
      "loss": 1.2706,
      "theoretical_loss": 3.3830934968706043,
      "tokens_seen": 2633695232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002039719157472417,
      "loss": 1.3699,
      "theoretical_loss": 3.3830869076827974,
      "tokens_seen": 2633760768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002039518555667001,
      "loss": 1.2627,
      "theoretical_loss": 3.383080318704855,
      "tokens_seen": 2633826304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020393179538615846,
      "loss": 1.237,
      "theoretical_loss": 3.383073729936765,
      "tokens_seen": 2633891840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020391173520561685,
      "loss": 1.3676,
      "theoretical_loss": 3.3830671413785147,
      "tokens_seen": 2633957376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020389167502507521,
      "loss": 1.3265,
      "theoretical_loss": 3.3830605530300932,
      "tokens_seen": 2634022912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002038716148445336,
      "loss": 1.329,
      "theoretical_loss": 3.383053964891489,
      "tokens_seen": 2634088448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020385155466399197,
      "loss": 1.2976,
      "theoretical_loss": 3.383047376962689,
      "tokens_seen": 2634153984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020383149448345036,
      "loss": 1.3263,
      "theoretical_loss": 3.3830407892436822,
      "tokens_seen": 2634219520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020381143430290874,
      "loss": 1.3919,
      "theoretical_loss": 3.3830342017344566,
      "tokens_seen": 2634285056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037913741223671,
      "loss": 1.2098,
      "theoretical_loss": 3.3830276144349996,
      "tokens_seen": 2634350592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037713139418255,
      "loss": 1.2501,
      "theoretical_loss": 3.3830210273452996,
      "tokens_seen": 2634416128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020375125376128386,
      "loss": 1.3143,
      "theoretical_loss": 3.3830144404653453,
      "tokens_seen": 2634481664
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020373119358074225,
      "loss": 1.2818,
      "theoretical_loss": 3.3830078537951245,
      "tokens_seen": 2634547200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037111334002006,
      "loss": 1.3101,
      "theoretical_loss": 3.383001267334625,
      "tokens_seen": 2634612736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000203691073219659,
      "loss": 1.3533,
      "theoretical_loss": 3.3829946810838347,
      "tokens_seen": 2634678272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020367101303911736,
      "loss": 1.2517,
      "theoretical_loss": 3.3829880950427427,
      "tokens_seen": 2634743808
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020365095285857575,
      "loss": 1.2034,
      "theoretical_loss": 3.3829815092113362,
      "tokens_seen": 2634809344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002036308926780341,
      "loss": 1.3102,
      "theoretical_loss": 3.382974923589604,
      "tokens_seen": 2634874880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020361083249749248,
      "loss": 1.2457,
      "theoretical_loss": 3.382968338177533,
      "tokens_seen": 2634940416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020359077231695086,
      "loss": 1.3118,
      "theoretical_loss": 3.382961752975113,
      "tokens_seen": 2635005952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020357071213640923,
      "loss": 1.3523,
      "theoretical_loss": 3.3829551679823306,
      "tokens_seen": 2635071488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020355065195586762,
      "loss": 1.3211,
      "theoretical_loss": 3.382948583199175,
      "tokens_seen": 2635137024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020353059177532598,
      "loss": 1.4232,
      "theoretical_loss": 3.3829419986256335,
      "tokens_seen": 2635202560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020351053159478437,
      "loss": 1.3094,
      "theoretical_loss": 3.382935414261695,
      "tokens_seen": 2635268096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020349047141424273,
      "loss": 1.3112,
      "theoretical_loss": 3.382928830107347,
      "tokens_seen": 2635333632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020347041123370112,
      "loss": 1.2773,
      "theoretical_loss": 3.3829222461625776,
      "tokens_seen": 2635399168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020345035105315948,
      "loss": 1.2746,
      "theoretical_loss": 3.3829156624273753,
      "tokens_seen": 2635464704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020343029087261784,
      "loss": 1.243,
      "theoretical_loss": 3.3829090789017284,
      "tokens_seen": 2635530240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020341023069207623,
      "loss": 1.2504,
      "theoretical_loss": 3.382902495585624,
      "tokens_seen": 2635595776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002033901705115346,
      "loss": 1.2461,
      "theoretical_loss": 3.3828959124790514,
      "tokens_seen": 2635661312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020337011033099298,
      "loss": 1.2918,
      "theoretical_loss": 3.382889329581998,
      "tokens_seen": 2635726848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020335005015045135,
      "loss": 1.2903,
      "theoretical_loss": 3.3828827468944525,
      "tokens_seen": 2635792384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020332998996990974,
      "loss": 1.2851,
      "theoretical_loss": 3.3828761644164023,
      "tokens_seen": 2635857920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002033099297893681,
      "loss": 1.2285,
      "theoretical_loss": 3.382869582147836,
      "tokens_seen": 2635923456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002032898696088265,
      "loss": 1.2742,
      "theoretical_loss": 3.3828630000887414,
      "tokens_seen": 2635988992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020326980942828485,
      "loss": 1.2928,
      "theoretical_loss": 3.382856418239107,
      "tokens_seen": 2636054528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020324974924774324,
      "loss": 1.101,
      "theoretical_loss": 3.382849836598921,
      "tokens_seen": 2636120064
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002032296890672016,
      "loss": 1.2575,
      "theoretical_loss": 3.382843255168171,
      "tokens_seen": 2636185600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020320962888665996,
      "loss": 1.1833,
      "theoretical_loss": 3.382836673946845,
      "tokens_seen": 2636251136
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 957995,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.201979160308838,
      "objective/train/theoretical_loss": 3.3828300929349324,
      "objective/train/tokens_used": 1006775776,
      "theoretical_loss": 3.3828300929349324,
      "tokens_seen": 2636316672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020318956870611835,
      "loss": 1.264,
      "theoretical_loss": 3.3828300929349324,
      "tokens_seen": 2636316672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020316950852557672,
      "loss": 1.2874,
      "theoretical_loss": 3.38282351213242,
      "tokens_seen": 2636382208
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002031494483450351,
      "loss": 1.3076,
      "theoretical_loss": 3.3828169315392964,
      "tokens_seen": 2636447744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020312938816449347,
      "loss": 1.3011,
      "theoretical_loss": 3.38281035115555,
      "tokens_seen": 2636513280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020310932798395186,
      "loss": 1.2906,
      "theoretical_loss": 3.3828037709811687,
      "tokens_seen": 2636578816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020308926780341022,
      "loss": 1.3196,
      "theoretical_loss": 3.3827971910161407,
      "tokens_seen": 2636644352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020306920762286864,
      "loss": 1.2277,
      "theoretical_loss": 3.3827906112604538,
      "tokens_seen": 2636709888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000203049147442327,
      "loss": 1.3058,
      "theoretical_loss": 3.3827840317140963,
      "tokens_seen": 2636775424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020302908726178536,
      "loss": 1.2802,
      "theoretical_loss": 3.3827774523770566,
      "tokens_seen": 2636840960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020300902708124375,
      "loss": 1.2996,
      "theoretical_loss": 3.3827708732493225,
      "tokens_seen": 2636906496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002029889669007021,
      "loss": 1.2672,
      "theoretical_loss": 3.382764294330882,
      "tokens_seen": 2636972032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002029689067201605,
      "loss": 1.2917,
      "theoretical_loss": 3.3827577156217243,
      "tokens_seen": 2637037568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020294884653961886,
      "loss": 1.2406,
      "theoretical_loss": 3.3827511371218364,
      "tokens_seen": 2637103104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020292878635907725,
      "loss": 1.2337,
      "theoretical_loss": 3.382744558831207,
      "tokens_seen": 2637168640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020290872617853561,
      "loss": 1.2899,
      "theoretical_loss": 3.382737980749824,
      "tokens_seen": 2637234176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202888665997994,
      "loss": 1.2923,
      "theoretical_loss": 3.3827314028776754,
      "tokens_seen": 2637299712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020286860581745237,
      "loss": 1.362,
      "theoretical_loss": 3.38272482521475,
      "tokens_seen": 2637365248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020284854563691073,
      "loss": 1.2281,
      "theoretical_loss": 3.382718247761035,
      "tokens_seen": 2637430784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020282848545636912,
      "loss": 1.272,
      "theoretical_loss": 3.382711670516519,
      "tokens_seen": 2637496320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020280842527582748,
      "loss": 1.2332,
      "theoretical_loss": 3.382705093481191,
      "tokens_seen": 2637561856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020278836509528587,
      "loss": 1.221,
      "theoretical_loss": 3.3826985166550374,
      "tokens_seen": 2637627392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020276830491474423,
      "loss": 1.2161,
      "theoretical_loss": 3.3826919400380477,
      "tokens_seen": 2637692928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020274824473420262,
      "loss": 1.3571,
      "theoretical_loss": 3.38268536363021,
      "tokens_seen": 2637758464
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020272818455366098,
      "loss": 1.2595,
      "theoretical_loss": 3.382678787431512,
      "tokens_seen": 2637824000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020270812437311937,
      "loss": 1.3324,
      "theoretical_loss": 3.3826722114419416,
      "tokens_seen": 2637889536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020268806419257773,
      "loss": 1.2602,
      "theoretical_loss": 3.3826656356614873,
      "tokens_seen": 2637955072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020266800401203612,
      "loss": 1.2448,
      "theoretical_loss": 3.3826590600901376,
      "tokens_seen": 2638020608
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020264794383149449,
      "loss": 1.2525,
      "theoretical_loss": 3.3826524847278803,
      "tokens_seen": 2638086144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020262788365095285,
      "loss": 1.1907,
      "theoretical_loss": 3.3826459095747037,
      "tokens_seen": 2638151680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020260782347041124,
      "loss": 1.2541,
      "theoretical_loss": 3.3826393346305954,
      "tokens_seen": 2638217216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002025877632898696,
      "loss": 1.3538,
      "theoretical_loss": 3.3826327598955443,
      "tokens_seen": 2638282752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202567703109328,
      "loss": 1.1886,
      "theoretical_loss": 3.382626185369538,
      "tokens_seen": 2638348288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020254764292878635,
      "loss": 1.312,
      "theoretical_loss": 3.382619611052565,
      "tokens_seen": 2638413824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020252758274824474,
      "loss": 1.2355,
      "theoretical_loss": 3.3826130369446137,
      "tokens_seen": 2638479360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002025075225677031,
      "loss": 1.3295,
      "theoretical_loss": 3.382606463045672,
      "tokens_seen": 2638544896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002024874623871615,
      "loss": 1.3521,
      "theoretical_loss": 3.3825998893557276,
      "tokens_seen": 2638610432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020246740220661985,
      "loss": 1.2493,
      "theoretical_loss": 3.382593315874769,
      "tokens_seen": 2638675968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020244734202607822,
      "loss": 1.3064,
      "theoretical_loss": 3.382586742602785,
      "tokens_seen": 2638741504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002024272818455366,
      "loss": 1.1987,
      "theoretical_loss": 3.3825801695397626,
      "tokens_seen": 2638807040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020240722166499497,
      "loss": 1.3061,
      "theoretical_loss": 3.382573596685691,
      "tokens_seen": 2638872576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020238716148445336,
      "loss": 1.2069,
      "theoretical_loss": 3.382567024040558,
      "tokens_seen": 2638938112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020236710130391172,
      "loss": 1.2174,
      "theoretical_loss": 3.3825604516043515,
      "tokens_seen": 2639003648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023470411233701,
      "loss": 1.2628,
      "theoretical_loss": 3.3825538793770598,
      "tokens_seen": 2639069184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023269809428285,
      "loss": 1.2979,
      "theoretical_loss": 3.3825473073586716,
      "tokens_seen": 2639134720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023069207622869,
      "loss": 1.3032,
      "theoretical_loss": 3.382540735549174,
      "tokens_seen": 2639200256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020228686058174525,
      "loss": 1.1514,
      "theoretical_loss": 3.382534163948556,
      "tokens_seen": 2639265792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020226680040120364,
      "loss": 1.3321,
      "theoretical_loss": 3.382527592556806,
      "tokens_seen": 2639331328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202246740220662,
      "loss": 1.2436,
      "theoretical_loss": 3.382521021373911,
      "tokens_seen": 2639396864
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020222668004012036,
      "loss": 1.2337,
      "theoretical_loss": 3.382514450399861,
      "tokens_seen": 2639462400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020220661985957875,
      "loss": 1.2755,
      "theoretical_loss": 3.3825078796346424,
      "tokens_seen": 2639527936
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 959162,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3204405307769775,
      "objective/train/theoretical_loss": 3.382501309078244,
      "objective/train/tokens_used": 1010052576,
      "theoretical_loss": 3.382501309078244,
      "tokens_seen": 2639593472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020218655967903712,
      "loss": 1.2452,
      "theoretical_loss": 3.382501309078244,
      "tokens_seen": 2639593472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002021664994984955,
      "loss": 1.2847,
      "theoretical_loss": 3.382494738730654,
      "tokens_seen": 2639659008
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020214643931795387,
      "loss": 1.1747,
      "theoretical_loss": 3.382488168591861,
      "tokens_seen": 2639724544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020212637913741226,
      "loss": 1.2614,
      "theoretical_loss": 3.382481598661853,
      "tokens_seen": 2639790080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020210631895687062,
      "loss": 1.2516,
      "theoretical_loss": 3.3824750289406174,
      "tokens_seen": 2639855616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202086258776329,
      "loss": 1.3555,
      "theoretical_loss": 3.3824684594281433,
      "tokens_seen": 2639921152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020206619859578737,
      "loss": 1.3245,
      "theoretical_loss": 3.382461890124419,
      "tokens_seen": 2639986688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020204613841524573,
      "loss": 1.2297,
      "theoretical_loss": 3.3824553210294317,
      "tokens_seen": 2640052224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020202607823470412,
      "loss": 1.2674,
      "theoretical_loss": 3.38244875214317,
      "tokens_seen": 2640117760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020200601805416248,
      "loss": 1.2367,
      "theoretical_loss": 3.382442183465623,
      "tokens_seen": 2640183296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020198595787362087,
      "loss": 1.1677,
      "theoretical_loss": 3.382435614996777,
      "tokens_seen": 2640248832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020196589769307924,
      "loss": 1.2467,
      "theoretical_loss": 3.3824290467366223,
      "tokens_seen": 2640314368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020194583751253763,
      "loss": 1.2175,
      "theoretical_loss": 3.382422478685146,
      "tokens_seen": 2640379904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201925777331996,
      "loss": 1.3024,
      "theoretical_loss": 3.3824159108423357,
      "tokens_seen": 2640445440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020190571715145438,
      "loss": 1.2672,
      "theoretical_loss": 3.382409343208181,
      "tokens_seen": 2640510976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020188565697091274,
      "loss": 1.302,
      "theoretical_loss": 3.382402775782669,
      "tokens_seen": 2640576512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002018655967903711,
      "loss": 1.2346,
      "theoretical_loss": 3.3823962085657886,
      "tokens_seen": 2640642048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002018455366098295,
      "loss": 1.2787,
      "theoretical_loss": 3.382389641557527,
      "tokens_seen": 2640707584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020182547642928785,
      "loss": 1.13,
      "theoretical_loss": 3.3823830747578736,
      "tokens_seen": 2640773120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020180541624874624,
      "loss": 1.2116,
      "theoretical_loss": 3.382376508166816,
      "tokens_seen": 2640838656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002017853560682046,
      "loss": 1.2709,
      "theoretical_loss": 3.3823699417843427,
      "tokens_seen": 2640904192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201765295887663,
      "loss": 1.415,
      "theoretical_loss": 3.382363375610441,
      "tokens_seen": 2640969728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020174523570712136,
      "loss": 1.297,
      "theoretical_loss": 3.3823568096451004,
      "tokens_seen": 2641035264
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020172517552657975,
      "loss": 1.305,
      "theoretical_loss": 3.3823502438883084,
      "tokens_seen": 2641100800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002017051153460381,
      "loss": 1.2318,
      "theoretical_loss": 3.3823436783400527,
      "tokens_seen": 2641166336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002016850551654965,
      "loss": 1.1838,
      "theoretical_loss": 3.3823371130003226,
      "tokens_seen": 2641231872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020166499498495486,
      "loss": 1.2678,
      "theoretical_loss": 3.3823305478691053,
      "tokens_seen": 2641297408
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020164493480441322,
      "loss": 1.2903,
      "theoretical_loss": 3.3823239829463896,
      "tokens_seen": 2641362944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002016248746238716,
      "loss": 1.239,
      "theoretical_loss": 3.382317418232164,
      "tokens_seen": 2641428480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020160481444332997,
      "loss": 1.2758,
      "theoretical_loss": 3.3823108537264157,
      "tokens_seen": 2641494016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020158475426278836,
      "loss": 1.285,
      "theoretical_loss": 3.382304289429134,
      "tokens_seen": 2641559552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020156469408224675,
      "loss": 1.275,
      "theoretical_loss": 3.3822977253403064,
      "tokens_seen": 2641625088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020154463390170514,
      "loss": 1.3044,
      "theoretical_loss": 3.382291161459921,
      "tokens_seen": 2641690624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002015245737211635,
      "loss": 1.2368,
      "theoretical_loss": 3.3822845977879665,
      "tokens_seen": 2641756160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002015045135406219,
      "loss": 1.3448,
      "theoretical_loss": 3.382278034324431,
      "tokens_seen": 2641821696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020148445336008026,
      "loss": 1.3418,
      "theoretical_loss": 3.3822714710693025,
      "tokens_seen": 2641887232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020146439317953862,
      "loss": 1.292,
      "theoretical_loss": 3.3822649080225693,
      "tokens_seen": 2641952768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201444332998997,
      "loss": 1.2346,
      "theoretical_loss": 3.38225834518422,
      "tokens_seen": 2642018304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020142427281845537,
      "loss": 1.1476,
      "theoretical_loss": 3.3822517825542424,
      "tokens_seen": 2642083840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020140421263791376,
      "loss": 1.1398,
      "theoretical_loss": 3.3822452201326247,
      "tokens_seen": 2642149376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020138415245737212,
      "loss": 1.2182,
      "theoretical_loss": 3.382238657919355,
      "tokens_seen": 2642214912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002013640922768305,
      "loss": 1.1945,
      "theoretical_loss": 3.3822320959144223,
      "tokens_seen": 2642280448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020134403209628887,
      "loss": 1.2641,
      "theoretical_loss": 3.3822255341178136,
      "tokens_seen": 2642345984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020132397191574726,
      "loss": 1.3163,
      "theoretical_loss": 3.382218972529518,
      "tokens_seen": 2642411520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020130391173520562,
      "loss": 1.2702,
      "theoretical_loss": 3.3822124111495233,
      "tokens_seen": 2642477056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020128385155466399,
      "loss": 1.2567,
      "theoretical_loss": 3.3822058499778183,
      "tokens_seen": 2642542592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020126379137412238,
      "loss": 1.2038,
      "theoretical_loss": 3.3821992890143906,
      "tokens_seen": 2642608128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020124373119358074,
      "loss": 1.276,
      "theoretical_loss": 3.382192728259229,
      "tokens_seen": 2642673664
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020122367101303913,
      "loss": 1.2247,
      "theoretical_loss": 3.3821861677123204,
      "tokens_seen": 2642739200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002012036108324975,
      "loss": 1.2217,
      "theoretical_loss": 3.382179607373655,
      "tokens_seen": 2642804736
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 960167,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1384999752044678,
      "objective/train/theoretical_loss": 3.38217304724322,
      "objective/train/tokens_used": 1013329376,
      "theoretical_loss": 3.38217304724322,
      "tokens_seen": 2642870272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020118355065195588,
      "loss": 1.2681,
      "theoretical_loss": 3.38217304724322,
      "tokens_seen": 2642870272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020116349047141424,
      "loss": 1.3025,
      "theoretical_loss": 3.382166487321003,
      "tokens_seen": 2642935808
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020114343029087263,
      "loss": 1.2767,
      "theoretical_loss": 3.382159927606993,
      "tokens_seen": 2643001344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201123370110331,
      "loss": 1.2844,
      "theoretical_loss": 3.3821533681011786,
      "tokens_seen": 2643066880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020110330992978938,
      "loss": 1.2512,
      "theoretical_loss": 3.3821468088035473,
      "tokens_seen": 2643132416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020108324974924774,
      "loss": 1.209,
      "theoretical_loss": 3.3821402497140873,
      "tokens_seen": 2643197952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002010631895687061,
      "loss": 1.2558,
      "theoretical_loss": 3.3821336908327875,
      "tokens_seen": 2643263488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002010431293881645,
      "loss": 1.323,
      "theoretical_loss": 3.382127132159636,
      "tokens_seen": 2643329024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020102306920762286,
      "loss": 1.265,
      "theoretical_loss": 3.38212057369462,
      "tokens_seen": 2643394560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020100300902708125,
      "loss": 1.2609,
      "theoretical_loss": 3.382114015437729,
      "tokens_seen": 2643460096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002009829488465396,
      "loss": 1.3538,
      "theoretical_loss": 3.3821074573889502,
      "tokens_seen": 2643525632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200962888665998,
      "loss": 1.2228,
      "theoretical_loss": 3.382100899548273,
      "tokens_seen": 2643591168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020094282848545636,
      "loss": 1.2541,
      "theoretical_loss": 3.382094341915685,
      "tokens_seen": 2643656704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020092276830491475,
      "loss": 1.3607,
      "theoretical_loss": 3.382087784491174,
      "tokens_seen": 2643722240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002009027081243731,
      "loss": 1.2506,
      "theoretical_loss": 3.382081227274729,
      "tokens_seen": 2643787776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020088264794383147,
      "loss": 1.3474,
      "theoretical_loss": 3.382074670266338,
      "tokens_seen": 2643853312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020086258776328986,
      "loss": 1.2217,
      "theoretical_loss": 3.3820681134659893,
      "tokens_seen": 2643918848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020084252758274823,
      "loss": 1.1686,
      "theoretical_loss": 3.3820615568736705,
      "tokens_seen": 2643984384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020082246740220664,
      "loss": 1.3207,
      "theoretical_loss": 3.382055000489371,
      "tokens_seen": 2644049920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200802407221665,
      "loss": 1.3747,
      "theoretical_loss": 3.382048444313078,
      "tokens_seen": 2644115456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007823470411234,
      "loss": 1.259,
      "theoretical_loss": 3.38204188834478,
      "tokens_seen": 2644180992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020076228686058176,
      "loss": 1.2697,
      "theoretical_loss": 3.382035332584466,
      "tokens_seen": 2644246528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020074222668004015,
      "loss": 1.2729,
      "theoretical_loss": 3.382028777032123,
      "tokens_seen": 2644312064
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007221664994985,
      "loss": 1.206,
      "theoretical_loss": 3.3820222216877402,
      "tokens_seen": 2644377600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007021063189569,
      "loss": 1.3485,
      "theoretical_loss": 3.3820156665513057,
      "tokens_seen": 2644443136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020068204613841526,
      "loss": 1.3715,
      "theoretical_loss": 3.3820091116228075,
      "tokens_seen": 2644508672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020066198595787362,
      "loss": 1.2732,
      "theoretical_loss": 3.3820025569022336,
      "tokens_seen": 2644574208
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200641925777332,
      "loss": 1.248,
      "theoretical_loss": 3.381996002389573,
      "tokens_seen": 2644639744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020062186559679037,
      "loss": 1.2418,
      "theoretical_loss": 3.3819894480848136,
      "tokens_seen": 2644705280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020060180541624876,
      "loss": 1.3882,
      "theoretical_loss": 3.381982893987943,
      "tokens_seen": 2644770816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020058174523570713,
      "loss": 1.263,
      "theoretical_loss": 3.3819763400989507,
      "tokens_seen": 2644836352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020056168505516551,
      "loss": 1.2397,
      "theoretical_loss": 3.381969786417824,
      "tokens_seen": 2644901888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020054162487462388,
      "loss": 1.2616,
      "theoretical_loss": 3.381963232944552,
      "tokens_seen": 2644967424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020052156469408227,
      "loss": 1.2899,
      "theoretical_loss": 3.3819566796791216,
      "tokens_seen": 2645032960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020050150451354063,
      "loss": 1.2283,
      "theoretical_loss": 3.381950126621523,
      "tokens_seen": 2645098496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200481444332999,
      "loss": 1.2003,
      "theoretical_loss": 3.3819435737717423,
      "tokens_seen": 2645164032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020046138415245738,
      "loss": 1.4152,
      "theoretical_loss": 3.3819370211297692,
      "tokens_seen": 2645229568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020044132397191574,
      "loss": 1.2271,
      "theoretical_loss": 3.3819304686955913,
      "tokens_seen": 2645295104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020042126379137413,
      "loss": 1.3159,
      "theoretical_loss": 3.381923916469198,
      "tokens_seen": 2645360640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002004012036108325,
      "loss": 1.2204,
      "theoretical_loss": 3.381917364450576,
      "tokens_seen": 2645426176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020038114343029088,
      "loss": 1.2656,
      "theoretical_loss": 3.381910812639714,
      "tokens_seen": 2645491712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020036108324974925,
      "loss": 1.1801,
      "theoretical_loss": 3.381904261036601,
      "tokens_seen": 2645557248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020034102306920763,
      "loss": 1.2248,
      "theoretical_loss": 3.3818977096412244,
      "tokens_seen": 2645622784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200320962888666,
      "loss": 1.2944,
      "theoretical_loss": 3.3818911584535734,
      "tokens_seen": 2645688320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020030090270812436,
      "loss": 1.207,
      "theoretical_loss": 3.3818846074736357,
      "tokens_seen": 2645753856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020028084252758275,
      "loss": 1.2354,
      "theoretical_loss": 3.381878056701399,
      "tokens_seen": 2645819392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002002607823470411,
      "loss": 1.2071,
      "theoretical_loss": 3.3818715061368527,
      "tokens_seen": 2645884928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002002407221664995,
      "loss": 1.2256,
      "theoretical_loss": 3.3818649557799847,
      "tokens_seen": 2645950464
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020022066198595786,
      "loss": 1.1436,
      "theoretical_loss": 3.3818584056307825,
      "tokens_seen": 2646016000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020020060180541625,
      "loss": 1.2193,
      "theoretical_loss": 3.3818518556892356,
      "tokens_seen": 2646081536
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 961281,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1973772048950195,
      "objective/train/theoretical_loss": 3.3818453059553315,
      "objective/train/tokens_used": 1016606176,
      "theoretical_loss": 3.3818453059553315,
      "tokens_seen": 2646147072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020018054162487461,
      "loss": 1.2111,
      "theoretical_loss": 3.3818453059553315,
      "tokens_seen": 2646147072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200160481444333,
      "loss": 1.1561,
      "theoretical_loss": 3.381838756429058,
      "tokens_seen": 2646212608
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020014042126379137,
      "loss": 1.2872,
      "theoretical_loss": 3.3818322071104046,
      "tokens_seen": 2646278144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020012036108324973,
      "loss": 1.3475,
      "theoretical_loss": 3.381825657999359,
      "tokens_seen": 2646343680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020010030090270812,
      "loss": 1.2583,
      "theoretical_loss": 3.3818191090959093,
      "tokens_seen": 2646409216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020008024072216648,
      "loss": 1.2005,
      "theoretical_loss": 3.381812560400044,
      "tokens_seen": 2646474752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002000601805416249,
      "loss": 1.2885,
      "theoretical_loss": 3.3818060119117517,
      "tokens_seen": 2646540288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020004012036108326,
      "loss": 1.2989,
      "theoretical_loss": 3.38179946363102,
      "tokens_seen": 2646605824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020002006018054165,
      "loss": 1.2161,
      "theoretical_loss": 3.3817929155578375,
      "tokens_seen": 2646671360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 1.2512,
      "theoretical_loss": 3.381786367692192,
      "tokens_seen": 2646736896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001999799398194584,
      "loss": 1.2643,
      "theoretical_loss": 3.381779820034073,
      "tokens_seen": 2646802432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019995987963891676,
      "loss": 1.3153,
      "theoretical_loss": 3.3817732725834677,
      "tokens_seen": 2646867968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019993981945837515,
      "loss": 1.2022,
      "theoretical_loss": 3.381766725340365,
      "tokens_seen": 2646933504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001999197592778335,
      "loss": 1.3213,
      "theoretical_loss": 3.3817601783047526,
      "tokens_seen": 2646999040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019989969909729187,
      "loss": 1.1823,
      "theoretical_loss": 3.3817536314766192,
      "tokens_seen": 2647064576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019987963891675026,
      "loss": 1.2881,
      "theoretical_loss": 3.3817470848559528,
      "tokens_seen": 2647130112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019985957873620863,
      "loss": 1.2931,
      "theoretical_loss": 3.381740538442742,
      "tokens_seen": 2647195648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019983951855566702,
      "loss": 1.2177,
      "theoretical_loss": 3.381733992236975,
      "tokens_seen": 2647261184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019981945837512538,
      "loss": 1.265,
      "theoretical_loss": 3.3817274462386404,
      "tokens_seen": 2647326720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019979939819458377,
      "loss": 1.2663,
      "theoretical_loss": 3.3817209004477258,
      "tokens_seen": 2647392256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019977933801404213,
      "loss": 1.2891,
      "theoretical_loss": 3.3817143548642195,
      "tokens_seen": 2647457792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019975927783350052,
      "loss": 1.2522,
      "theoretical_loss": 3.381707809488111,
      "tokens_seen": 2647523328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019973921765295888,
      "loss": 1.2463,
      "theoretical_loss": 3.381701264319387,
      "tokens_seen": 2647588864
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019971915747241724,
      "loss": 1.2642,
      "theoretical_loss": 3.3816947193580367,
      "tokens_seen": 2647654400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019969909729187563,
      "loss": 1.272,
      "theoretical_loss": 3.381688174604048,
      "tokens_seen": 2647719936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000199679037111334,
      "loss": 1.2931,
      "theoretical_loss": 3.38168163005741,
      "tokens_seen": 2647785472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019965897693079238,
      "loss": 1.1697,
      "theoretical_loss": 3.3816750857181104,
      "tokens_seen": 2647851008
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019963891675025075,
      "loss": 1.3464,
      "theoretical_loss": 3.381668541586137,
      "tokens_seen": 2647916544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019961885656970914,
      "loss": 1.1962,
      "theoretical_loss": 3.381661997661479,
      "tokens_seen": 2647982080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001995987963891675,
      "loss": 1.3695,
      "theoretical_loss": 3.3816554539441244,
      "tokens_seen": 2648047616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001995787362086259,
      "loss": 1.3091,
      "theoretical_loss": 3.3816489104340612,
      "tokens_seen": 2648113152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019955867602808425,
      "loss": 1.1998,
      "theoretical_loss": 3.381642367131278,
      "tokens_seen": 2648178688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019953861584754264,
      "loss": 1.2733,
      "theoretical_loss": 3.381635824035763,
      "tokens_seen": 2648244224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000199518555667001,
      "loss": 1.2639,
      "theoretical_loss": 3.3816292811475046,
      "tokens_seen": 2648309760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019949849548645936,
      "loss": 1.2089,
      "theoretical_loss": 3.3816227384664908,
      "tokens_seen": 2648375296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019947843530591775,
      "loss": 1.1908,
      "theoretical_loss": 3.3816161959927107,
      "tokens_seen": 2648440832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019945837512537612,
      "loss": 1.1739,
      "theoretical_loss": 3.3816096537261515,
      "tokens_seen": 2648506368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001994383149448345,
      "loss": 1.3263,
      "theoretical_loss": 3.381603111666802,
      "tokens_seen": 2648571904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019941825476429287,
      "loss": 1.2777,
      "theoretical_loss": 3.381596569814651,
      "tokens_seen": 2648637440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019939819458375126,
      "loss": 1.1741,
      "theoretical_loss": 3.3815900281696867,
      "tokens_seen": 2648702976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019937813440320962,
      "loss": 1.162,
      "theoretical_loss": 3.3815834867318966,
      "tokens_seen": 2648768512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000199358074222668,
      "loss": 1.2242,
      "theoretical_loss": 3.3815769455012696,
      "tokens_seen": 2648834048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019933801404212637,
      "loss": 1.2056,
      "theoretical_loss": 3.381570404477794,
      "tokens_seen": 2648899584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019931795386158476,
      "loss": 1.3062,
      "theoretical_loss": 3.381563863661458,
      "tokens_seen": 2648965120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019929789368104315,
      "loss": 1.2106,
      "theoretical_loss": 3.3815573230522498,
      "tokens_seen": 2649030656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001992778335005015,
      "loss": 1.1869,
      "theoretical_loss": 3.3815507826501583,
      "tokens_seen": 2649096192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001992577733199599,
      "loss": 1.3313,
      "theoretical_loss": 3.381544242455171,
      "tokens_seen": 2649161728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019923771313941826,
      "loss": 1.1158,
      "theoretical_loss": 3.381537702467277,
      "tokens_seen": 2649227264
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019921765295887665,
      "loss": 1.2932,
      "theoretical_loss": 3.3815311626864637,
      "tokens_seen": 2649292800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019919759277833501,
      "loss": 1.2912,
      "theoretical_loss": 3.3815246231127203,
      "tokens_seen": 2649358336
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 962524,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.213138461112976,
      "objective/train/theoretical_loss": 3.381518083746035,
      "objective/train/tokens_used": 1019882976,
      "theoretical_loss": 3.381518083746035,
      "tokens_seen": 2649423872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001991775325977934,
      "loss": 1.1698,
      "theoretical_loss": 3.381518083746035,
      "tokens_seen": 2649423872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019915747241725177,
      "loss": 1.3079,
      "theoretical_loss": 3.3815115445863952,
      "tokens_seen": 2649489408
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019913741223671013,
      "loss": 1.2524,
      "theoretical_loss": 3.3815050056337905,
      "tokens_seen": 2649554944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019911735205616852,
      "loss": 1.2441,
      "theoretical_loss": 3.3814984668882087,
      "tokens_seen": 2649620480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019909729187562688,
      "loss": 1.138,
      "theoretical_loss": 3.3814919283496376,
      "tokens_seen": 2649686016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019907723169508527,
      "loss": 1.3142,
      "theoretical_loss": 3.3814853900180664,
      "tokens_seen": 2649751552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019905717151454363,
      "loss": 1.2079,
      "theoretical_loss": 3.3814788518934833,
      "tokens_seen": 2649817088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019903711133400202,
      "loss": 1.3174,
      "theoretical_loss": 3.3814723139758756,
      "tokens_seen": 2649882624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019901705115346038,
      "loss": 1.2018,
      "theoretical_loss": 3.381465776265233,
      "tokens_seen": 2649948160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019899699097291877,
      "loss": 1.2357,
      "theoretical_loss": 3.381459238761543,
      "tokens_seen": 2650013696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019897693079237713,
      "loss": 1.3107,
      "theoretical_loss": 3.381452701464794,
      "tokens_seen": 2650079232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019895687061183552,
      "loss": 1.2369,
      "theoretical_loss": 3.3814461643749745,
      "tokens_seen": 2650144768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019893681043129389,
      "loss": 1.3116,
      "theoretical_loss": 3.381439627492073,
      "tokens_seen": 2650210304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019891675025075225,
      "loss": 1.2234,
      "theoretical_loss": 3.3814330908160777,
      "tokens_seen": 2650275840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019889669007021064,
      "loss": 1.2238,
      "theoretical_loss": 3.3814265543469766,
      "tokens_seen": 2650341376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000198876629889669,
      "loss": 1.3086,
      "theoretical_loss": 3.381420018084759,
      "tokens_seen": 2650406912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001988565697091274,
      "loss": 1.2966,
      "theoretical_loss": 3.3814134820294117,
      "tokens_seen": 2650472448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019883650952858575,
      "loss": 1.187,
      "theoretical_loss": 3.381406946180924,
      "tokens_seen": 2650537984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019881644934804414,
      "loss": 1.2762,
      "theoretical_loss": 3.3814004105392845,
      "tokens_seen": 2650603520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001987963891675025,
      "loss": 1.37,
      "theoretical_loss": 3.3813938751044814,
      "tokens_seen": 2650669056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001987763289869609,
      "loss": 1.2544,
      "theoretical_loss": 3.381387339876502,
      "tokens_seen": 2650734592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019875626880641925,
      "loss": 1.3817,
      "theoretical_loss": 3.381380804855336,
      "tokens_seen": 2650800128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019873620862587762,
      "loss": 1.1681,
      "theoretical_loss": 3.381374270040971,
      "tokens_seen": 2650865664
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000198716148445336,
      "loss": 1.2488,
      "theoretical_loss": 3.381367735433396,
      "tokens_seen": 2650931200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019869608826479437,
      "loss": 1.2115,
      "theoretical_loss": 3.381361201032598,
      "tokens_seen": 2650996736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019867602808425276,
      "loss": 1.1659,
      "theoretical_loss": 3.381354666838567,
      "tokens_seen": 2651062272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019865596790371112,
      "loss": 1.4403,
      "theoretical_loss": 3.3813481328512904,
      "tokens_seen": 2651127808
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001986359077231695,
      "loss": 1.2798,
      "theoretical_loss": 3.3813415990707565,
      "tokens_seen": 2651193344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019861584754262787,
      "loss": 1.2298,
      "theoretical_loss": 3.381335065496954,
      "tokens_seen": 2651258880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019859578736208626,
      "loss": 1.2182,
      "theoretical_loss": 3.381328532129871,
      "tokens_seen": 2651324416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019857572718154462,
      "loss": 1.3436,
      "theoretical_loss": 3.3813219989694963,
      "tokens_seen": 2651389952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019855566700100304,
      "loss": 1.1535,
      "theoretical_loss": 3.381315466015818,
      "tokens_seen": 2651455488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001985356068204614,
      "loss": 1.2648,
      "theoretical_loss": 3.3813089332688238,
      "tokens_seen": 2651521024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001985356068204614,
      "loss": 1.3039,
      "theoretical_loss": 3.381302400728503,
      "tokens_seen": 2651586560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019851554663991976,
      "loss": 1.2558,
      "theoretical_loss": 3.3812958683948438,
      "tokens_seen": 2651652096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019849548645937815,
      "loss": 1.1892,
      "theoretical_loss": 3.381289336267834,
      "tokens_seen": 2651717632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019847542627883652,
      "loss": 1.3813,
      "theoretical_loss": 3.381282804347462,
      "tokens_seen": 2651783168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001984553660982949,
      "loss": 1.275,
      "theoretical_loss": 3.3812762726337167,
      "tokens_seen": 2651848704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019843530591775327,
      "loss": 1.3222,
      "theoretical_loss": 3.3812697411265864,
      "tokens_seen": 2651914240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019841524573721166,
      "loss": 1.1559,
      "theoretical_loss": 3.381263209826059,
      "tokens_seen": 2651979776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019839518555667002,
      "loss": 1.3131,
      "theoretical_loss": 3.3812566787321234,
      "tokens_seen": 2652045312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001983751253761284,
      "loss": 1.3973,
      "theoretical_loss": 3.381250147844767,
      "tokens_seen": 2652110848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019835506519558677,
      "loss": 1.2412,
      "theoretical_loss": 3.38124361716398,
      "tokens_seen": 2652176384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019833500501504513,
      "loss": 1.2847,
      "theoretical_loss": 3.3812370866897488,
      "tokens_seen": 2652241920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019831494483450352,
      "loss": 1.1661,
      "theoretical_loss": 3.3812305564220626,
      "tokens_seen": 2652307456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019829488465396188,
      "loss": 1.3592,
      "theoretical_loss": 3.38122402636091,
      "tokens_seen": 2652372992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019827482447342027,
      "loss": 1.2045,
      "theoretical_loss": 3.3812174965062787,
      "tokens_seen": 2652438528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019825476429287864,
      "loss": 1.2422,
      "theoretical_loss": 3.381210966858158,
      "tokens_seen": 2652504064
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019823470411233703,
      "loss": 1.1748,
      "theoretical_loss": 3.3812044374165353,
      "tokens_seen": 2652569600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001982146439317954,
      "loss": 1.2377,
      "theoretical_loss": 3.3811979081813996,
      "tokens_seen": 2652635136
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 963694,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.239452600479126,
      "objective/train/theoretical_loss": 3.381191379152739,
      "objective/train/tokens_used": 1023159776,
      "theoretical_loss": 3.381191379152739,
      "tokens_seen": 2652700672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019819458375125378,
      "loss": 1.3148,
      "theoretical_loss": 3.381191379152739,
      "tokens_seen": 2652700672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019817452357071214,
      "loss": 1.2615,
      "theoretical_loss": 3.381184850330542,
      "tokens_seen": 2652766208
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001981544633901705,
      "loss": 1.2956,
      "theoretical_loss": 3.381178321714797,
      "tokens_seen": 2652831744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001981344032096289,
      "loss": 1.2904,
      "theoretical_loss": 3.3811717933054917,
      "tokens_seen": 2652897280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019811434302908725,
      "loss": 1.3746,
      "theoretical_loss": 3.3811652651026156,
      "tokens_seen": 2652962816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019809428284854564,
      "loss": 1.1559,
      "theoretical_loss": 3.3811587371061567,
      "tokens_seen": 2653028352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000198074222668004,
      "loss": 1.4098,
      "theoretical_loss": 3.381152209316103,
      "tokens_seen": 2653093888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001980541624874624,
      "loss": 1.2169,
      "theoretical_loss": 3.3811456817324426,
      "tokens_seen": 2653159424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019803410230692076,
      "loss": 1.2452,
      "theoretical_loss": 3.3811391543551648,
      "tokens_seen": 2653224960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019801404212637915,
      "loss": 1.1884,
      "theoretical_loss": 3.3811326271842574,
      "tokens_seen": 2653290496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001979939819458375,
      "loss": 1.2662,
      "theoretical_loss": 3.381126100219709,
      "tokens_seen": 2653356032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001979739217652959,
      "loss": 1.3056,
      "theoretical_loss": 3.381119573461508,
      "tokens_seen": 2653421568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019795386158475426,
      "loss": 1.4675,
      "theoretical_loss": 3.3811130469096424,
      "tokens_seen": 2653487104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019793380140421262,
      "loss": 1.2084,
      "theoretical_loss": 3.381106520564101,
      "tokens_seen": 2653552640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000197913741223671,
      "loss": 1.2547,
      "theoretical_loss": 3.381099994424872,
      "tokens_seen": 2653618176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019789368104312937,
      "loss": 1.2575,
      "theoretical_loss": 3.381093468491944,
      "tokens_seen": 2653683712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019787362086258776,
      "loss": 1.2532,
      "theoretical_loss": 3.3810869427653047,
      "tokens_seen": 2653749248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019785356068204612,
      "loss": 1.3338,
      "theoretical_loss": 3.3810804172449433,
      "tokens_seen": 2653814784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019783350050150451,
      "loss": 1.2655,
      "theoretical_loss": 3.381073891930848,
      "tokens_seen": 2653880320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001978134403209629,
      "loss": 1.2915,
      "theoretical_loss": 3.381067366823007,
      "tokens_seen": 2653945856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001977933801404213,
      "loss": 1.2569,
      "theoretical_loss": 3.3810608419214088,
      "tokens_seen": 2654011392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019777331995987965,
      "loss": 1.3045,
      "theoretical_loss": 3.381054317226041,
      "tokens_seen": 2654076928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019775325977933802,
      "loss": 1.2777,
      "theoretical_loss": 3.3810477927368936,
      "tokens_seen": 2654142464
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001977331995987964,
      "loss": 1.3121,
      "theoretical_loss": 3.381041268453954,
      "tokens_seen": 2654208000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019771313941825477,
      "loss": 1.3335,
      "theoretical_loss": 3.38103474437721,
      "tokens_seen": 2654273536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019769307923771316,
      "loss": 1.263,
      "theoretical_loss": 3.3810282205066517,
      "tokens_seen": 2654339072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019767301905717152,
      "loss": 1.2393,
      "theoretical_loss": 3.381021696842266,
      "tokens_seen": 2654404608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001976529588766299,
      "loss": 1.3143,
      "theoretical_loss": 3.3810151733840415,
      "tokens_seen": 2654470144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019763289869608827,
      "loss": 1.2448,
      "theoretical_loss": 3.3810086501319674,
      "tokens_seen": 2654535680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019761283851554666,
      "loss": 1.3655,
      "theoretical_loss": 3.381002127086031,
      "tokens_seen": 2654601216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019759277833500502,
      "loss": 1.3713,
      "theoretical_loss": 3.380995604246222,
      "tokens_seen": 2654666752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019757271815446339,
      "loss": 1.345,
      "theoretical_loss": 3.3809890816125274,
      "tokens_seen": 2654732288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019755265797392178,
      "loss": 1.2556,
      "theoretical_loss": 3.3809825591849365,
      "tokens_seen": 2654797824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019753259779338014,
      "loss": 1.3493,
      "theoretical_loss": 3.3809760369634376,
      "tokens_seen": 2654863360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019751253761283853,
      "loss": 1.3571,
      "theoretical_loss": 3.380969514948019,
      "tokens_seen": 2654928896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001974924774322969,
      "loss": 1.3099,
      "theoretical_loss": 3.380962993138669,
      "tokens_seen": 2654994432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019747241725175528,
      "loss": 1.262,
      "theoretical_loss": 3.380956471535376,
      "tokens_seen": 2655059968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019745235707121364,
      "loss": 1.2547,
      "theoretical_loss": 3.3809499501381284,
      "tokens_seen": 2655125504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019743229689067203,
      "loss": 1.2093,
      "theoretical_loss": 3.3809434289469147,
      "tokens_seen": 2655191040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001974122367101304,
      "loss": 1.282,
      "theoretical_loss": 3.3809369079617233,
      "tokens_seen": 2655256576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019739217652958878,
      "loss": 1.2582,
      "theoretical_loss": 3.3809303871825427,
      "tokens_seen": 2655322112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019737211634904714,
      "loss": 1.263,
      "theoretical_loss": 3.380923866609361,
      "tokens_seen": 2655387648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001973520561685055,
      "loss": 1.2264,
      "theoretical_loss": 3.3809173462421667,
      "tokens_seen": 2655453184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001973319959879639,
      "loss": 1.3503,
      "theoretical_loss": 3.3809108260809486,
      "tokens_seen": 2655518720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019731193580742226,
      "loss": 1.2345,
      "theoretical_loss": 3.380904306125695,
      "tokens_seen": 2655584256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019729187562688065,
      "loss": 1.4298,
      "theoretical_loss": 3.380897786376394,
      "tokens_seen": 2655649792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000197271815446339,
      "loss": 1.3091,
      "theoretical_loss": 3.380891266833034,
      "tokens_seen": 2655715328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001972517552657974,
      "loss": 1.2336,
      "theoretical_loss": 3.3808847474956036,
      "tokens_seen": 2655780864
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019723169508525576,
      "loss": 1.3105,
      "theoretical_loss": 3.3808782283640912,
      "tokens_seen": 2655846400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019721163490471415,
      "loss": 1.3683,
      "theoretical_loss": 3.3808717094384853,
      "tokens_seen": 2655911936
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 964907,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.267056941986084,
      "objective/train/theoretical_loss": 3.380865190718774,
      "objective/train/tokens_used": 1026436576,
      "theoretical_loss": 3.380865190718774,
      "tokens_seen": 2655977472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001971915747241725,
      "loss": 1.2506,
      "theoretical_loss": 3.380865190718774,
      "tokens_seen": 2655977472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019717151454363087,
      "loss": 1.3529,
      "theoretical_loss": 3.380858672204946,
      "tokens_seen": 2656043008
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019715145436308926,
      "loss": 1.2973,
      "theoretical_loss": 3.38085215389699,
      "tokens_seen": 2656108544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019713139418254763,
      "loss": 1.3239,
      "theoretical_loss": 3.3808456357948935,
      "tokens_seen": 2656174080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019711133400200602,
      "loss": 1.2802,
      "theoretical_loss": 3.380839117898646,
      "tokens_seen": 2656239616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019709127382146438,
      "loss": 1.2926,
      "theoretical_loss": 3.380832600208235,
      "tokens_seen": 2656305152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001970712136409228,
      "loss": 1.297,
      "theoretical_loss": 3.380826082723649,
      "tokens_seen": 2656370688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019705115346038116,
      "loss": 1.2773,
      "theoretical_loss": 3.3808195654448774,
      "tokens_seen": 2656436224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019703109327983955,
      "loss": 1.2835,
      "theoretical_loss": 3.3808130483719077,
      "tokens_seen": 2656501760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001970110330992979,
      "loss": 1.2787,
      "theoretical_loss": 3.380806531504729,
      "tokens_seen": 2656567296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001969909729187563,
      "loss": 1.3826,
      "theoretical_loss": 3.380800014843329,
      "tokens_seen": 2656632832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019697091273821466,
      "loss": 1.2839,
      "theoretical_loss": 3.380793498387696,
      "tokens_seen": 2656698368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019695085255767302,
      "loss": 1.3612,
      "theoretical_loss": 3.3807869821378196,
      "tokens_seen": 2656763904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001969307923771314,
      "loss": 1.263,
      "theoretical_loss": 3.380780466093687,
      "tokens_seen": 2656829440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019691073219658977,
      "loss": 1.3475,
      "theoretical_loss": 3.3807739502552874,
      "tokens_seen": 2656894976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019689067201604816,
      "loss": 1.3144,
      "theoretical_loss": 3.380767434622609,
      "tokens_seen": 2656960512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019687061183550652,
      "loss": 1.1881,
      "theoretical_loss": 3.38076091919564,
      "tokens_seen": 2657026048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019685055165496491,
      "loss": 1.2939,
      "theoretical_loss": 3.380754403974369,
      "tokens_seen": 2657091584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019683049147442328,
      "loss": 1.34,
      "theoretical_loss": 3.3807478889587848,
      "tokens_seen": 2657157120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019681043129388167,
      "loss": 1.4229,
      "theoretical_loss": 3.380741374148875,
      "tokens_seen": 2657222656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019679037111334003,
      "loss": 1.2202,
      "theoretical_loss": 3.3807348595446287,
      "tokens_seen": 2657288192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001967703109327984,
      "loss": 1.2942,
      "theoretical_loss": 3.3807283451460344,
      "tokens_seen": 2657353728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019675025075225678,
      "loss": 1.2736,
      "theoretical_loss": 3.38072183095308,
      "tokens_seen": 2657419264
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019673019057171514,
      "loss": 1.3408,
      "theoretical_loss": 3.3807153169657544,
      "tokens_seen": 2657484800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019671013039117353,
      "loss": 1.3451,
      "theoretical_loss": 3.3807088031840458,
      "tokens_seen": 2657550336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001966900702106319,
      "loss": 1.2663,
      "theoretical_loss": 3.380702289607943,
      "tokens_seen": 2657615872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019667001003009028,
      "loss": 1.2517,
      "theoretical_loss": 3.3806957762374337,
      "tokens_seen": 2657681408
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019664994984954865,
      "loss": 1.3547,
      "theoretical_loss": 3.3806892630725067,
      "tokens_seen": 2657746944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019662988966900703,
      "loss": 1.3876,
      "theoretical_loss": 3.380682750113151,
      "tokens_seen": 2657812480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001966098294884654,
      "loss": 1.2723,
      "theoretical_loss": 3.380676237359354,
      "tokens_seen": 2657878016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019658976930792376,
      "loss": 1.1649,
      "theoretical_loss": 3.3806697248111055,
      "tokens_seen": 2657943552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019656970912738215,
      "loss": 1.243,
      "theoretical_loss": 3.3806632124683924,
      "tokens_seen": 2658009088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001965496489468405,
      "loss": 1.185,
      "theoretical_loss": 3.3806567003312047,
      "tokens_seen": 2658074624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001965295887662989,
      "loss": 1.1815,
      "theoretical_loss": 3.3806501883995295,
      "tokens_seen": 2658140160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019650952858575726,
      "loss": 1.321,
      "theoretical_loss": 3.3806436766733556,
      "tokens_seen": 2658205696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019648946840521565,
      "loss": 1.3233,
      "theoretical_loss": 3.380637165152672,
      "tokens_seen": 2658271232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019646940822467401,
      "loss": 1.251,
      "theoretical_loss": 3.380630653837467,
      "tokens_seen": 2658336768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001964493480441324,
      "loss": 1.3714,
      "theoretical_loss": 3.3806241427277284,
      "tokens_seen": 2658402304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019642928786359077,
      "loss": 1.3176,
      "theoretical_loss": 3.3806176318234455,
      "tokens_seen": 2658467840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019640922768304913,
      "loss": 1.2722,
      "theoretical_loss": 3.3806111211246064,
      "tokens_seen": 2658533376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019638916750250752,
      "loss": 1.2552,
      "theoretical_loss": 3.380604610631199,
      "tokens_seen": 2658598912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019636910732196588,
      "loss": 1.3284,
      "theoretical_loss": 3.380598100343213,
      "tokens_seen": 2658664448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019634904714142427,
      "loss": 1.3202,
      "theoretical_loss": 3.3805915902606354,
      "tokens_seen": 2658729984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019632898696088263,
      "loss": 1.2703,
      "theoretical_loss": 3.3805850803834554,
      "tokens_seen": 2658795520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019630892678034105,
      "loss": 1.2509,
      "theoretical_loss": 3.380578570711662,
      "tokens_seen": 2658861056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001962888665997994,
      "loss": 1.3554,
      "theoretical_loss": 3.3805720612452426,
      "tokens_seen": 2658926592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001962688064192578,
      "loss": 1.3777,
      "theoretical_loss": 3.380565551984186,
      "tokens_seen": 2658992128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019624874623871616,
      "loss": 1.3427,
      "theoretical_loss": 3.3805590429284815,
      "tokens_seen": 2659057664
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019622868605817455,
      "loss": 1.216,
      "theoretical_loss": 3.3805525340781166,
      "tokens_seen": 2659123200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001962086258776329,
      "loss": 1.2495,
      "theoretical_loss": 3.38054602543308,
      "tokens_seen": 2659188736
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 966064,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.339836597442627,
      "objective/train/theoretical_loss": 3.38053951699336,
      "objective/train/tokens_used": 1029713376,
      "theoretical_loss": 3.38053951699336,
      "tokens_seen": 2659254272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019618856569709127,
      "loss": 1.3582,
      "theoretical_loss": 3.38053951699336,
      "tokens_seen": 2659254272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019616850551654966,
      "loss": 1.1905,
      "theoretical_loss": 3.3805330087589454,
      "tokens_seen": 2659319808
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019614844533600803,
      "loss": 1.1893,
      "theoretical_loss": 3.3805265007298244,
      "tokens_seen": 2659385344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019612838515546642,
      "loss": 1.3688,
      "theoretical_loss": 3.380519992905986,
      "tokens_seen": 2659450880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019610832497492478,
      "loss": 1.3907,
      "theoretical_loss": 3.380513485287418,
      "tokens_seen": 2659516416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019608826479438317,
      "loss": 1.223,
      "theoretical_loss": 3.380506977874109,
      "tokens_seen": 2659581952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019606820461384153,
      "loss": 1.3153,
      "theoretical_loss": 3.3805004706660475,
      "tokens_seen": 2659647488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019604814443329992,
      "loss": 1.156,
      "theoretical_loss": 3.3804939636632225,
      "tokens_seen": 2659713024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019602808425275828,
      "loss": 1.2414,
      "theoretical_loss": 3.3804874568656214,
      "tokens_seen": 2659778560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019600802407221664,
      "loss": 1.2743,
      "theoretical_loss": 3.380480950273234,
      "tokens_seen": 2659844096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019598796389167503,
      "loss": 1.2559,
      "theoretical_loss": 3.3804744438860475,
      "tokens_seen": 2659909632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001959679037111334,
      "loss": 1.3221,
      "theoretical_loss": 3.380467937704051,
      "tokens_seen": 2659975168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019594784353059178,
      "loss": 1.2654,
      "theoretical_loss": 3.3804614317272335,
      "tokens_seen": 2660040704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019592778335005015,
      "loss": 1.1443,
      "theoretical_loss": 3.3804549259555823,
      "tokens_seen": 2660106240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019590772316950854,
      "loss": 1.2079,
      "theoretical_loss": 3.3804484203890866,
      "tokens_seen": 2660171776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001958876629889669,
      "loss": 1.339,
      "theoretical_loss": 3.3804419150277347,
      "tokens_seen": 2660237312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001958676028084253,
      "loss": 1.3511,
      "theoretical_loss": 3.3804354098715153,
      "tokens_seen": 2660302848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019584754262788365,
      "loss": 1.2761,
      "theoretical_loss": 3.3804289049204166,
      "tokens_seen": 2660368384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019582748244734204,
      "loss": 1.2553,
      "theoretical_loss": 3.380422400174427,
      "tokens_seen": 2660433920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001958074222668004,
      "loss": 1.3845,
      "theoretical_loss": 3.3804158956335355,
      "tokens_seen": 2660499456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019578736208625876,
      "loss": 1.2584,
      "theoretical_loss": 3.3804093912977304,
      "tokens_seen": 2660564992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019576730190571715,
      "loss": 1.148,
      "theoretical_loss": 3.3804028871669995,
      "tokens_seen": 2660630528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019574724172517552,
      "loss": 1.3555,
      "theoretical_loss": 3.380396383241332,
      "tokens_seen": 2660696064
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001957271815446339,
      "loss": 1.1839,
      "theoretical_loss": 3.3803898795207163,
      "tokens_seen": 2660761600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019570712136409227,
      "loss": 1.2782,
      "theoretical_loss": 3.380383376005141,
      "tokens_seen": 2660827136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019568706118355066,
      "loss": 1.3598,
      "theoretical_loss": 3.380376872694594,
      "tokens_seen": 2660892672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019566700100300902,
      "loss": 1.2497,
      "theoretical_loss": 3.380370369589064,
      "tokens_seen": 2660958208
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001956469408224674,
      "loss": 1.2487,
      "theoretical_loss": 3.3803638666885396,
      "tokens_seen": 2661023744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019562688064192577,
      "loss": 1.2954,
      "theoretical_loss": 3.38035736399301,
      "tokens_seen": 2661089280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019560682046138413,
      "loss": 1.2797,
      "theoretical_loss": 3.3803508615024627,
      "tokens_seen": 2661154816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019558676028084252,
      "loss": 1.2637,
      "theoretical_loss": 3.380344359216886,
      "tokens_seen": 2661220352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001955667001003009,
      "loss": 1.3957,
      "theoretical_loss": 3.3803378571362694,
      "tokens_seen": 2661285888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001955466399197593,
      "loss": 1.1916,
      "theoretical_loss": 3.380331355260601,
      "tokens_seen": 2661351424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019552657973921766,
      "loss": 1.2735,
      "theoretical_loss": 3.380324853589869,
      "tokens_seen": 2661416960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019550651955867605,
      "loss": 1.2175,
      "theoretical_loss": 3.380318352124062,
      "tokens_seen": 2661482496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019548645937813441,
      "loss": 1.2067,
      "theoretical_loss": 3.3803118508631687,
      "tokens_seen": 2661548032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001954663991975928,
      "loss": 1.2864,
      "theoretical_loss": 3.3803053498071773,
      "tokens_seen": 2661613568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019544633901705117,
      "loss": 1.3328,
      "theoretical_loss": 3.3802988489560764,
      "tokens_seen": 2661679104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019542627883650953,
      "loss": 1.1952,
      "theoretical_loss": 3.3802923483098546,
      "tokens_seen": 2661744640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019540621865596792,
      "loss": 1.3026,
      "theoretical_loss": 3.3802858478685005,
      "tokens_seen": 2661810176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019538615847542628,
      "loss": 1.2715,
      "theoretical_loss": 3.3802793476320026,
      "tokens_seen": 2661875712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019536609829488467,
      "loss": 1.2725,
      "theoretical_loss": 3.380272847600349,
      "tokens_seen": 2661941248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019534603811434303,
      "loss": 1.2106,
      "theoretical_loss": 3.380266347773529,
      "tokens_seen": 2662006784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019532597793380142,
      "loss": 1.2455,
      "theoretical_loss": 3.38025984815153,
      "tokens_seen": 2662072320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019530591775325978,
      "loss": 1.1722,
      "theoretical_loss": 3.3802533487343407,
      "tokens_seen": 2662137856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019528585757271817,
      "loss": 1.3761,
      "theoretical_loss": 3.380246849521951,
      "tokens_seen": 2662203392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019526579739217653,
      "loss": 1.2397,
      "theoretical_loss": 3.380240350514348,
      "tokens_seen": 2662268928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019524573721163492,
      "loss": 1.2704,
      "theoretical_loss": 3.38023385171152,
      "tokens_seen": 2662334464
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019522567703109329,
      "loss": 1.3713,
      "theoretical_loss": 3.3802273531134563,
      "tokens_seen": 2662400000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019520561685055165,
      "loss": 1.2335,
      "theoretical_loss": 3.3802208547201458,
      "tokens_seen": 2662465536
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 967352,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4056241512298584,
      "objective/train/theoretical_loss": 3.3802143565315763,
      "objective/train/tokens_used": 1032990176,
      "theoretical_loss": 3.3802143565315763,
      "tokens_seen": 2662531072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019518555667001004,
      "loss": 1.3436,
      "theoretical_loss": 3.3802143565315763,
      "tokens_seen": 2662531072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001951654964894684,
      "loss": 1.2872,
      "theoretical_loss": 3.380207858547736,
      "tokens_seen": 2662596608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001951454363089268,
      "loss": 1.1963,
      "theoretical_loss": 3.380201360768614,
      "tokens_seen": 2662662144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019512537612838515,
      "loss": 1.2783,
      "theoretical_loss": 3.3801948631941987,
      "tokens_seen": 2662727680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019510531594784354,
      "loss": 1.3089,
      "theoretical_loss": 3.3801883658244782,
      "tokens_seen": 2662793216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001950852557673019,
      "loss": 1.2035,
      "theoretical_loss": 3.380181868659442,
      "tokens_seen": 2662858752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001950651955867603,
      "loss": 1.2592,
      "theoretical_loss": 3.3801753716990777,
      "tokens_seen": 2662924288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019504513540621865,
      "loss": 1.2461,
      "theoretical_loss": 3.380168874943374,
      "tokens_seen": 2662989824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019502507522567702,
      "loss": 1.3318,
      "theoretical_loss": 3.3801623783923196,
      "tokens_seen": 2663055360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001950050150451354,
      "loss": 1.2828,
      "theoretical_loss": 3.380155882045903,
      "tokens_seen": 2663120896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019498495486459377,
      "loss": 1.3562,
      "theoretical_loss": 3.3801493859041125,
      "tokens_seen": 2663186432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019496489468405216,
      "loss": 1.3365,
      "theoretical_loss": 3.380142889966937,
      "tokens_seen": 2663251968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019494483450351052,
      "loss": 1.2243,
      "theoretical_loss": 3.3801363942343645,
      "tokens_seen": 2663317504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001949247743229689,
      "loss": 1.1459,
      "theoretical_loss": 3.380129898706384,
      "tokens_seen": 2663383040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019490471414242727,
      "loss": 1.2395,
      "theoretical_loss": 3.3801234033829837,
      "tokens_seen": 2663448576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019488465396188566,
      "loss": 1.2678,
      "theoretical_loss": 3.3801169082641525,
      "tokens_seen": 2663514112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019486459378134402,
      "loss": 1.2651,
      "theoretical_loss": 3.3801104133498785,
      "tokens_seen": 2663579648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019484453360080239,
      "loss": 1.2437,
      "theoretical_loss": 3.3801039186401507,
      "tokens_seen": 2663645184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019482447342026077,
      "loss": 1.1319,
      "theoretical_loss": 3.3800974241349566,
      "tokens_seen": 2663710720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019480441323971916,
      "loss": 1.4213,
      "theoretical_loss": 3.380090929834286,
      "tokens_seen": 2663776256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019478435305917755,
      "loss": 1.313,
      "theoretical_loss": 3.380084435738127,
      "tokens_seen": 2663841792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019476429287863592,
      "loss": 1.2526,
      "theoretical_loss": 3.380077941846468,
      "tokens_seen": 2663907328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001947442326980943,
      "loss": 1.3835,
      "theoretical_loss": 3.3800714481592973,
      "tokens_seen": 2663972864
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019472417251755267,
      "loss": 1.2958,
      "theoretical_loss": 3.3800649546766035,
      "tokens_seen": 2664038400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019470411233701106,
      "loss": 1.2478,
      "theoretical_loss": 3.380058461398376,
      "tokens_seen": 2664103936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019468405215646942,
      "loss": 1.2686,
      "theoretical_loss": 3.380051968324602,
      "tokens_seen": 2664169472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001946639919759278,
      "loss": 1.1908,
      "theoretical_loss": 3.380045475455271,
      "tokens_seen": 2664235008
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019464393179538617,
      "loss": 1.2109,
      "theoretical_loss": 3.380038982790371,
      "tokens_seen": 2664300544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019462387161484453,
      "loss": 1.2909,
      "theoretical_loss": 3.380032490329891,
      "tokens_seen": 2664366080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019460381143430292,
      "loss": 1.4067,
      "theoretical_loss": 3.380025998073819,
      "tokens_seen": 2664431616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019458375125376128,
      "loss": 1.2901,
      "theoretical_loss": 3.380019506022144,
      "tokens_seen": 2664497152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019456369107321967,
      "loss": 1.3075,
      "theoretical_loss": 3.380013014174854,
      "tokens_seen": 2664562688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019454363089267804,
      "loss": 1.2623,
      "theoretical_loss": 3.3800065225319385,
      "tokens_seen": 2664628224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019452357071213642,
      "loss": 1.2744,
      "theoretical_loss": 3.380000031093385,
      "tokens_seen": 2664693760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001945035105315948,
      "loss": 1.2315,
      "theoretical_loss": 3.379993539859183,
      "tokens_seen": 2664759296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019448345035105318,
      "loss": 1.3334,
      "theoretical_loss": 3.37998704882932,
      "tokens_seen": 2664824832
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019446339017051154,
      "loss": 1.3622,
      "theoretical_loss": 3.379980558003785,
      "tokens_seen": 2664890368
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001944433299899699,
      "loss": 1.3244,
      "theoretical_loss": 3.379974067382567,
      "tokens_seen": 2664955904
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001944232698094283,
      "loss": 1.3377,
      "theoretical_loss": 3.3799675769656536,
      "tokens_seen": 2665021440
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019440320962888665,
      "loss": 1.2742,
      "theoretical_loss": 3.3799610867530343,
      "tokens_seen": 2665086976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019438314944834504,
      "loss": 1.3016,
      "theoretical_loss": 3.379954596744697,
      "tokens_seen": 2665152512
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001943630892678034,
      "loss": 1.2693,
      "theoretical_loss": 3.3799481069406307,
      "tokens_seen": 2665218048
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001943430290872618,
      "loss": 1.2442,
      "theoretical_loss": 3.3799416173408234,
      "tokens_seen": 2665283584
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019432296890672016,
      "loss": 1.3383,
      "theoretical_loss": 3.3799351279452647,
      "tokens_seen": 2665349120
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019430290872617855,
      "loss": 1.3045,
      "theoretical_loss": 3.379928638753942,
      "tokens_seen": 2665414656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001942828485456369,
      "loss": 1.324,
      "theoretical_loss": 3.379922149766844,
      "tokens_seen": 2665480192
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001942627883650953,
      "loss": 1.2815,
      "theoretical_loss": 3.37991566098396,
      "tokens_seen": 2665545728
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019424272818455366,
      "loss": 1.2694,
      "theoretical_loss": 3.3799091724052777,
      "tokens_seen": 2665611264
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019422266800401202,
      "loss": 1.3483,
      "theoretical_loss": 3.379902684030786,
      "tokens_seen": 2665676800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001942026078234704,
      "loss": 1.2199,
      "theoretical_loss": 3.379896195860474,
      "tokens_seen": 2665742336
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 968517,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1904771327972412,
      "objective/train/theoretical_loss": 3.3798897078943293,
      "objective/train/tokens_used": 1036266976,
      "theoretical_loss": 3.3798897078943293,
      "tokens_seen": 2665807872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019418254764292877,
      "loss": 1.1881,
      "theoretical_loss": 3.3798897078943293,
      "tokens_seen": 2665807872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019416248746238716,
      "loss": 1.3074,
      "theoretical_loss": 3.379883220132341,
      "tokens_seen": 2665873408
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019414242728184552,
      "loss": 1.2525,
      "theoretical_loss": 3.379876732574498,
      "tokens_seen": 2665938944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019412236710130391,
      "loss": 1.2101,
      "theoretical_loss": 3.3798702452207876,
      "tokens_seen": 2666004480
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019410230692076228,
      "loss": 1.3973,
      "theoretical_loss": 3.3798637580712,
      "tokens_seen": 2666070016
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019408224674022067,
      "loss": 1.2646,
      "theoretical_loss": 3.379857271125722,
      "tokens_seen": 2666135552
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019406218655967905,
      "loss": 1.2617,
      "theoretical_loss": 3.379850784384344,
      "tokens_seen": 2666201088
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019404212637913742,
      "loss": 1.3228,
      "theoretical_loss": 3.379844297847053,
      "tokens_seen": 2666266624
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001940220661985958,
      "loss": 1.4132,
      "theoretical_loss": 3.379837811513839,
      "tokens_seen": 2666332160
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019400200601805417,
      "loss": 1.3857,
      "theoretical_loss": 3.3798313253846892,
      "tokens_seen": 2666397696
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019398194583751256,
      "loss": 1.2102,
      "theoretical_loss": 3.3798248394595927,
      "tokens_seen": 2666463232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019396188565697092,
      "loss": 1.3052,
      "theoretical_loss": 3.3798183537385382,
      "tokens_seen": 2666528768
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001939418254764293,
      "loss": 1.3127,
      "theoretical_loss": 3.3798118682215144,
      "tokens_seen": 2666594304
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019392176529588767,
      "loss": 1.3574,
      "theoretical_loss": 3.3798053829085095,
      "tokens_seen": 2666659840
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019390170511534606,
      "loss": 1.2186,
      "theoretical_loss": 3.3797988977995126,
      "tokens_seen": 2666725376
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019388164493480442,
      "loss": 1.2625,
      "theoretical_loss": 3.3797924128945116,
      "tokens_seen": 2666790912
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019386158475426279,
      "loss": 1.1881,
      "theoretical_loss": 3.379785928193495,
      "tokens_seen": 2666856448
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019384152457372117,
      "loss": 1.294,
      "theoretical_loss": 3.379779443696452,
      "tokens_seen": 2666921984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019382146439317954,
      "loss": 1.2021,
      "theoretical_loss": 3.379772959403371,
      "tokens_seen": 2666987520
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019380140421263793,
      "loss": 1.2596,
      "theoretical_loss": 3.3797664753142405,
      "tokens_seen": 2667053056
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001937813440320963,
      "loss": 1.1942,
      "theoretical_loss": 3.379759991429049,
      "tokens_seen": 2667118592
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019376128385155468,
      "loss": 1.4046,
      "theoretical_loss": 3.379753507747785,
      "tokens_seen": 2667184128
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019374122367101304,
      "loss": 1.346,
      "theoretical_loss": 3.3797470242704373,
      "tokens_seen": 2667249664
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019372116349047143,
      "loss": 1.1822,
      "theoretical_loss": 3.3797405409969947,
      "tokens_seen": 2667315200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001937011033099298,
      "loss": 1.2917,
      "theoretical_loss": 3.379734057927445,
      "tokens_seen": 2667380736
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019368104312938818,
      "loss": 1.2507,
      "theoretical_loss": 3.3797275750617772,
      "tokens_seen": 2667446272
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019366098294884654,
      "loss": 1.2957,
      "theoretical_loss": 3.37972109239998,
      "tokens_seen": 2667511808
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001936409227683049,
      "loss": 1.3227,
      "theoretical_loss": 3.379714609942042,
      "tokens_seen": 2667577344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001936208625877633,
      "loss": 1.2974,
      "theoretical_loss": 3.3797081276879517,
      "tokens_seen": 2667642880
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019360080240722166,
      "loss": 1.2986,
      "theoretical_loss": 3.379701645637698,
      "tokens_seen": 2667708416
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019358074222668005,
      "loss": 1.3659,
      "theoretical_loss": 3.3796951637912684,
      "tokens_seen": 2667773952
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001935606820461384,
      "loss": 1.3245,
      "theoretical_loss": 3.3796886821486525,
      "tokens_seen": 2667839488
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001935406218655968,
      "loss": 1.3009,
      "theoretical_loss": 3.379682200709839,
      "tokens_seen": 2667905024
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019352056168505516,
      "loss": 1.1695,
      "theoretical_loss": 3.3796757194748155,
      "tokens_seen": 2667970560
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019350050150451355,
      "loss": 1.1743,
      "theoretical_loss": 3.3796692384435714,
      "tokens_seen": 2668036096
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001934804413239719,
      "loss": 1.2831,
      "theoretical_loss": 3.379662757616095,
      "tokens_seen": 2668101632
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019346038114343027,
      "loss": 1.328,
      "theoretical_loss": 3.379656276992375,
      "tokens_seen": 2668167168
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019344032096288866,
      "loss": 1.2585,
      "theoretical_loss": 3.3796497965724,
      "tokens_seen": 2668232704
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019342026078234703,
      "loss": 1.2463,
      "theoretical_loss": 3.3796433163561588,
      "tokens_seen": 2668298240
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019340020060180542,
      "loss": 1.3478,
      "theoretical_loss": 3.3796368363436393,
      "tokens_seen": 2668363776
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019338014042126378,
      "loss": 1.2647,
      "theoretical_loss": 3.3796303565348307,
      "tokens_seen": 2668429312
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019336008024072217,
      "loss": 1.261,
      "theoretical_loss": 3.3796238769297213,
      "tokens_seen": 2668494848
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019334002006018053,
      "loss": 1.1797,
      "theoretical_loss": 3.3796173975282997,
      "tokens_seen": 2668560384
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019331995987963892,
      "loss": 1.2988,
      "theoretical_loss": 3.379610918330555,
      "tokens_seen": 2668625920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001932998996990973,
      "loss": 1.321,
      "theoretical_loss": 3.3796044393364753,
      "tokens_seen": 2668691456
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001932798395185557,
      "loss": 1.3668,
      "theoretical_loss": 3.3795979605460493,
      "tokens_seen": 2668756992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019325977933801406,
      "loss": 1.1865,
      "theoretical_loss": 3.3795914819592654,
      "tokens_seen": 2668822528
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019323971915747242,
      "loss": 1.151,
      "theoretical_loss": 3.3795850035761124,
      "tokens_seen": 2668888064
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001932196589769308,
      "loss": 1.3627,
      "theoretical_loss": 3.379578525396579,
      "tokens_seen": 2668953600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019319959879638917,
      "loss": 1.1747,
      "theoretical_loss": 3.3795720474206536,
      "tokens_seen": 2669019136
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 969669,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2841310501098633,
      "objective/train/theoretical_loss": 3.379565569648325,
      "objective/train/tokens_used": 1039543776,
      "theoretical_loss": 3.379565569648325,
      "tokens_seen": 2669084672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019317953861584756,
      "loss": 1.2036,
      "theoretical_loss": 3.379565569648325,
      "tokens_seen": 2669084672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019315947843530592,
      "loss": 1.1719,
      "theoretical_loss": 3.3795590920795817,
      "tokens_seen": 2669150208
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019313941825476431,
      "loss": 1.3,
      "theoretical_loss": 3.379552614714412,
      "tokens_seen": 2669215744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019311935807422268,
      "loss": 1.2812,
      "theoretical_loss": 3.379546137552805,
      "tokens_seen": 2669281280
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019309929789368107,
      "loss": 1.2456,
      "theoretical_loss": 3.3795396605947494,
      "tokens_seen": 2669346816
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019307923771313943,
      "loss": 1.2806,
      "theoretical_loss": 3.379533183840233,
      "tokens_seen": 2669412352
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001930591775325978,
      "loss": 1.2262,
      "theoretical_loss": 3.379526707289245,
      "tokens_seen": 2669477888
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019303911735205618,
      "loss": 1.3211,
      "theoretical_loss": 3.3795202309417745,
      "tokens_seen": 2669543424
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019301905717151454,
      "loss": 1.2184,
      "theoretical_loss": 3.3795137547978094,
      "tokens_seen": 2669608960
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019299899699097293,
      "loss": 1.2998,
      "theoretical_loss": 3.379507278857338,
      "tokens_seen": 2669674496
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001929789368104313,
      "loss": 1.3029,
      "theoretical_loss": 3.3795008031203495,
      "tokens_seen": 2669740032
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019295887662988968,
      "loss": 1.155,
      "theoretical_loss": 3.3794943275868325,
      "tokens_seen": 2669805568
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019293881644934804,
      "loss": 1.2162,
      "theoretical_loss": 3.3794878522567755,
      "tokens_seen": 2669871104
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019291875626880643,
      "loss": 1.4354,
      "theoretical_loss": 3.379481377130167,
      "tokens_seen": 2669936640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001928986960882648,
      "loss": 1.3583,
      "theoretical_loss": 3.3794749022069954,
      "tokens_seen": 2670002176
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019287863590772316,
      "loss": 1.1993,
      "theoretical_loss": 3.37946842748725,
      "tokens_seen": 2670067712
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019285857572718155,
      "loss": 1.3174,
      "theoretical_loss": 3.379461952970919,
      "tokens_seen": 2670133248
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001928385155466399,
      "loss": 1.3602,
      "theoretical_loss": 3.3794554786579907,
      "tokens_seen": 2670198784
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001928184553660983,
      "loss": 1.274,
      "theoretical_loss": 3.3794490045484546,
      "tokens_seen": 2670264320
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019279839518555666,
      "loss": 1.2479,
      "theoretical_loss": 3.3794425306422986,
      "tokens_seen": 2670329856
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019277833500501505,
      "loss": 1.1793,
      "theoretical_loss": 3.3794360569395114,
      "tokens_seen": 2670395392
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001927582748244734,
      "loss": 1.3215,
      "theoretical_loss": 3.379429583440082,
      "tokens_seen": 2670460928
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001927382146439318,
      "loss": 1.2288,
      "theoretical_loss": 3.3794231101439984,
      "tokens_seen": 2670526464
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019271815446339017,
      "loss": 1.2267,
      "theoretical_loss": 3.37941663705125,
      "tokens_seen": 2670592000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019269809428284853,
      "loss": 1.2771,
      "theoretical_loss": 3.3794101641618246,
      "tokens_seen": 2670657536
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019267803410230692,
      "loss": 1.2863,
      "theoretical_loss": 3.3794036914757113,
      "tokens_seen": 2670723072
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019265797392176528,
      "loss": 1.3475,
      "theoretical_loss": 3.379397218992899,
      "tokens_seen": 2670788608
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019263791374122367,
      "loss": 1.2057,
      "theoretical_loss": 3.3793907467133755,
      "tokens_seen": 2670854144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019261785356068203,
      "loss": 1.2344,
      "theoretical_loss": 3.3793842746371303,
      "tokens_seen": 2670919680
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019259779338014042,
      "loss": 1.275,
      "theoretical_loss": 3.3793778027641515,
      "tokens_seen": 2670985216
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019257773319959878,
      "loss": 1.1845,
      "theoretical_loss": 3.3793713310944278,
      "tokens_seen": 2671050752
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001925576730190572,
      "loss": 1.3653,
      "theoretical_loss": 3.3793648596279477,
      "tokens_seen": 2671116288
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019253761283851556,
      "loss": 1.2878,
      "theoretical_loss": 3.3793583883647003,
      "tokens_seen": 2671181824
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019251755265797395,
      "loss": 1.3495,
      "theoretical_loss": 3.3793519173046738,
      "tokens_seen": 2671247360
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001924974924774323,
      "loss": 1.2718,
      "theoretical_loss": 3.379345446447857,
      "tokens_seen": 2671312896
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019247743229689067,
      "loss": 1.2726,
      "theoretical_loss": 3.379338975794239,
      "tokens_seen": 2671378432
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019245737211634906,
      "loss": 1.3146,
      "theoretical_loss": 3.3793325053438075,
      "tokens_seen": 2671443968
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019243731193580743,
      "loss": 1.3239,
      "theoretical_loss": 3.3793260350965513,
      "tokens_seen": 2671509504
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019241725175526582,
      "loss": 1.3136,
      "theoretical_loss": 3.3793195650524597,
      "tokens_seen": 2671575040
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019239719157472418,
      "loss": 1.2227,
      "theoretical_loss": 3.379313095211521,
      "tokens_seen": 2671640576
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019237713139418257,
      "loss": 1.2616,
      "theoretical_loss": 3.3793066255737236,
      "tokens_seen": 2671706112
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019235707121364093,
      "loss": 1.2453,
      "theoretical_loss": 3.3793001561390565,
      "tokens_seen": 2671771648
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019233701103309932,
      "loss": 1.3529,
      "theoretical_loss": 3.3792936869075083,
      "tokens_seen": 2671837184
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019231695085255768,
      "loss": 1.2903,
      "theoretical_loss": 3.3792872178790674,
      "tokens_seen": 2671902720
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019229689067201604,
      "loss": 1.2907,
      "theoretical_loss": 3.379280749053722,
      "tokens_seen": 2671968256
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019227683049147443,
      "loss": 1.2259,
      "theoretical_loss": 3.379274280431462,
      "tokens_seen": 2672033792
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001922567703109328,
      "loss": 1.2363,
      "theoretical_loss": 3.379267812012275,
      "tokens_seen": 2672099328
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019223671013039118,
      "loss": 1.3612,
      "theoretical_loss": 3.3792613437961503,
      "tokens_seen": 2672164864
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019221664994984955,
      "loss": 1.2963,
      "theoretical_loss": 3.3792548757830763,
      "tokens_seen": 2672230400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019219658976930794,
      "loss": 1.2402,
      "theoretical_loss": 3.3792484079730416,
      "tokens_seen": 2672295936
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 970931,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2702627182006836,
      "objective/train/theoretical_loss": 3.3792419403660343,
      "objective/train/tokens_used": 1042820576,
      "theoretical_loss": 3.3792419403660343,
      "tokens_seen": 2672361472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001921765295887663,
      "loss": 1.2611,
      "theoretical_loss": 3.3792419403660343,
      "tokens_seen": 2672361472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001921564694082247,
      "loss": 1.291,
      "theoretical_loss": 3.379235472962044,
      "tokens_seen": 2672427008
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019213640922768305,
      "loss": 1.3085,
      "theoretical_loss": 3.379229005761059,
      "tokens_seen": 2672492544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019211634904714144,
      "loss": 1.201,
      "theoretical_loss": 3.3792225387630674,
      "tokens_seen": 2672558080
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001920962888665998,
      "loss": 1.3575,
      "theoretical_loss": 3.3792160719680586,
      "tokens_seen": 2672623616
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019207622868605816,
      "loss": 1.3064,
      "theoretical_loss": 3.379209605376021,
      "tokens_seen": 2672689152
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019205616850551655,
      "loss": 1.2878,
      "theoretical_loss": 3.3792031389869432,
      "tokens_seen": 2672754688
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019203610832497491,
      "loss": 1.2692,
      "theoretical_loss": 3.379196672800814,
      "tokens_seen": 2672820224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001920160481444333,
      "loss": 1.2673,
      "theoretical_loss": 3.379190206817622,
      "tokens_seen": 2672885760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019199598796389167,
      "loss": 1.2377,
      "theoretical_loss": 3.3791837410373553,
      "tokens_seen": 2672951296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019197592778335006,
      "loss": 1.2548,
      "theoretical_loss": 3.3791772754600036,
      "tokens_seen": 2673016832
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019195586760280842,
      "loss": 1.4395,
      "theoretical_loss": 3.379170810085555,
      "tokens_seen": 2673082368
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001919358074222668,
      "loss": 1.2573,
      "theoretical_loss": 3.379164344913998,
      "tokens_seen": 2673147904
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019191574724172517,
      "loss": 1.2524,
      "theoretical_loss": 3.3791578799453212,
      "tokens_seen": 2673213440
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019189568706118353,
      "loss": 1.28,
      "theoretical_loss": 3.3791514151795137,
      "tokens_seen": 2673278976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019187562688064192,
      "loss": 1.3391,
      "theoretical_loss": 3.3791449506165643,
      "tokens_seen": 2673344512
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019185556670010028,
      "loss": 1.3871,
      "theoretical_loss": 3.3791384862564606,
      "tokens_seen": 2673410048
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019183550651955867,
      "loss": 1.3004,
      "theoretical_loss": 3.3791320220991925,
      "tokens_seen": 2673475584
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019181544633901706,
      "loss": 1.3787,
      "theoretical_loss": 3.379125558144748,
      "tokens_seen": 2673541120
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019179538615847545,
      "loss": 1.29,
      "theoretical_loss": 3.3791190943931158,
      "tokens_seen": 2673606656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019177532597793381,
      "loss": 1.3001,
      "theoretical_loss": 3.3791126308442845,
      "tokens_seen": 2673672192
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001917552657973922,
      "loss": 1.2542,
      "theoretical_loss": 3.3791061674982434,
      "tokens_seen": 2673737728
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019173520561685057,
      "loss": 1.322,
      "theoretical_loss": 3.3790997043549806,
      "tokens_seen": 2673803264
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019171514543630893,
      "loss": 1.3317,
      "theoretical_loss": 3.3790932414144845,
      "tokens_seen": 2673868800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019169508525576732,
      "loss": 1.2408,
      "theoretical_loss": 3.3790867786767445,
      "tokens_seen": 2673934336
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019167502507522568,
      "loss": 1.3049,
      "theoretical_loss": 3.379080316141749,
      "tokens_seen": 2673999872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019165496489468407,
      "loss": 1.2413,
      "theoretical_loss": 3.379073853809486,
      "tokens_seen": 2674065408
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019163490471414243,
      "loss": 1.329,
      "theoretical_loss": 3.3790673916799454,
      "tokens_seen": 2674130944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019161484453360082,
      "loss": 1.3022,
      "theoretical_loss": 3.379060929753115,
      "tokens_seen": 2674196480
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019159478435305918,
      "loss": 1.3978,
      "theoretical_loss": 3.3790544680289836,
      "tokens_seen": 2674262016
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019157472417251757,
      "loss": 1.2625,
      "theoretical_loss": 3.37904800650754,
      "tokens_seen": 2674327552
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019155466399197593,
      "loss": 1.3073,
      "theoretical_loss": 3.3790415451887728,
      "tokens_seen": 2674393088
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019153460381143432,
      "loss": 1.305,
      "theoretical_loss": 3.3790350840726706,
      "tokens_seen": 2674458624
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019151454363089269,
      "loss": 1.241,
      "theoretical_loss": 3.3790286231592224,
      "tokens_seen": 2674524160
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019149448345035105,
      "loss": 1.3244,
      "theoretical_loss": 3.3790221624484165,
      "tokens_seen": 2674589696
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019147442326980944,
      "loss": 1.2436,
      "theoretical_loss": 3.379015701940242,
      "tokens_seen": 2674655232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001914543630892678,
      "loss": 1.3866,
      "theoretical_loss": 3.3790092416346873,
      "tokens_seen": 2674720768
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001914343029087262,
      "loss": 1.1416,
      "theoretical_loss": 3.379002781531741,
      "tokens_seen": 2674786304
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019141424272818455,
      "loss": 1.2542,
      "theoretical_loss": 3.378996321631392,
      "tokens_seen": 2674851840
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019139418254764294,
      "loss": 1.3604,
      "theoretical_loss": 3.3789898619336287,
      "tokens_seen": 2674917376
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001913741223671013,
      "loss": 1.2589,
      "theoretical_loss": 3.37898340243844,
      "tokens_seen": 2674982912
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001913540621865597,
      "loss": 1.1792,
      "theoretical_loss": 3.378976943145815,
      "tokens_seen": 2675048448
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019133400200601805,
      "loss": 1.1825,
      "theoretical_loss": 3.3789704840557415,
      "tokens_seen": 2675113984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019131394182547642,
      "loss": 1.1713,
      "theoretical_loss": 3.3789640251682087,
      "tokens_seen": 2675179520
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001912938816449348,
      "loss": 1.3156,
      "theoretical_loss": 3.378957566483205,
      "tokens_seen": 2675245056
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019127382146439317,
      "loss": 1.2106,
      "theoretical_loss": 3.3789511080007197,
      "tokens_seen": 2675310592
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019125376128385156,
      "loss": 1.3434,
      "theoretical_loss": 3.378944649720741,
      "tokens_seen": 2675376128
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019123370110330992,
      "loss": 1.2389,
      "theoretical_loss": 3.3789381916432575,
      "tokens_seen": 2675441664
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001912136409227683,
      "loss": 1.2723,
      "theoretical_loss": 3.3789317337682583,
      "tokens_seen": 2675507200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019119358074222667,
      "loss": 1.1906,
      "theoretical_loss": 3.378925276095732,
      "tokens_seen": 2675572736
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 972181,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.390811562538147,
      "objective/train/theoretical_loss": 3.3789188186256665,
      "objective/train/tokens_used": 1046097376,
      "theoretical_loss": 3.3789188186256665,
      "tokens_seen": 2675638272
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019117352056168506,
      "loss": 1.3517,
      "theoretical_loss": 3.3789188186256665,
      "tokens_seen": 2675638272
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019115346038114342,
      "loss": 1.2414,
      "theoretical_loss": 3.3789123613580516,
      "tokens_seen": 2675703808
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019113340020060178,
      "loss": 1.3185,
      "theoretical_loss": 3.3789059042928753,
      "tokens_seen": 2675769344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019111334002006017,
      "loss": 1.236,
      "theoretical_loss": 3.378899447430127,
      "tokens_seen": 2675834880
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019109327983951854,
      "loss": 1.3724,
      "theoretical_loss": 3.378892990769794,
      "tokens_seen": 2675900416
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019107321965897693,
      "loss": 1.3447,
      "theoretical_loss": 3.378886534311867,
      "tokens_seen": 2675965952
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019105315947843532,
      "loss": 1.3479,
      "theoretical_loss": 3.378880078056333,
      "tokens_seen": 2676031488
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001910330992978937,
      "loss": 1.2817,
      "theoretical_loss": 3.378873622003182,
      "tokens_seen": 2676097024
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019101303911735207,
      "loss": 1.2947,
      "theoretical_loss": 3.3788671661524012,
      "tokens_seen": 2676162560
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019099297893681046,
      "loss": 1.347,
      "theoretical_loss": 3.3788607105039805,
      "tokens_seen": 2676228096
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019097291875626882,
      "loss": 1.3454,
      "theoretical_loss": 3.378854255057908,
      "tokens_seen": 2676293632
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001909528585757272,
      "loss": 1.3315,
      "theoretical_loss": 3.378847799814173,
      "tokens_seen": 2676359168
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019093279839518557,
      "loss": 1.1803,
      "theoretical_loss": 3.3788413447727637,
      "tokens_seen": 2676424704
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019091273821464393,
      "loss": 1.2725,
      "theoretical_loss": 3.378834889933669,
      "tokens_seen": 2676490240
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019089267803410232,
      "loss": 1.3039,
      "theoretical_loss": 3.378828435296877,
      "tokens_seen": 2676555776
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019087261785356068,
      "loss": 1.2414,
      "theoretical_loss": 3.378821980862378,
      "tokens_seen": 2676621312
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019085255767301907,
      "loss": 1.2447,
      "theoretical_loss": 3.3788155266301585,
      "tokens_seen": 2676686848
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019083249749247744,
      "loss": 1.3125,
      "theoretical_loss": 3.378809072600209,
      "tokens_seen": 2676752384
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019081243731193582,
      "loss": 1.3048,
      "theoretical_loss": 3.3788026187725175,
      "tokens_seen": 2676817920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001907923771313942,
      "loss": 1.2108,
      "theoretical_loss": 3.378796165147073,
      "tokens_seen": 2676883456
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019077231695085258,
      "loss": 1.3193,
      "theoretical_loss": 3.3787897117238637,
      "tokens_seen": 2676948992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019075225677031094,
      "loss": 1.2521,
      "theoretical_loss": 3.378783258502879,
      "tokens_seen": 2677014528
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001907321965897693,
      "loss": 1.3527,
      "theoretical_loss": 3.3787768054841067,
      "tokens_seen": 2677080064
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001907121364092277,
      "loss": 1.2616,
      "theoretical_loss": 3.3787703526675363,
      "tokens_seen": 2677145600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019069207622868605,
      "loss": 1.2198,
      "theoretical_loss": 3.378763900053156,
      "tokens_seen": 2677211136
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019067201604814444,
      "loss": 1.1887,
      "theoretical_loss": 3.3787574476409548,
      "tokens_seen": 2677276672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001906519558676028,
      "loss": 1.2067,
      "theoretical_loss": 3.378750995430922,
      "tokens_seen": 2677342208
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001906318956870612,
      "loss": 1.2238,
      "theoretical_loss": 3.378744543423045,
      "tokens_seen": 2677407744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019061183550651956,
      "loss": 1.2413,
      "theoretical_loss": 3.3787380916173135,
      "tokens_seen": 2677473280
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019059177532597794,
      "loss": 1.3954,
      "theoretical_loss": 3.378731640013716,
      "tokens_seen": 2677538816
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001905717151454363,
      "loss": 1.2245,
      "theoretical_loss": 3.3787251886122407,
      "tokens_seen": 2677604352
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019055165496489467,
      "loss": 1.1951,
      "theoretical_loss": 3.3787187374128775,
      "tokens_seen": 2677669888
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019053159478435306,
      "loss": 1.2822,
      "theoretical_loss": 3.378712286415614,
      "tokens_seen": 2677735424
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019051153460381142,
      "loss": 1.149,
      "theoretical_loss": 3.3787058356204396,
      "tokens_seen": 2677800960
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001904914744232698,
      "loss": 1.2555,
      "theoretical_loss": 3.3786993850273426,
      "tokens_seen": 2677866496
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019047141424272817,
      "loss": 1.2333,
      "theoretical_loss": 3.3786929346363115,
      "tokens_seen": 2677932032
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019045135406218656,
      "loss": 1.3094,
      "theoretical_loss": 3.3786864844473357,
      "tokens_seen": 2677997568
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019043129388164492,
      "loss": 1.2878,
      "theoretical_loss": 3.3786800344604035,
      "tokens_seen": 2678063104
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001904112337011033,
      "loss": 1.2926,
      "theoretical_loss": 3.3786735846755036,
      "tokens_seen": 2678128640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019039117352056168,
      "loss": 1.2735,
      "theoretical_loss": 3.378667135092625,
      "tokens_seen": 2678194176
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019037111334002007,
      "loss": 1.4145,
      "theoretical_loss": 3.3786606857117567,
      "tokens_seen": 2678259712
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019035105315947843,
      "loss": 1.3226,
      "theoretical_loss": 3.3786542365328867,
      "tokens_seen": 2678325248
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001903309929789368,
      "loss": 1.3814,
      "theoretical_loss": 3.3786477875560044,
      "tokens_seen": 2678390784
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001903109327983952,
      "loss": 1.2356,
      "theoretical_loss": 3.378641338781098,
      "tokens_seen": 2678456320
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019029087261785357,
      "loss": 1.3251,
      "theoretical_loss": 3.378634890208156,
      "tokens_seen": 2678521856
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019027081243731196,
      "loss": 1.3197,
      "theoretical_loss": 3.378628441837168,
      "tokens_seen": 2678587392
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019025075225677032,
      "loss": 1.2695,
      "theoretical_loss": 3.3786219936681223,
      "tokens_seen": 2678652928
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001902306920762287,
      "loss": 1.2139,
      "theoretical_loss": 3.3786155457010074,
      "tokens_seen": 2678718464
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019021063189568707,
      "loss": 1.4032,
      "theoretical_loss": 3.3786090979358123,
      "tokens_seen": 2678784000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019019057171514546,
      "loss": 1.3137,
      "theoretical_loss": 3.378602650372526,
      "tokens_seen": 2678849536
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 973380,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1937341690063477,
      "objective/train/theoretical_loss": 3.3785962030111367,
      "objective/train/tokens_used": 1049374176,
      "theoretical_loss": 3.3785962030111367,
      "tokens_seen": 2678915072
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019017051153460382,
      "loss": 1.29,
      "theoretical_loss": 3.3785962030111367,
      "tokens_seen": 2678915072
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019015045135406219,
      "loss": 1.2726,
      "theoretical_loss": 3.3785897558516336,
      "tokens_seen": 2678980608
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019013039117352057,
      "loss": 1.1968,
      "theoretical_loss": 3.378583308894005,
      "tokens_seen": 2679046144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019011033099297894,
      "loss": 1.1832,
      "theoretical_loss": 3.37857686213824,
      "tokens_seen": 2679111680
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019009027081243733,
      "loss": 1.2947,
      "theoretical_loss": 3.3785704155843272,
      "tokens_seen": 2679177216
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001900702106318957,
      "loss": 1.3409,
      "theoretical_loss": 3.378563969232255,
      "tokens_seen": 2679242752
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019005015045135408,
      "loss": 1.1919,
      "theoretical_loss": 3.378557523082013,
      "tokens_seen": 2679308288
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019003009027081244,
      "loss": 1.2641,
      "theoretical_loss": 3.378551077133589,
      "tokens_seen": 2679373824
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019001003009027083,
      "loss": 1.2909,
      "theoretical_loss": 3.3785446313869727,
      "tokens_seen": 2679439360
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001899899699097292,
      "loss": 1.2462,
      "theoretical_loss": 3.378538185842152,
      "tokens_seen": 2679504896
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018996990972918758,
      "loss": 1.2626,
      "theoretical_loss": 3.378531740499116,
      "tokens_seen": 2679570432
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018994984954864594,
      "loss": 1.2488,
      "theoretical_loss": 3.3785252953578535,
      "tokens_seen": 2679635968
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001899297893681043,
      "loss": 1.263,
      "theoretical_loss": 3.378518850418353,
      "tokens_seen": 2679701504
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001899097291875627,
      "loss": 1.3377,
      "theoretical_loss": 3.3785124056806035,
      "tokens_seen": 2679767040
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018988966900702106,
      "loss": 1.3389,
      "theoretical_loss": 3.378505961144594,
      "tokens_seen": 2679832576
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018986960882647945,
      "loss": 1.2219,
      "theoretical_loss": 3.3784995168103125,
      "tokens_seen": 2679898112
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001898495486459378,
      "loss": 1.1344,
      "theoretical_loss": 3.378493072677748,
      "tokens_seen": 2679963648
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001898294884653962,
      "loss": 1.3718,
      "theoretical_loss": 3.37848662874689,
      "tokens_seen": 2680029184
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018980942828485456,
      "loss": 1.269,
      "theoretical_loss": 3.378480185017726,
      "tokens_seen": 2680094720
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018978936810431295,
      "loss": 1.3966,
      "theoretical_loss": 3.378473741490246,
      "tokens_seen": 2680160256
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001897693079237713,
      "loss": 1.3913,
      "theoretical_loss": 3.3784672981644386,
      "tokens_seen": 2680225792
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018974924774322967,
      "loss": 1.2809,
      "theoretical_loss": 3.3784608550402915,
      "tokens_seen": 2680291328
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018972918756268806,
      "loss": 1.3273,
      "theoretical_loss": 3.378454412117794,
      "tokens_seen": 2680356864
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018970912738214643,
      "loss": 1.3014,
      "theoretical_loss": 3.378447969396935,
      "tokens_seen": 2680422400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018968906720160481,
      "loss": 1.3661,
      "theoretical_loss": 3.3784415268777037,
      "tokens_seen": 2680487936
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018966900702106318,
      "loss": 1.3925,
      "theoretical_loss": 3.3784350845600883,
      "tokens_seen": 2680553472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018964894684052157,
      "loss": 1.1533,
      "theoretical_loss": 3.378428642444077,
      "tokens_seen": 2680619008
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018962888665997993,
      "loss": 1.2372,
      "theoretical_loss": 3.37842220052966,
      "tokens_seen": 2680684544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018960882647943832,
      "loss": 1.2516,
      "theoretical_loss": 3.378415758816825,
      "tokens_seen": 2680750080
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018958876629889668,
      "loss": 1.2879,
      "theoretical_loss": 3.378409317305561,
      "tokens_seen": 2680815616
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018956870611835504,
      "loss": 1.3018,
      "theoretical_loss": 3.3784028759958566,
      "tokens_seen": 2680881152
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018954864593781346,
      "loss": 1.2701,
      "theoretical_loss": 3.378396434887701,
      "tokens_seen": 2680946688
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018952858575727182,
      "loss": 1.2391,
      "theoretical_loss": 3.3783899939810826,
      "tokens_seen": 2681012224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001895085255767302,
      "loss": 1.3227,
      "theoretical_loss": 3.3783835532759903,
      "tokens_seen": 2681077760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018948846539618857,
      "loss": 1.1852,
      "theoretical_loss": 3.378377112772413,
      "tokens_seen": 2681143296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018946840521564696,
      "loss": 1.2738,
      "theoretical_loss": 3.3783706724703393,
      "tokens_seen": 2681208832
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018944834503510532,
      "loss": 1.2638,
      "theoretical_loss": 3.3783642323697585,
      "tokens_seen": 2681274368
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018942828485456371,
      "loss": 1.2643,
      "theoretical_loss": 3.3783577924706583,
      "tokens_seen": 2681339904
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018940822467402208,
      "loss": 1.2847,
      "theoretical_loss": 3.378351352773028,
      "tokens_seen": 2681405440
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018938816449348047,
      "loss": 1.2619,
      "theoretical_loss": 3.3783449132768566,
      "tokens_seen": 2681470976
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018936810431293883,
      "loss": 1.347,
      "theoretical_loss": 3.378338473982133,
      "tokens_seen": 2681536512
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001893480441323972,
      "loss": 1.3082,
      "theoretical_loss": 3.378332034888845,
      "tokens_seen": 2681602048
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018932798395185558,
      "loss": 1.2288,
      "theoretical_loss": 3.3783255959969827,
      "tokens_seen": 2681667584
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018930792377131394,
      "loss": 1.3175,
      "theoretical_loss": 3.378319157306534,
      "tokens_seen": 2681733120
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018928786359077233,
      "loss": 1.1632,
      "theoretical_loss": 3.378312718817488,
      "tokens_seen": 2681798656
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001892678034102307,
      "loss": 1.378,
      "theoretical_loss": 3.3783062805298334,
      "tokens_seen": 2681864192
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018924774322968908,
      "loss": 1.2017,
      "theoretical_loss": 3.3782998424435586,
      "tokens_seen": 2681929728
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018922768304914744,
      "loss": 1.2376,
      "theoretical_loss": 3.378293404558653,
      "tokens_seen": 2681995264
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018920762286860583,
      "loss": 1.2911,
      "theoretical_loss": 3.3782869668751054,
      "tokens_seen": 2682060800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001891875626880642,
      "loss": 1.3274,
      "theoretical_loss": 3.378280529392904,
      "tokens_seen": 2682126336
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 974516,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2817437648773193,
      "objective/train/theoretical_loss": 3.3782740921120378,
      "objective/train/tokens_used": 1052650976,
      "theoretical_loss": 3.3782740921120378,
      "tokens_seen": 2682191872
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018916750250752256,
      "loss": 1.3338,
      "theoretical_loss": 3.3782740921120378,
      "tokens_seen": 2682191872
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018914744232698095,
      "loss": 1.2995,
      "theoretical_loss": 3.378267655032496,
      "tokens_seen": 2682257408
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001891273821464393,
      "loss": 1.3301,
      "theoretical_loss": 3.378261218154267,
      "tokens_seen": 2682322944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001891073219658977,
      "loss": 1.2205,
      "theoretical_loss": 3.3782547814773394,
      "tokens_seen": 2682388480
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018908726178535606,
      "loss": 1.201,
      "theoretical_loss": 3.3782483450017025,
      "tokens_seen": 2682454016
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018906720160481445,
      "loss": 1.3596,
      "theoretical_loss": 3.3782419087273445,
      "tokens_seen": 2682519552
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001890471414242728,
      "loss": 1.3146,
      "theoretical_loss": 3.378235472654255,
      "tokens_seen": 2682585088
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001890270812437312,
      "loss": 1.1883,
      "theoretical_loss": 3.3782290367824217,
      "tokens_seen": 2682650624
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018900702106318956,
      "loss": 1.2277,
      "theoretical_loss": 3.3782226011118346,
      "tokens_seen": 2682716160
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018898696088264793,
      "loss": 1.1723,
      "theoretical_loss": 3.3782161656424816,
      "tokens_seen": 2682781696
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018896690070210632,
      "loss": 1.2408,
      "theoretical_loss": 3.3782097303743517,
      "tokens_seen": 2682847232
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018894684052156468,
      "loss": 1.3188,
      "theoretical_loss": 3.3782032953074337,
      "tokens_seen": 2682912768
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018892678034102307,
      "loss": 1.3007,
      "theoretical_loss": 3.378196860441717,
      "tokens_seen": 2682978304
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018890672016048143,
      "loss": 1.2857,
      "theoretical_loss": 3.3781904257771895,
      "tokens_seen": 2683043840
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018888665997993982,
      "loss": 1.2931,
      "theoretical_loss": 3.3781839913138403,
      "tokens_seen": 2683109376
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018886659979939818,
      "loss": 1.3403,
      "theoretical_loss": 3.378177557051658,
      "tokens_seen": 2683174912
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018884653961885657,
      "loss": 1.3411,
      "theoretical_loss": 3.378171122990632,
      "tokens_seen": 2683240448
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018882647943831493,
      "loss": 1.2592,
      "theoretical_loss": 3.3781646891307506,
      "tokens_seen": 2683305984
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018880641925777335,
      "loss": 1.2874,
      "theoretical_loss": 3.378158255472003,
      "tokens_seen": 2683371520
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001887863590772317,
      "loss": 1.2748,
      "theoretical_loss": 3.378151822014378,
      "tokens_seen": 2683437056
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018876629889669007,
      "loss": 1.2951,
      "theoretical_loss": 3.378145388757863,
      "tokens_seen": 2683502592
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018874623871614846,
      "loss": 1.3252,
      "theoretical_loss": 3.378138955702449,
      "tokens_seen": 2683568128
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018872617853560683,
      "loss": 1.2277,
      "theoretical_loss": 3.378132522848124,
      "tokens_seen": 2683633664
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018870611835506522,
      "loss": 1.2829,
      "theoretical_loss": 3.3781260901948755,
      "tokens_seen": 2683699200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018868605817452358,
      "loss": 1.3172,
      "theoretical_loss": 3.378119657742694,
      "tokens_seen": 2683764736
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018866599799398197,
      "loss": 1.2989,
      "theoretical_loss": 3.3781132254915676,
      "tokens_seen": 2683830272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018864593781344033,
      "loss": 1.3358,
      "theoretical_loss": 3.378106793441485,
      "tokens_seen": 2683895808
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018862587763289872,
      "loss": 1.1881,
      "theoretical_loss": 3.3781003615924354,
      "tokens_seen": 2683961344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018860581745235708,
      "loss": 1.2759,
      "theoretical_loss": 3.378093929944407,
      "tokens_seen": 2684026880
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018858575727181544,
      "loss": 1.3154,
      "theoretical_loss": 3.3780874984973894,
      "tokens_seen": 2684092416
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018856569709127383,
      "loss": 1.3181,
      "theoretical_loss": 3.378081067251371,
      "tokens_seen": 2684157952
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001885456369107322,
      "loss": 1.3171,
      "theoretical_loss": 3.3780746362063403,
      "tokens_seen": 2684223488
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018852557673019058,
      "loss": 1.2124,
      "theoretical_loss": 3.3780682053622866,
      "tokens_seen": 2684289024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018850551654964895,
      "loss": 1.2622,
      "theoretical_loss": 3.3780617747191988,
      "tokens_seen": 2684354560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018848545636910734,
      "loss": 1.3523,
      "theoretical_loss": 3.378055344277065,
      "tokens_seen": 2684420096
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001884653961885657,
      "loss": 1.333,
      "theoretical_loss": 3.3780489140358747,
      "tokens_seen": 2684485632
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001884453360080241,
      "loss": 1.2463,
      "theoretical_loss": 3.3780424839956167,
      "tokens_seen": 2684551168
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018842527582748245,
      "loss": 1.4101,
      "theoretical_loss": 3.3780360541562793,
      "tokens_seen": 2684616704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018840521564694084,
      "loss": 1.2575,
      "theoretical_loss": 3.3780296245178514,
      "tokens_seen": 2684682240
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001883851554663992,
      "loss": 1.3139,
      "theoretical_loss": 3.3780231950803223,
      "tokens_seen": 2684747776
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018836509528585756,
      "loss": 1.2919,
      "theoretical_loss": 3.3780167658436806,
      "tokens_seen": 2684813312
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018834503510531595,
      "loss": 1.296,
      "theoretical_loss": 3.3780103368079155,
      "tokens_seen": 2684878848
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018832497492477431,
      "loss": 1.2919,
      "theoretical_loss": 3.3780039079730146,
      "tokens_seen": 2684944384
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001883049147442327,
      "loss": 1.1995,
      "theoretical_loss": 3.377997479338968,
      "tokens_seen": 2685009920
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018828485456369107,
      "loss": 1.2028,
      "theoretical_loss": 3.377991050905764,
      "tokens_seen": 2685075456
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018826479438314946,
      "loss": 1.2834,
      "theoretical_loss": 3.3779846226733916,
      "tokens_seen": 2685140992
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018824473420260782,
      "loss": 1.2431,
      "theoretical_loss": 3.3779781946418392,
      "tokens_seen": 2685206528
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001882246740220662,
      "loss": 1.3404,
      "theoretical_loss": 3.377971766811096,
      "tokens_seen": 2685272064
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018820461384152457,
      "loss": 1.2511,
      "theoretical_loss": 3.3779653391811504,
      "tokens_seen": 2685337600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018818455366098293,
      "loss": 1.1932,
      "theoretical_loss": 3.377958911751992,
      "tokens_seen": 2685403136
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 975809,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2196687459945679,
      "objective/train/theoretical_loss": 3.3779524845236093,
      "objective/train/tokens_used": 1055927776,
      "theoretical_loss": 3.3779524845236093,
      "tokens_seen": 2685468672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018816449348044132,
      "loss": 1.1863,
      "theoretical_loss": 3.3779524845236093,
      "tokens_seen": 2685468672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018814443329989968,
      "loss": 1.257,
      "theoretical_loss": 3.3779460574959908,
      "tokens_seen": 2685534208
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018812437311935807,
      "loss": 1.2884,
      "theoretical_loss": 3.3779396306691254,
      "tokens_seen": 2685599744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018810431293881643,
      "loss": 1.3721,
      "theoretical_loss": 3.3779332040430026,
      "tokens_seen": 2685665280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018808425275827482,
      "loss": 1.3789,
      "theoretical_loss": 3.3779267776176103,
      "tokens_seen": 2685730816
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018806419257773319,
      "loss": 1.2599,
      "theoretical_loss": 3.377920351392938,
      "tokens_seen": 2685796352
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001880441323971916,
      "loss": 1.2731,
      "theoretical_loss": 3.377913925368974,
      "tokens_seen": 2685861888
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018802407221664997,
      "loss": 1.3289,
      "theoretical_loss": 3.3779074995457075,
      "tokens_seen": 2685927424
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018800401203610833,
      "loss": 1.2614,
      "theoretical_loss": 3.377901073923127,
      "tokens_seen": 2685992960
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018798395185556672,
      "loss": 1.2449,
      "theoretical_loss": 3.377894648501222,
      "tokens_seen": 2686058496
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018796389167502508,
      "loss": 1.2554,
      "theoretical_loss": 3.3778882232799807,
      "tokens_seen": 2686124032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018794383149448347,
      "loss": 1.2933,
      "theoretical_loss": 3.3778817982593923,
      "tokens_seen": 2686189568
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018792377131394183,
      "loss": 1.2699,
      "theoretical_loss": 3.3778753734394455,
      "tokens_seen": 2686255104
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018790371113340022,
      "loss": 1.3444,
      "theoretical_loss": 3.3778689488201294,
      "tokens_seen": 2686320640
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018788365095285858,
      "loss": 1.3619,
      "theoretical_loss": 3.377862524401432,
      "tokens_seen": 2686386176
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018786359077231697,
      "loss": 1.2479,
      "theoretical_loss": 3.377856100183343,
      "tokens_seen": 2686451712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018784353059177533,
      "loss": 1.253,
      "theoretical_loss": 3.377849676165851,
      "tokens_seen": 2686517248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018782347041123372,
      "loss": 1.1751,
      "theoretical_loss": 3.377843252348945,
      "tokens_seen": 2686582784
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018780341023069209,
      "loss": 1.2837,
      "theoretical_loss": 3.3778368287326135,
      "tokens_seen": 2686648320
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018778335005015045,
      "loss": 1.3053,
      "theoretical_loss": 3.3778304053168453,
      "tokens_seen": 2686713856
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018776328986960884,
      "loss": 1.207,
      "theoretical_loss": 3.3778239821016296,
      "tokens_seen": 2686779392
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001877432296890672,
      "loss": 1.3989,
      "theoretical_loss": 3.377817559086955,
      "tokens_seen": 2686844928
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001877231695085256,
      "loss": 1.319,
      "theoretical_loss": 3.3778111362728103,
      "tokens_seen": 2686910464
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018770310932798395,
      "loss": 1.2088,
      "theoretical_loss": 3.377804713659185,
      "tokens_seen": 2686976000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018768304914744234,
      "loss": 1.2801,
      "theoretical_loss": 3.377798291246067,
      "tokens_seen": 2687041536
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001876629889669007,
      "loss": 1.3427,
      "theoretical_loss": 3.3777918690334454,
      "tokens_seen": 2687107072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001876429287863591,
      "loss": 1.264,
      "theoretical_loss": 3.37778544702131,
      "tokens_seen": 2687172608
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018762286860581745,
      "loss": 1.1464,
      "theoretical_loss": 3.377779025209648,
      "tokens_seen": 2687238144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018760280842527582,
      "loss": 1.3858,
      "theoretical_loss": 3.3777726035984497,
      "tokens_seen": 2687303680
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001875827482447342,
      "loss": 1.2449,
      "theoretical_loss": 3.3777661821877034,
      "tokens_seen": 2687369216
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018756268806419257,
      "loss": 1.1721,
      "theoretical_loss": 3.3777597609773977,
      "tokens_seen": 2687434752
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018754262788365096,
      "loss": 1.4086,
      "theoretical_loss": 3.377753339967522,
      "tokens_seen": 2687500288
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018752256770310932,
      "loss": 1.2726,
      "theoretical_loss": 3.3777469191580645,
      "tokens_seen": 2687565824
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001875025075225677,
      "loss": 1.3571,
      "theoretical_loss": 3.3777404985490147,
      "tokens_seen": 2687631360
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018748244734202607,
      "loss": 1.3415,
      "theoretical_loss": 3.377734078140361,
      "tokens_seen": 2687696896
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018746238716148446,
      "loss": 1.3169,
      "theoretical_loss": 3.3777276579320925,
      "tokens_seen": 2687762432
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018744232698094282,
      "loss": 1.2626,
      "theoretical_loss": 3.377721237924198,
      "tokens_seen": 2687827968
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018742226680040118,
      "loss": 1.1911,
      "theoretical_loss": 3.3777148181166665,
      "tokens_seen": 2687893504
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018740220661985957,
      "loss": 1.1932,
      "theoretical_loss": 3.3777083985094865,
      "tokens_seen": 2687959040
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018738214643931794,
      "loss": 1.2838,
      "theoretical_loss": 3.377701979102647,
      "tokens_seen": 2688024576
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018736208625877633,
      "loss": 1.2482,
      "theoretical_loss": 3.3776955598961376,
      "tokens_seen": 2688090112
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001873420260782347,
      "loss": 1.3211,
      "theoretical_loss": 3.3776891408899457,
      "tokens_seen": 2688155648
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018732196589769308,
      "loss": 1.2152,
      "theoretical_loss": 3.377682722084061,
      "tokens_seen": 2688221184
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018730190571715147,
      "loss": 1.2591,
      "theoretical_loss": 3.3776763034784727,
      "tokens_seen": 2688286720
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018728184553660986,
      "loss": 1.3185,
      "theoretical_loss": 3.3776698850731695,
      "tokens_seen": 2688352256
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018726178535606822,
      "loss": 1.355,
      "theoretical_loss": 3.37766346686814,
      "tokens_seen": 2688417792
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001872417251755266,
      "loss": 1.3688,
      "theoretical_loss": 3.377657048863373,
      "tokens_seen": 2688483328
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018722166499498497,
      "loss": 1.2499,
      "theoretical_loss": 3.377650631058857,
      "tokens_seen": 2688548864
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018720160481444333,
      "loss": 1.2978,
      "theoretical_loss": 3.377644213454582,
      "tokens_seen": 2688614400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018718154463390172,
      "loss": 1.3456,
      "theoretical_loss": 3.377637796050536,
      "tokens_seen": 2688679936
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 976981,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.30851411819458,
      "objective/train/theoretical_loss": 3.3776313788467087,
      "objective/train/tokens_used": 1059204576,
      "theoretical_loss": 3.3776313788467087,
      "tokens_seen": 2688745472
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018716148445336008,
      "loss": 1.3226,
      "theoretical_loss": 3.3776313788467087,
      "tokens_seen": 2688745472
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018714142427281847,
      "loss": 1.2651,
      "theoretical_loss": 3.377624961843088,
      "tokens_seen": 2688811008
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018712136409227684,
      "loss": 1.1601,
      "theoretical_loss": 3.377618545039663,
      "tokens_seen": 2688876544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018710130391173522,
      "loss": 1.3678,
      "theoretical_loss": 3.377612128436423,
      "tokens_seen": 2688942080
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001870812437311936,
      "loss": 1.3709,
      "theoretical_loss": 3.3776057120333567,
      "tokens_seen": 2689007616
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018706118355065198,
      "loss": 1.2364,
      "theoretical_loss": 3.3775992958304526,
      "tokens_seen": 2689073152
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018704112337011034,
      "loss": 1.1755,
      "theoretical_loss": 3.3775928798277,
      "tokens_seen": 2689138688
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001870210631895687,
      "loss": 1.395,
      "theoretical_loss": 3.3775864640250877,
      "tokens_seen": 2689204224
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001870010030090271,
      "loss": 1.3095,
      "theoretical_loss": 3.377580048422604,
      "tokens_seen": 2689269760
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018698094282848545,
      "loss": 1.2174,
      "theoretical_loss": 3.3775736330202393,
      "tokens_seen": 2689335296
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018696088264794384,
      "loss": 1.1834,
      "theoretical_loss": 3.377567217817981,
      "tokens_seen": 2689400832
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001869408224674022,
      "loss": 1.2467,
      "theoretical_loss": 3.3775608028158186,
      "tokens_seen": 2689466368
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001869207622868606,
      "loss": 1.3043,
      "theoretical_loss": 3.3775543880137406,
      "tokens_seen": 2689531904
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018690070210631896,
      "loss": 1.2881,
      "theoretical_loss": 3.3775479734117364,
      "tokens_seen": 2689597440
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018688064192577734,
      "loss": 1.1605,
      "theoretical_loss": 3.377541559009795,
      "tokens_seen": 2689662976
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001868605817452357,
      "loss": 1.1807,
      "theoretical_loss": 3.3775351448079043,
      "tokens_seen": 2689728512
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018684052156469407,
      "loss": 1.2473,
      "theoretical_loss": 3.3775287308060538,
      "tokens_seen": 2689794048
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018682046138415246,
      "loss": 1.2656,
      "theoretical_loss": 3.3775223170042326,
      "tokens_seen": 2689859584
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018680040120361082,
      "loss": 1.3143,
      "theoretical_loss": 3.3775159034024296,
      "tokens_seen": 2689925120
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001867803410230692,
      "loss": 1.3432,
      "theoretical_loss": 3.3775094900006333,
      "tokens_seen": 2689990656
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018676028084252757,
      "loss": 1.2542,
      "theoretical_loss": 3.377503076798833,
      "tokens_seen": 2690056192
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018674022066198596,
      "loss": 1.3057,
      "theoretical_loss": 3.377496663797017,
      "tokens_seen": 2690121728
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018672016048144432,
      "loss": 1.3686,
      "theoretical_loss": 3.3774902509951747,
      "tokens_seen": 2690187264
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001867001003009027,
      "loss": 1.3084,
      "theoretical_loss": 3.377483838393295,
      "tokens_seen": 2690252800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018668004012036108,
      "loss": 1.2785,
      "theoretical_loss": 3.3774774259913665,
      "tokens_seen": 2690318336
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018665997993981946,
      "loss": 1.3614,
      "theoretical_loss": 3.377471013789378,
      "tokens_seen": 2690383872
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018663991975927783,
      "loss": 1.385,
      "theoretical_loss": 3.377464601787319,
      "tokens_seen": 2690449408
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001866198595787362,
      "loss": 1.3776,
      "theoretical_loss": 3.3774581899851777,
      "tokens_seen": 2690514944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018659979939819458,
      "loss": 1.1836,
      "theoretical_loss": 3.3774517783829436,
      "tokens_seen": 2690580480
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018657973921765294,
      "loss": 1.3361,
      "theoretical_loss": 3.3774453669806053,
      "tokens_seen": 2690646016
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018655967903711136,
      "loss": 1.2696,
      "theoretical_loss": 3.3774389557781515,
      "tokens_seen": 2690711552
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018653961885656972,
      "loss": 1.3906,
      "theoretical_loss": 3.3774325447755715,
      "tokens_seen": 2690777088
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001865195586760281,
      "loss": 1.2825,
      "theoretical_loss": 3.377426133972854,
      "tokens_seen": 2690842624
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018649949849548647,
      "loss": 1.2613,
      "theoretical_loss": 3.3774197233699876,
      "tokens_seen": 2690908160
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018647943831494486,
      "loss": 1.3839,
      "theoretical_loss": 3.377413312966962,
      "tokens_seen": 2690973696
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018645937813440322,
      "loss": 1.3058,
      "theoretical_loss": 3.3774069027637657,
      "tokens_seen": 2691039232
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018643931795386159,
      "loss": 1.2099,
      "theoretical_loss": 3.377400492760387,
      "tokens_seen": 2691104768
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018641925777331997,
      "loss": 1.2881,
      "theoretical_loss": 3.3773940829568154,
      "tokens_seen": 2691170304
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018639919759277834,
      "loss": 1.2965,
      "theoretical_loss": 3.37738767335304,
      "tokens_seen": 2691235840
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018637913741223673,
      "loss": 1.2947,
      "theoretical_loss": 3.3773812639490495,
      "tokens_seen": 2691301376
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001863590772316951,
      "loss": 1.3024,
      "theoretical_loss": 3.377374854744833,
      "tokens_seen": 2691366912
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018633901705115348,
      "loss": 1.3444,
      "theoretical_loss": 3.3773684457403785,
      "tokens_seen": 2691432448
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018631895687061184,
      "loss": 1.3721,
      "theoretical_loss": 3.377362036935676,
      "tokens_seen": 2691497984
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018629889669007023,
      "loss": 1.3998,
      "theoretical_loss": 3.3773556283307142,
      "tokens_seen": 2691563520
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001862788365095286,
      "loss": 1.2232,
      "theoretical_loss": 3.377349219925481,
      "tokens_seen": 2691629056
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018625877632898698,
      "loss": 1.2628,
      "theoretical_loss": 3.377342811719967,
      "tokens_seen": 2691694592
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018623871614844534,
      "loss": 1.3611,
      "theoretical_loss": 3.3773364037141596,
      "tokens_seen": 2691760128
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001862186559679037,
      "loss": 1.3419,
      "theoretical_loss": 3.377329995908049,
      "tokens_seen": 2691825664
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001861985957873621,
      "loss": 1.2183,
      "theoretical_loss": 3.377323588301623,
      "tokens_seen": 2691891200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018617853560682046,
      "loss": 1.2865,
      "theoretical_loss": 3.377317180894871,
      "tokens_seen": 2691956736
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 978239,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.130199909210205,
      "objective/train/theoretical_loss": 3.377310773687782,
      "objective/train/tokens_used": 1062481376,
      "theoretical_loss": 3.377310773687782,
      "tokens_seen": 2692022272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018615847542627885,
      "loss": 1.2411,
      "theoretical_loss": 3.377310773687782,
      "tokens_seen": 2692022272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001861384152457372,
      "loss": 1.3594,
      "theoretical_loss": 3.3773043666803444,
      "tokens_seen": 2692087808
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001861183550651956,
      "loss": 1.2251,
      "theoretical_loss": 3.377297959872548,
      "tokens_seen": 2692153344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018609829488465396,
      "loss": 1.2129,
      "theoretical_loss": 3.377291553264381,
      "tokens_seen": 2692218880
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018607823470411235,
      "loss": 1.3608,
      "theoretical_loss": 3.377285146855833,
      "tokens_seen": 2692284416
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001860581745235707,
      "loss": 1.2893,
      "theoretical_loss": 3.377278740646892,
      "tokens_seen": 2692349952
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018603811434302907,
      "loss": 1.3382,
      "theoretical_loss": 3.3772723346375475,
      "tokens_seen": 2692415488
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018601805416248746,
      "loss": 1.3454,
      "theoretical_loss": 3.3772659288277884,
      "tokens_seen": 2692481024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018599799398194583,
      "loss": 1.3267,
      "theoretical_loss": 3.3772595232176035,
      "tokens_seen": 2692546560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018597793380140421,
      "loss": 1.2875,
      "theoretical_loss": 3.377253117806982,
      "tokens_seen": 2692612096
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018595787362086258,
      "loss": 1.3767,
      "theoretical_loss": 3.3772467125959125,
      "tokens_seen": 2692677632
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018593781344032097,
      "loss": 1.2801,
      "theoretical_loss": 3.3772403075843838,
      "tokens_seen": 2692743168
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018591775325977933,
      "loss": 1.3334,
      "theoretical_loss": 3.377233902772385,
      "tokens_seen": 2692808704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018589769307923772,
      "loss": 1.438,
      "theoretical_loss": 3.3772274981599057,
      "tokens_seen": 2692874240
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018587763289869608,
      "loss": 1.3733,
      "theoretical_loss": 3.377221093746934,
      "tokens_seen": 2692939776
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018585757271815444,
      "loss": 1.2483,
      "theoretical_loss": 3.3772146895334587,
      "tokens_seen": 2693005312
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018583751253761283,
      "loss": 1.3103,
      "theoretical_loss": 3.377208285519469,
      "tokens_seen": 2693070848
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001858174523570712,
      "loss": 1.2367,
      "theoretical_loss": 3.3772018817049543,
      "tokens_seen": 2693136384
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001857973921765296,
      "loss": 1.3222,
      "theoretical_loss": 3.377195478089903,
      "tokens_seen": 2693201920
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018577733199598797,
      "loss": 1.3909,
      "theoretical_loss": 3.3771890746743045,
      "tokens_seen": 2693267456
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018575727181544636,
      "loss": 1.2548,
      "theoretical_loss": 3.377182671458147,
      "tokens_seen": 2693332992
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018573721163490472,
      "loss": 1.3134,
      "theoretical_loss": 3.3771762684414197,
      "tokens_seen": 2693398528
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018571715145436311,
      "loss": 1.3278,
      "theoretical_loss": 3.3771698656241123,
      "tokens_seen": 2693464064
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018569709127382148,
      "loss": 1.1707,
      "theoretical_loss": 3.377163463006213,
      "tokens_seen": 2693529600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018567703109327987,
      "loss": 1.3272,
      "theoretical_loss": 3.3771570605877104,
      "tokens_seen": 2693595136
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018565697091273823,
      "loss": 1.3815,
      "theoretical_loss": 3.377150658368594,
      "tokens_seen": 2693660672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001856369107321966,
      "loss": 1.435,
      "theoretical_loss": 3.3771442563488527,
      "tokens_seen": 2693726208
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018561685055165498,
      "loss": 1.2747,
      "theoretical_loss": 3.377137854528476,
      "tokens_seen": 2693791744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018559679037111334,
      "loss": 1.2391,
      "theoretical_loss": 3.3771314529074514,
      "tokens_seen": 2693857280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018557673019057173,
      "loss": 1.2485,
      "theoretical_loss": 3.377125051485769,
      "tokens_seen": 2693922816
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001855566700100301,
      "loss": 1.3246,
      "theoretical_loss": 3.377118650263417,
      "tokens_seen": 2693988352
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018553660982948848,
      "loss": 1.3096,
      "theoretical_loss": 3.377112249240385,
      "tokens_seen": 2694053888
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018551654964894684,
      "loss": 1.2391,
      "theoretical_loss": 3.377105848416662,
      "tokens_seen": 2694119424
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018549648946840523,
      "loss": 1.2733,
      "theoretical_loss": 3.3770994477922365,
      "tokens_seen": 2694184960
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001854764292878636,
      "loss": 1.3157,
      "theoretical_loss": 3.3770930473670977,
      "tokens_seen": 2694250496
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018545636910732196,
      "loss": 1.3173,
      "theoretical_loss": 3.377086647141234,
      "tokens_seen": 2694316032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018543630892678035,
      "loss": 1.3513,
      "theoretical_loss": 3.3770802471146353,
      "tokens_seen": 2694381568
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001854162487462387,
      "loss": 1.3295,
      "theoretical_loss": 3.37707384728729,
      "tokens_seen": 2694447104
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001853961885656971,
      "loss": 1.3679,
      "theoretical_loss": 3.3770674476591864,
      "tokens_seen": 2694512640
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018537612838515546,
      "loss": 1.3452,
      "theoretical_loss": 3.3770610482303147,
      "tokens_seen": 2694578176
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018535606820461385,
      "loss": 1.3549,
      "theoretical_loss": 3.377054649000663,
      "tokens_seen": 2694643712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001853360080240722,
      "loss": 1.2412,
      "theoretical_loss": 3.3770482499702212,
      "tokens_seen": 2694709248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001853159478435306,
      "loss": 1.2788,
      "theoretical_loss": 3.3770418511389773,
      "tokens_seen": 2694774784
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018529588766298896,
      "loss": 1.2072,
      "theoretical_loss": 3.37703545250692,
      "tokens_seen": 2694840320
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018527582748244733,
      "loss": 1.3047,
      "theoretical_loss": 3.3770290540740393,
      "tokens_seen": 2694905856
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018525576730190572,
      "loss": 1.2673,
      "theoretical_loss": 3.3770226558403236,
      "tokens_seen": 2694971392
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018523570712136408,
      "loss": 1.2546,
      "theoretical_loss": 3.377016257805762,
      "tokens_seen": 2695036928
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018521564694082247,
      "loss": 1.3489,
      "theoretical_loss": 3.377009859970343,
      "tokens_seen": 2695102464
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018519558676028083,
      "loss": 1.3243,
      "theoretical_loss": 3.3770034623340566,
      "tokens_seen": 2695168000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018517552657973922,
      "loss": 1.2213,
      "theoretical_loss": 3.3769970648968908,
      "tokens_seen": 2695233536
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 979344,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.375655174255371,
      "objective/train/theoretical_loss": 3.3769906676588346,
      "objective/train/tokens_used": 1065758176,
      "theoretical_loss": 3.3769906676588346,
      "tokens_seen": 2695299072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018515546639919758,
      "loss": 1.3557,
      "theoretical_loss": 3.3769906676588346,
      "tokens_seen": 2695299072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018513540621865597,
      "loss": 1.2539,
      "theoretical_loss": 3.3769842706198774,
      "tokens_seen": 2695364608
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018511534603811433,
      "loss": 1.2233,
      "theoretical_loss": 3.376977873780008,
      "tokens_seen": 2695430144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018509528585757272,
      "loss": 1.293,
      "theoretical_loss": 3.3769714771392154,
      "tokens_seen": 2695495680
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018507522567703108,
      "loss": 1.3252,
      "theoretical_loss": 3.3769650806974885,
      "tokens_seen": 2695561216
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018505516549648947,
      "loss": 1.2073,
      "theoretical_loss": 3.376958684454816,
      "tokens_seen": 2695626752
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018503510531594786,
      "loss": 1.2002,
      "theoretical_loss": 3.3769522884111876,
      "tokens_seen": 2695692288
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018501504513540623,
      "loss": 1.3129,
      "theoretical_loss": 3.3769458925665914,
      "tokens_seen": 2695757824
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018499498495486462,
      "loss": 1.2587,
      "theoretical_loss": 3.3769394969210174,
      "tokens_seen": 2695823360
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018497492477432298,
      "loss": 1.2704,
      "theoretical_loss": 3.376933101474453,
      "tokens_seen": 2695888896
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018495486459378137,
      "loss": 1.1931,
      "theoretical_loss": 3.3769267062268886,
      "tokens_seen": 2695954432
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018493480441323973,
      "loss": 1.2588,
      "theoretical_loss": 3.376920311178313,
      "tokens_seen": 2696019968
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018491474423269812,
      "loss": 1.1945,
      "theoretical_loss": 3.3769139163287147,
      "tokens_seen": 2696085504
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018489468405215648,
      "loss": 1.2936,
      "theoretical_loss": 3.3769075216780826,
      "tokens_seen": 2696151040
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018487462387161484,
      "loss": 1.2884,
      "theoretical_loss": 3.376901127226406,
      "tokens_seen": 2696216576
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018485456369107323,
      "loss": 1.4073,
      "theoretical_loss": 3.376894732973674,
      "tokens_seen": 2696282112
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001848345035105316,
      "loss": 1.3461,
      "theoretical_loss": 3.376888338919875,
      "tokens_seen": 2696347648
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018481444332998998,
      "loss": 1.3021,
      "theoretical_loss": 3.3768819450649987,
      "tokens_seen": 2696413184
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018479438314944835,
      "loss": 1.3153,
      "theoretical_loss": 3.376875551409033,
      "tokens_seen": 2696478720
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018477432296890674,
      "loss": 1.3215,
      "theoretical_loss": 3.3768691579519685,
      "tokens_seen": 2696544256
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001847542627883651,
      "loss": 1.257,
      "theoretical_loss": 3.3768627646937928,
      "tokens_seen": 2696609792
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001847342026078235,
      "loss": 1.2458,
      "theoretical_loss": 3.3768563716344957,
      "tokens_seen": 2696675328
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018471414242728185,
      "loss": 1.3325,
      "theoretical_loss": 3.3768499787740653,
      "tokens_seen": 2696740864
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018469408224674024,
      "loss": 1.3164,
      "theoretical_loss": 3.3768435861124915,
      "tokens_seen": 2696806400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001846740220661986,
      "loss": 1.2158,
      "theoretical_loss": 3.3768371936497625,
      "tokens_seen": 2696871936
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018465396188565696,
      "loss": 1.3166,
      "theoretical_loss": 3.376830801385868,
      "tokens_seen": 2696937472
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018463390170511535,
      "loss": 1.2954,
      "theoretical_loss": 3.376824409320797,
      "tokens_seen": 2697003008
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018461384152457371,
      "loss": 1.2463,
      "theoretical_loss": 3.3768180174545375,
      "tokens_seen": 2697068544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001845937813440321,
      "loss": 1.2675,
      "theoretical_loss": 3.376811625787079,
      "tokens_seen": 2697134080
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018457372116349047,
      "loss": 1.2966,
      "theoretical_loss": 3.376805234318411,
      "tokens_seen": 2697199616
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018455366098294886,
      "loss": 1.2181,
      "theoretical_loss": 3.3767988430485225,
      "tokens_seen": 2697265152
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018453360080240722,
      "loss": 1.2312,
      "theoretical_loss": 3.3767924519774013,
      "tokens_seen": 2697330688
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001845135406218656,
      "loss": 1.2595,
      "theoretical_loss": 3.3767860611050375,
      "tokens_seen": 2697396224
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018449348044132397,
      "loss": 1.2673,
      "theoretical_loss": 3.3767796704314197,
      "tokens_seen": 2697461760
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018447342026078233,
      "loss": 1.2563,
      "theoretical_loss": 3.376773279956537,
      "tokens_seen": 2697527296
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018445336008024072,
      "loss": 1.1277,
      "theoretical_loss": 3.3767668896803786,
      "tokens_seen": 2697592832
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018443329989969908,
      "loss": 1.3412,
      "theoretical_loss": 3.3767604996029332,
      "tokens_seen": 2697658368
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018441323971915747,
      "loss": 1.3468,
      "theoretical_loss": 3.3767541097241898,
      "tokens_seen": 2697723904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018439317953861583,
      "loss": 1.2762,
      "theoretical_loss": 3.376747720044137,
      "tokens_seen": 2697789440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018437311935807422,
      "loss": 1.2268,
      "theoretical_loss": 3.3767413305627647,
      "tokens_seen": 2697854976
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018435305917753259,
      "loss": 1.263,
      "theoretical_loss": 3.3767349412800614,
      "tokens_seen": 2697920512
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018433299899699098,
      "loss": 1.2317,
      "theoretical_loss": 3.376728552196016,
      "tokens_seen": 2697986048
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018431293881644934,
      "loss": 1.3815,
      "theoretical_loss": 3.3767221633106175,
      "tokens_seen": 2698051584
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018429287863590773,
      "loss": 1.2771,
      "theoretical_loss": 3.3767157746238547,
      "tokens_seen": 2698117120
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018427281845536612,
      "loss": 1.3258,
      "theoretical_loss": 3.3767093861357176,
      "tokens_seen": 2698182656
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018425275827482448,
      "loss": 1.3029,
      "theoretical_loss": 3.3767029978461944,
      "tokens_seen": 2698248192
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018423269809428287,
      "loss": 1.1993,
      "theoretical_loss": 3.3766966097552737,
      "tokens_seen": 2698313728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018421263791374123,
      "loss": 1.1792,
      "theoretical_loss": 3.3766902218629458,
      "tokens_seen": 2698379264
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018419257773319962,
      "loss": 1.2025,
      "theoretical_loss": 3.3766838341691985,
      "tokens_seen": 2698444800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018417251755265798,
      "loss": 1.2751,
      "theoretical_loss": 3.3766774466740213,
      "tokens_seen": 2698510336
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 980656,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4430756568908691,
      "objective/train/theoretical_loss": 3.376671059377403,
      "objective/train/tokens_used": 1069034976,
      "theoretical_loss": 3.376671059377403,
      "tokens_seen": 2698575872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018415245737211637,
      "loss": 1.4259,
      "theoretical_loss": 3.376671059377403,
      "tokens_seen": 2698575872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018413239719157473,
      "loss": 1.1579,
      "theoretical_loss": 3.376664672279333,
      "tokens_seen": 2698641408
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018411233701103312,
      "loss": 1.3168,
      "theoretical_loss": 3.3766582853797997,
      "tokens_seen": 2698706944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018409227683049149,
      "loss": 1.3789,
      "theoretical_loss": 3.3766518986787926,
      "tokens_seen": 2698772480
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018407221664994985,
      "loss": 1.3199,
      "theoretical_loss": 3.3766455121763004,
      "tokens_seen": 2698838016
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018405215646940824,
      "loss": 1.2997,
      "theoretical_loss": 3.3766391258723125,
      "tokens_seen": 2698903552
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001840320962888666,
      "loss": 1.3263,
      "theoretical_loss": 3.376632739766818,
      "tokens_seen": 2698969088
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000184012036108325,
      "loss": 1.4628,
      "theoretical_loss": 3.376626353859805,
      "tokens_seen": 2699034624
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018399197592778335,
      "loss": 1.2597,
      "theoretical_loss": 3.376619968151263,
      "tokens_seen": 2699100160
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018397191574724174,
      "loss": 1.2576,
      "theoretical_loss": 3.376613582641182,
      "tokens_seen": 2699165696
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001839518555667001,
      "loss": 1.2593,
      "theoretical_loss": 3.376607197329549,
      "tokens_seen": 2699231232
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001839317953861585,
      "loss": 1.3755,
      "theoretical_loss": 3.376600812216355,
      "tokens_seen": 2699296768
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018391173520561685,
      "loss": 1.3459,
      "theoretical_loss": 3.3765944273015878,
      "tokens_seen": 2699362304
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018389167502507522,
      "loss": 1.3657,
      "theoretical_loss": 3.376588042585237,
      "tokens_seen": 2699427840
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001838716148445336,
      "loss": 1.2252,
      "theoretical_loss": 3.376581658067291,
      "tokens_seen": 2699493376
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018385155466399197,
      "loss": 1.2095,
      "theoretical_loss": 3.3765752737477395,
      "tokens_seen": 2699558912
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018383149448345036,
      "loss": 1.3188,
      "theoretical_loss": 3.3765688896265713,
      "tokens_seen": 2699624448
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018381143430290872,
      "loss": 1.3971,
      "theoretical_loss": 3.3765625057037756,
      "tokens_seen": 2699689984
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001837913741223671,
      "loss": 1.2705,
      "theoretical_loss": 3.3765561219793403,
      "tokens_seen": 2699755520
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018377131394182547,
      "loss": 1.1753,
      "theoretical_loss": 3.376549738453256,
      "tokens_seen": 2699821056
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018375125376128386,
      "loss": 1.3216,
      "theoretical_loss": 3.376543355125511,
      "tokens_seen": 2699886592
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018373119358074222,
      "loss": 1.2014,
      "theoretical_loss": 3.376536971996094,
      "tokens_seen": 2699952128
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018371113340020058,
      "loss": 1.3409,
      "theoretical_loss": 3.376530589064995,
      "tokens_seen": 2700017664
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018369107321965897,
      "loss": 1.2144,
      "theoretical_loss": 3.3765242063322014,
      "tokens_seen": 2700083200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018367101303911734,
      "loss": 1.2837,
      "theoretical_loss": 3.376517823797704,
      "tokens_seen": 2700148736
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018365095285857573,
      "loss": 1.3162,
      "theoretical_loss": 3.3765114414614907,
      "tokens_seen": 2700214272
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001836308926780341,
      "loss": 1.2485,
      "theoretical_loss": 3.376505059323551,
      "tokens_seen": 2700279808
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018361083249749248,
      "loss": 1.2465,
      "theoretical_loss": 3.3764986773838737,
      "tokens_seen": 2700345344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018359077231695084,
      "loss": 1.3127,
      "theoretical_loss": 3.3764922956424477,
      "tokens_seen": 2700410880
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018357071213640923,
      "loss": 1.2483,
      "theoretical_loss": 3.3764859140992627,
      "tokens_seen": 2700476416
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018355065195586762,
      "loss": 1.3849,
      "theoretical_loss": 3.376479532754307,
      "tokens_seen": 2700541952
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000183530591775326,
      "loss": 1.3264,
      "theoretical_loss": 3.37647315160757,
      "tokens_seen": 2700607488
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018351053159478437,
      "loss": 1.32,
      "theoretical_loss": 3.3764667706590408,
      "tokens_seen": 2700673024
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018349047141424273,
      "loss": 1.3413,
      "theoretical_loss": 3.3764603899087082,
      "tokens_seen": 2700738560
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018347041123370112,
      "loss": 1.373,
      "theoretical_loss": 3.376454009356561,
      "tokens_seen": 2700804096
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018345035105315948,
      "loss": 1.383,
      "theoretical_loss": 3.3764476290025893,
      "tokens_seen": 2700869632
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018343029087261787,
      "loss": 1.2484,
      "theoretical_loss": 3.3764412488467808,
      "tokens_seen": 2700935168
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018341023069207623,
      "loss": 1.3402,
      "theoretical_loss": 3.3764348688891253,
      "tokens_seen": 2701000704
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018339017051153462,
      "loss": 1.3182,
      "theoretical_loss": 3.3764284891296117,
      "tokens_seen": 2701066240
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000183370110330993,
      "loss": 1.1654,
      "theoretical_loss": 3.376422109568229,
      "tokens_seen": 2701131776
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018335005015045138,
      "loss": 1.255,
      "theoretical_loss": 3.376415730204966,
      "tokens_seen": 2701197312
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018332998996990974,
      "loss": 1.4261,
      "theoretical_loss": 3.376409351039812,
      "tokens_seen": 2701262848
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001833099297893681,
      "loss": 1.3415,
      "theoretical_loss": 3.3764029720727566,
      "tokens_seen": 2701328384
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001832898696088265,
      "loss": 1.2836,
      "theoretical_loss": 3.376396593303788,
      "tokens_seen": 2701393920
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018326980942828485,
      "loss": 1.2847,
      "theoretical_loss": 3.3763902147328952,
      "tokens_seen": 2701459456
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018324974924774324,
      "loss": 1.3633,
      "theoretical_loss": 3.376383836360068,
      "tokens_seen": 2701524992
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001832296890672016,
      "loss": 1.2633,
      "theoretical_loss": 3.3763774581852948,
      "tokens_seen": 2701590528
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018320962888666,
      "loss": 1.3437,
      "theoretical_loss": 3.3763710802085645,
      "tokens_seen": 2701656064
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018318956870611836,
      "loss": 1.2914,
      "theoretical_loss": 3.376364702429867,
      "tokens_seen": 2701721600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018316950852557674,
      "loss": 1.3869,
      "theoretical_loss": 3.376358324849191,
      "tokens_seen": 2701787136
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 981962,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.268550157546997,
      "objective/train/theoretical_loss": 3.3763519474665245,
      "objective/train/tokens_used": 1072311776,
      "theoretical_loss": 3.3763519474665245,
      "tokens_seen": 2701852672
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001831494483450351,
      "loss": 1.2926,
      "theoretical_loss": 3.3763519474665245,
      "tokens_seen": 2701852672
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018312938816449347,
      "loss": 1.2001,
      "theoretical_loss": 3.3763455702818583,
      "tokens_seen": 2701918208
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018310932798395186,
      "loss": 1.3362,
      "theoretical_loss": 3.37633919329518,
      "tokens_seen": 2701983744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018308926780341022,
      "loss": 1.3439,
      "theoretical_loss": 3.37633281650648,
      "tokens_seen": 2702049280
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001830692076228686,
      "loss": 1.2448,
      "theoretical_loss": 3.3763264399157458,
      "tokens_seen": 2702114816
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018304914744232697,
      "loss": 1.3015,
      "theoretical_loss": 3.3763200635229675,
      "tokens_seen": 2702180352
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018302908726178536,
      "loss": 1.194,
      "theoretical_loss": 3.376313687328134,
      "tokens_seen": 2702245888
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018300902708124372,
      "loss": 1.3356,
      "theoretical_loss": 3.3763073113312343,
      "tokens_seen": 2702311424
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001829889669007021,
      "loss": 1.2151,
      "theoretical_loss": 3.3763009355322575,
      "tokens_seen": 2702376960
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018296890672016048,
      "loss": 1.2108,
      "theoretical_loss": 3.3762945599311927,
      "tokens_seen": 2702442496
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018294884653961886,
      "loss": 1.3062,
      "theoretical_loss": 3.376288184528028,
      "tokens_seen": 2702508032
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018292878635907723,
      "loss": 1.3143,
      "theoretical_loss": 3.3762818093227542,
      "tokens_seen": 2702573568
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001829087261785356,
      "loss": 1.2904,
      "theoretical_loss": 3.3762754343153594,
      "tokens_seen": 2702639104
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018288866599799398,
      "loss": 1.3086,
      "theoretical_loss": 3.376269059505832,
      "tokens_seen": 2702704640
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018286860581745234,
      "loss": 1.2936,
      "theoretical_loss": 3.3762626848941624,
      "tokens_seen": 2702770176
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018284854563691073,
      "loss": 1.2879,
      "theoretical_loss": 3.376256310480339,
      "tokens_seen": 2702835712
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001828284854563691,
      "loss": 1.4082,
      "theoretical_loss": 3.3762499362643505,
      "tokens_seen": 2702901248
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018280842527582748,
      "loss": 1.3968,
      "theoretical_loss": 3.3762435622461866,
      "tokens_seen": 2702966784
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018278836509528587,
      "loss": 1.2412,
      "theoretical_loss": 3.376237188425836,
      "tokens_seen": 2703032320
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018276830491474426,
      "loss": 1.2635,
      "theoretical_loss": 3.376230814803288,
      "tokens_seen": 2703097856
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018274824473420262,
      "loss": 1.2619,
      "theoretical_loss": 3.3762244413785316,
      "tokens_seen": 2703163392
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018272818455366098,
      "loss": 1.3502,
      "theoretical_loss": 3.3762180681515557,
      "tokens_seen": 2703228928
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018270812437311937,
      "loss": 1.2567,
      "theoretical_loss": 3.3762116951223495,
      "tokens_seen": 2703294464
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018268806419257774,
      "loss": 1.2372,
      "theoretical_loss": 3.376205322290902,
      "tokens_seen": 2703360000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018266800401203613,
      "loss": 1.2614,
      "theoretical_loss": 3.3761989496572022,
      "tokens_seen": 2703425536
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001826479438314945,
      "loss": 1.3517,
      "theoretical_loss": 3.3761925772212398,
      "tokens_seen": 2703491072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018262788365095288,
      "loss": 1.3059,
      "theoretical_loss": 3.376186204983003,
      "tokens_seen": 2703556608
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018260782347041124,
      "loss": 1.3102,
      "theoretical_loss": 3.3761798329424813,
      "tokens_seen": 2703622144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018258776328986963,
      "loss": 1.1371,
      "theoretical_loss": 3.376173461099664,
      "tokens_seen": 2703687680
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000182567703109328,
      "loss": 1.2977,
      "theoretical_loss": 3.3761670894545395,
      "tokens_seen": 2703753216
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018254764292878638,
      "loss": 1.2429,
      "theoretical_loss": 3.376160718007097,
      "tokens_seen": 2703818752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018252758274824474,
      "loss": 1.3842,
      "theoretical_loss": 3.3761543467573265,
      "tokens_seen": 2703884288
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001825075225677031,
      "loss": 1.3175,
      "theoretical_loss": 3.3761479757052157,
      "tokens_seen": 2703949824
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001824874623871615,
      "loss": 1.3692,
      "theoretical_loss": 3.376141604850755,
      "tokens_seen": 2704015360
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018246740220661986,
      "loss": 1.3141,
      "theoretical_loss": 3.3761352341939324,
      "tokens_seen": 2704080896
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018244734202607825,
      "loss": 1.3081,
      "theoretical_loss": 3.3761288637347375,
      "tokens_seen": 2704146432
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001824272818455366,
      "loss": 1.3467,
      "theoretical_loss": 3.3761224934731597,
      "tokens_seen": 2704211968
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000182407221664995,
      "loss": 1.3624,
      "theoretical_loss": 3.3761161234091874,
      "tokens_seen": 2704277504
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018238716148445336,
      "loss": 1.337,
      "theoretical_loss": 3.3761097535428104,
      "tokens_seen": 2704343040
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018236710130391175,
      "loss": 1.3665,
      "theoretical_loss": 3.376103383874017,
      "tokens_seen": 2704408576
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001823470411233701,
      "loss": 1.3692,
      "theoretical_loss": 3.376097014402797,
      "tokens_seen": 2704474112
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018232698094282847,
      "loss": 1.2274,
      "theoretical_loss": 3.3760906451291386,
      "tokens_seen": 2704539648
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018230692076228686,
      "loss": 1.3324,
      "theoretical_loss": 3.3760842760530316,
      "tokens_seen": 2704605184
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018228686058174523,
      "loss": 1.2264,
      "theoretical_loss": 3.376077907174465,
      "tokens_seen": 2704670720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018226680040120361,
      "loss": 1.46,
      "theoretical_loss": 3.376071538493428,
      "tokens_seen": 2704736256
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018224674022066198,
      "loss": 1.3338,
      "theoretical_loss": 3.376065170009909,
      "tokens_seen": 2704801792
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018222668004012037,
      "loss": 1.3225,
      "theoretical_loss": 3.376058801723898,
      "tokens_seen": 2704867328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018220661985957873,
      "loss": 1.2547,
      "theoretical_loss": 3.376052433635383,
      "tokens_seen": 2704932864
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018218655967903712,
      "loss": 1.2694,
      "theoretical_loss": 3.3760460657443545,
      "tokens_seen": 2704998400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018216649949849548,
      "loss": 1.4207,
      "theoretical_loss": 3.3760396980508007,
      "tokens_seen": 2705063936
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 983087,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3899855613708496,
      "objective/train/theoretical_loss": 3.3760333305547108,
      "objective/train/tokens_used": 1075588576,
      "theoretical_loss": 3.3760333305547108,
      "tokens_seen": 2705129472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018214643931795384,
      "loss": 1.429,
      "theoretical_loss": 3.3760333305547108,
      "tokens_seen": 2705129472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018212637913741223,
      "loss": 1.2938,
      "theoretical_loss": 3.376026963256074,
      "tokens_seen": 2705195008
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001821063189568706,
      "loss": 1.3034,
      "theoretical_loss": 3.3760205961548793,
      "tokens_seen": 2705260544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018208625877632898,
      "loss": 1.2755,
      "theoretical_loss": 3.3760142292511155,
      "tokens_seen": 2705326080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018206619859578735,
      "loss": 1.2975,
      "theoretical_loss": 3.376007862544772,
      "tokens_seen": 2705391616
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018204613841524576,
      "loss": 1.25,
      "theoretical_loss": 3.376001496035838,
      "tokens_seen": 2705457152
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018202607823470412,
      "loss": 1.3093,
      "theoretical_loss": 3.375995129724303,
      "tokens_seen": 2705522688
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001820060180541625,
      "loss": 1.2637,
      "theoretical_loss": 3.375988763610155,
      "tokens_seen": 2705588224
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018198595787362088,
      "loss": 1.2303,
      "theoretical_loss": 3.375982397693384,
      "tokens_seen": 2705653760
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018196589769307926,
      "loss": 1.256,
      "theoretical_loss": 3.3759760319739787,
      "tokens_seen": 2705719296
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018194583751253763,
      "loss": 1.2352,
      "theoretical_loss": 3.3759696664519288,
      "tokens_seen": 2705784832
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000181925777331996,
      "loss": 1.3654,
      "theoretical_loss": 3.3759633011272223,
      "tokens_seen": 2705850368
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018190571715145438,
      "loss": 1.2762,
      "theoretical_loss": 3.3759569359998487,
      "tokens_seen": 2705915904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018188565697091274,
      "loss": 1.3371,
      "theoretical_loss": 3.3759505710697977,
      "tokens_seen": 2705981440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018186559679037113,
      "loss": 1.3019,
      "theoretical_loss": 3.375944206337058,
      "tokens_seen": 2706046976
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001818455366098295,
      "loss": 1.2163,
      "theoretical_loss": 3.3759378418016186,
      "tokens_seen": 2706112512
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018182547642928788,
      "loss": 1.2684,
      "theoretical_loss": 3.375931477463469,
      "tokens_seen": 2706178048
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018180541624874624,
      "loss": 1.2597,
      "theoretical_loss": 3.375925113322598,
      "tokens_seen": 2706243584
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018178535606820463,
      "loss": 1.2853,
      "theoretical_loss": 3.3759187493789944,
      "tokens_seen": 2706309120
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000181765295887663,
      "loss": 1.2427,
      "theoretical_loss": 3.375912385632648,
      "tokens_seen": 2706374656
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018174523570712136,
      "loss": 1.2836,
      "theoretical_loss": 3.375906022083547,
      "tokens_seen": 2706440192
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018172517552657975,
      "loss": 1.3165,
      "theoretical_loss": 3.3758996587316816,
      "tokens_seen": 2706505728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001817051153460381,
      "loss": 1.3351,
      "theoretical_loss": 3.37589329557704,
      "tokens_seen": 2706571264
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001816850551654965,
      "loss": 1.288,
      "theoretical_loss": 3.375886932619612,
      "tokens_seen": 2706636800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018166499498495486,
      "loss": 1.4257,
      "theoretical_loss": 3.375880569859387,
      "tokens_seen": 2706702336
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018164493480441325,
      "loss": 1.3389,
      "theoretical_loss": 3.3758742072963526,
      "tokens_seen": 2706767872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001816248746238716,
      "loss": 1.3796,
      "theoretical_loss": 3.3758678449304993,
      "tokens_seen": 2706833408
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018160481444333,
      "loss": 1.3624,
      "theoretical_loss": 3.375861482761815,
      "tokens_seen": 2706898944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018158475426278836,
      "loss": 1.3978,
      "theoretical_loss": 3.3758551207902903,
      "tokens_seen": 2706964480
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018156469408224673,
      "loss": 1.3236,
      "theoretical_loss": 3.3758487590159136,
      "tokens_seen": 2707030016
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018154463390170512,
      "loss": 1.3101,
      "theoretical_loss": 3.3758423974386735,
      "tokens_seen": 2707095552
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018152457372116348,
      "loss": 1.4664,
      "theoretical_loss": 3.37583603605856,
      "tokens_seen": 2707161088
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018150451354062187,
      "loss": 1.3504,
      "theoretical_loss": 3.3758296748755616,
      "tokens_seen": 2707226624
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018148445336008023,
      "loss": 1.2378,
      "theoretical_loss": 3.375823313889668,
      "tokens_seen": 2707292160
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018146439317953862,
      "loss": 1.2975,
      "theoretical_loss": 3.375816953100868,
      "tokens_seen": 2707357696
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018144433299899698,
      "loss": 1.3198,
      "theoretical_loss": 3.37581059250915,
      "tokens_seen": 2707423232
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018142427281845537,
      "loss": 1.4959,
      "theoretical_loss": 3.3758042321145045,
      "tokens_seen": 2707488768
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018140421263791373,
      "loss": 1.3353,
      "theoretical_loss": 3.3757978719169195,
      "tokens_seen": 2707554304
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018138415245737212,
      "loss": 1.3857,
      "theoretical_loss": 3.3757915119163853,
      "tokens_seen": 2707619840
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018136409227683048,
      "loss": 1.2183,
      "theoretical_loss": 3.37578515211289,
      "tokens_seen": 2707685376
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018134403209628885,
      "loss": 1.3534,
      "theoretical_loss": 3.3757787925064227,
      "tokens_seen": 2707750912
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018132397191574724,
      "loss": 1.2593,
      "theoretical_loss": 3.375772433096973,
      "tokens_seen": 2707816448
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018130391173520563,
      "loss": 1.3121,
      "theoretical_loss": 3.3757660738845296,
      "tokens_seen": 2707881984
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018128385155466401,
      "loss": 1.1856,
      "theoretical_loss": 3.375759714869082,
      "tokens_seen": 2707947520
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018126379137412238,
      "loss": 1.3425,
      "theoretical_loss": 3.3757533560506197,
      "tokens_seen": 2708013056
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018124373119358077,
      "loss": 1.2115,
      "theoretical_loss": 3.375746997429131,
      "tokens_seen": 2708078592
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018122367101303913,
      "loss": 1.3329,
      "theoretical_loss": 3.375740639004606,
      "tokens_seen": 2708144128
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018120361083249752,
      "loss": 1.3617,
      "theoretical_loss": 3.3757342807770323,
      "tokens_seen": 2708209664
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018118355065195588,
      "loss": 1.4144,
      "theoretical_loss": 3.3757279227464005,
      "tokens_seen": 2708275200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018116349047141424,
      "loss": 1.3063,
      "theoretical_loss": 3.3757215649126993,
      "tokens_seen": 2708340736
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 984234,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.212406039237976,
      "objective/train/theoretical_loss": 3.375715207275918,
      "objective/train/tokens_used": 1078865376,
      "theoretical_loss": 3.375715207275918,
      "tokens_seen": 2708406272
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018114343029087263,
      "loss": 1.3512,
      "theoretical_loss": 3.375715207275918,
      "tokens_seen": 2708406272
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000181123370110331,
      "loss": 1.3705,
      "theoretical_loss": 3.3757088498360446,
      "tokens_seen": 2708471808
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018110330992978938,
      "loss": 1.355,
      "theoretical_loss": 3.37570249259307,
      "tokens_seen": 2708537344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018108324974924775,
      "loss": 1.2609,
      "theoretical_loss": 3.3756961355469817,
      "tokens_seen": 2708602880
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018106318956870613,
      "loss": 1.3343,
      "theoretical_loss": 3.3756897786977698,
      "tokens_seen": 2708668416
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001810431293881645,
      "loss": 1.2499,
      "theoretical_loss": 3.3756834220454235,
      "tokens_seen": 2708733952
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001810230692076229,
      "loss": 1.1874,
      "theoretical_loss": 3.3756770655899313,
      "tokens_seen": 2708799488
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018100300902708125,
      "loss": 1.4343,
      "theoretical_loss": 3.375670709331283,
      "tokens_seen": 2708865024
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001809829488465396,
      "loss": 1.3782,
      "theoretical_loss": 3.3756643532694675,
      "tokens_seen": 2708930560
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000180962888665998,
      "loss": 1.4183,
      "theoretical_loss": 3.375657997404474,
      "tokens_seen": 2708996096
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018094282848545636,
      "loss": 1.3337,
      "theoretical_loss": 3.375651641736291,
      "tokens_seen": 2709061632
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018092276830491475,
      "loss": 1.4367,
      "theoretical_loss": 3.3756452862649087,
      "tokens_seen": 2709127168
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018090270812437311,
      "loss": 1.3099,
      "theoretical_loss": 3.3756389309903154,
      "tokens_seen": 2709192704
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001808826479438315,
      "loss": 1.3732,
      "theoretical_loss": 3.375632575912501,
      "tokens_seen": 2709258240
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018086258776328987,
      "loss": 1.3269,
      "theoretical_loss": 3.375626221031454,
      "tokens_seen": 2709323776
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018084252758274826,
      "loss": 1.2156,
      "theoretical_loss": 3.3756198663471637,
      "tokens_seen": 2709389312
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018082246740220662,
      "loss": 1.4591,
      "theoretical_loss": 3.375613511859619,
      "tokens_seen": 2709454848
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000180802407221665,
      "loss": 1.4467,
      "theoretical_loss": 3.37560715756881,
      "tokens_seen": 2709520384
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018078234704112337,
      "loss": 1.2421,
      "theoretical_loss": 3.375600803474725,
      "tokens_seen": 2709585920
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018076228686058173,
      "loss": 1.38,
      "theoretical_loss": 3.3755944495773536,
      "tokens_seen": 2709651456
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018074222668004012,
      "loss": 1.3948,
      "theoretical_loss": 3.3755880958766844,
      "tokens_seen": 2709716992
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018072216649949848,
      "loss": 1.3127,
      "theoretical_loss": 3.375581742372707,
      "tokens_seen": 2709782528
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018070210631895687,
      "loss": 1.3331,
      "theoretical_loss": 3.37557538906541,
      "tokens_seen": 2709848064
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018068204613841523,
      "loss": 1.207,
      "theoretical_loss": 3.3755690359547836,
      "tokens_seen": 2709913600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018066198595787362,
      "loss": 1.3121,
      "theoretical_loss": 3.375562683040816,
      "tokens_seen": 2709979136
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018064192577733199,
      "loss": 1.2909,
      "theoretical_loss": 3.375556330323497,
      "tokens_seen": 2710044672
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018062186559679038,
      "loss": 1.3125,
      "theoretical_loss": 3.375549977802815,
      "tokens_seen": 2710110208
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018060180541624874,
      "loss": 1.34,
      "theoretical_loss": 3.37554362547876,
      "tokens_seen": 2710175744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001805817452357071,
      "loss": 1.358,
      "theoretical_loss": 3.3755372733513207,
      "tokens_seen": 2710241280
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001805616850551655,
      "loss": 1.2636,
      "theoretical_loss": 3.3755309214204865,
      "tokens_seen": 2710306816
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018054162487462388,
      "loss": 1.3878,
      "theoretical_loss": 3.375524569686246,
      "tokens_seen": 2710372352
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018052156469408227,
      "loss": 1.3536,
      "theoretical_loss": 3.375518218148589,
      "tokens_seen": 2710437888
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018050150451354063,
      "loss": 1.266,
      "theoretical_loss": 3.3755118668075044,
      "tokens_seen": 2710503424
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018048144433299902,
      "loss": 1.2186,
      "theoretical_loss": 3.3755055156629816,
      "tokens_seen": 2710568960
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018046138415245738,
      "loss": 1.3233,
      "theoretical_loss": 3.375499164715009,
      "tokens_seen": 2710634496
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018044132397191577,
      "loss": 1.3231,
      "theoretical_loss": 3.375492813963577,
      "tokens_seen": 2710700032
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018042126379137413,
      "loss": 1.472,
      "theoretical_loss": 3.3754864634086736,
      "tokens_seen": 2710765568
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018040120361083252,
      "loss": 1.3516,
      "theoretical_loss": 3.3754801130502887,
      "tokens_seen": 2710831104
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018038114343029088,
      "loss": 1.3306,
      "theoretical_loss": 3.3754737628884115,
      "tokens_seen": 2710896640
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018036108324974925,
      "loss": 1.4419,
      "theoretical_loss": 3.3754674129230304,
      "tokens_seen": 2710962176
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018034102306920764,
      "loss": 1.2292,
      "theoretical_loss": 3.375461063154135,
      "tokens_seen": 2711027712
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000180320962888666,
      "loss": 1.2546,
      "theoretical_loss": 3.375454713581715,
      "tokens_seen": 2711093248
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001803009027081244,
      "loss": 1.3587,
      "theoretical_loss": 3.3754483642057584,
      "tokens_seen": 2711158784
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018028084252758275,
      "loss": 1.3103,
      "theoretical_loss": 3.375442015026256,
      "tokens_seen": 2711224320
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018026078234704114,
      "loss": 1.1019,
      "theoretical_loss": 3.3754356660431952,
      "tokens_seen": 2711289856
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001802407221664995,
      "loss": 1.3324,
      "theoretical_loss": 3.3754293172565664,
      "tokens_seen": 2711355392
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001802206619859579,
      "loss": 1.3489,
      "theoretical_loss": 3.375422968666358,
      "tokens_seen": 2711420928
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018020060180541625,
      "loss": 1.4152,
      "theoretical_loss": 3.3754166202725604,
      "tokens_seen": 2711486464
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018018054162487462,
      "loss": 1.4148,
      "theoretical_loss": 3.3754102720751615,
      "tokens_seen": 2711552000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000180160481444333,
      "loss": 1.3362,
      "theoretical_loss": 3.375403924074151,
      "tokens_seen": 2711617536
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 985372,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3839335441589355,
      "objective/train/theoretical_loss": 3.375397576269518,
      "objective/train/tokens_used": 1082142176,
      "theoretical_loss": 3.375397576269518,
      "tokens_seen": 2711683072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018014042126379137,
      "loss": 1.298,
      "theoretical_loss": 3.375397576269518,
      "tokens_seen": 2711683072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018012036108324976,
      "loss": 1.2714,
      "theoretical_loss": 3.3753912286612513,
      "tokens_seen": 2711748608
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018010030090270812,
      "loss": 1.2916,
      "theoretical_loss": 3.3753848812493406,
      "tokens_seen": 2711814144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001800802407221665,
      "loss": 1.2888,
      "theoretical_loss": 3.3753785340337754,
      "tokens_seen": 2711879680
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018006018054162487,
      "loss": 1.2693,
      "theoretical_loss": 3.375372187014544,
      "tokens_seen": 2711945216
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018004012036108326,
      "loss": 1.2787,
      "theoretical_loss": 3.375365840191636,
      "tokens_seen": 2712010752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018002006018054162,
      "loss": 1.2967,
      "theoretical_loss": 3.375359493565041,
      "tokens_seen": 2712076288
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017999999999999998,
      "loss": 1.1786,
      "theoretical_loss": 3.3753531471347475,
      "tokens_seen": 2712141824
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017997993981945837,
      "loss": 1.4275,
      "theoretical_loss": 3.375346800900745,
      "tokens_seen": 2712207360
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017995987963891674,
      "loss": 1.2934,
      "theoretical_loss": 3.3753404548630224,
      "tokens_seen": 2712272896
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017993981945837513,
      "loss": 1.3358,
      "theoretical_loss": 3.375334109021569,
      "tokens_seen": 2712338432
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001799197592778335,
      "loss": 1.3402,
      "theoretical_loss": 3.375327763376375,
      "tokens_seen": 2712403968
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017989969909729188,
      "loss": 1.3088,
      "theoretical_loss": 3.3753214179274282,
      "tokens_seen": 2712469504
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017987963891675024,
      "loss": 1.3327,
      "theoretical_loss": 3.3753150726747183,
      "tokens_seen": 2712535040
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017985957873620863,
      "loss": 1.344,
      "theoretical_loss": 3.3753087276182345,
      "tokens_seen": 2712600576
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000179839518555667,
      "loss": 1.5,
      "theoretical_loss": 3.375302382757966,
      "tokens_seen": 2712666112
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017981945837512538,
      "loss": 1.4291,
      "theoretical_loss": 3.375296038093902,
      "tokens_seen": 2712731648
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017979939819458377,
      "loss": 1.2942,
      "theoretical_loss": 3.3752896936260313,
      "tokens_seen": 2712797184
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017977933801404213,
      "loss": 1.3567,
      "theoretical_loss": 3.3752833493543437,
      "tokens_seen": 2712862720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017975927783350052,
      "loss": 1.44,
      "theoretical_loss": 3.3752770052788286,
      "tokens_seen": 2712928256
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017973921765295888,
      "loss": 1.2012,
      "theoretical_loss": 3.3752706613994743,
      "tokens_seen": 2712993792
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017971915747241727,
      "loss": 1.2104,
      "theoretical_loss": 3.3752643177162702,
      "tokens_seen": 2713059328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017969909729187563,
      "loss": 1.324,
      "theoretical_loss": 3.375257974229206,
      "tokens_seen": 2713124864
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017967903711133402,
      "loss": 1.1641,
      "theoretical_loss": 3.375251630938271,
      "tokens_seen": 2713190400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017965897693079239,
      "loss": 1.2993,
      "theoretical_loss": 3.3752452878434536,
      "tokens_seen": 2713255936
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017963891675025078,
      "loss": 1.4211,
      "theoretical_loss": 3.3752389449447433,
      "tokens_seen": 2713321472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017961885656970914,
      "loss": 1.3962,
      "theoretical_loss": 3.37523260224213,
      "tokens_seen": 2713387008
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001795987963891675,
      "loss": 1.3953,
      "theoretical_loss": 3.375226259735602,
      "tokens_seen": 2713452544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001795787362086259,
      "loss": 1.299,
      "theoretical_loss": 3.3752199174251487,
      "tokens_seen": 2713518080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017955867602808425,
      "loss": 1.3269,
      "theoretical_loss": 3.37521357531076,
      "tokens_seen": 2713583616
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017953861584754264,
      "loss": 1.184,
      "theoretical_loss": 3.375207233392424,
      "tokens_seen": 2713649152
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000179518555667001,
      "loss": 1.2643,
      "theoretical_loss": 3.3752008916701306,
      "tokens_seen": 2713714688
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001794984954864594,
      "loss": 1.5399,
      "theoretical_loss": 3.375194550143869,
      "tokens_seen": 2713780224
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017947843530591775,
      "loss": 1.3362,
      "theoretical_loss": 3.3751882088136282,
      "tokens_seen": 2713845760
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017945837512537614,
      "loss": 1.2977,
      "theoretical_loss": 3.375181867679397,
      "tokens_seen": 2713911296
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001794383149448345,
      "loss": 1.2893,
      "theoretical_loss": 3.375175526741166,
      "tokens_seen": 2713976832
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017941825476429287,
      "loss": 1.3033,
      "theoretical_loss": 3.375169185998923,
      "tokens_seen": 2714042368
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017939819458375126,
      "loss": 1.3452,
      "theoretical_loss": 3.3751628454526577,
      "tokens_seen": 2714107904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017937813440320962,
      "loss": 1.2127,
      "theoretical_loss": 3.3751565051023595,
      "tokens_seen": 2714173440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000179358074222668,
      "loss": 1.3915,
      "theoretical_loss": 3.3751501649480176,
      "tokens_seen": 2714238976
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017933801404212637,
      "loss": 1.1754,
      "theoretical_loss": 3.3751438249896206,
      "tokens_seen": 2714304512
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017931795386158476,
      "loss": 1.2748,
      "theoretical_loss": 3.375137485227158,
      "tokens_seen": 2714370048
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017929789368104312,
      "loss": 1.3254,
      "theoretical_loss": 3.3751311456606197,
      "tokens_seen": 2714435584
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001792778335005015,
      "loss": 1.3941,
      "theoretical_loss": 3.375124806289994,
      "tokens_seen": 2714501120
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017925777331995988,
      "loss": 1.3558,
      "theoretical_loss": 3.3751184671152705,
      "tokens_seen": 2714566656
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017923771313941826,
      "loss": 1.3026,
      "theoretical_loss": 3.375112128136439,
      "tokens_seen": 2714632192
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017921765295887663,
      "loss": 1.359,
      "theoretical_loss": 3.3751057893534875,
      "tokens_seen": 2714697728
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000179197592778335,
      "loss": 1.3657,
      "theoretical_loss": 3.375099450766406,
      "tokens_seen": 2714763264
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017917753259779338,
      "loss": 1.351,
      "theoretical_loss": 3.3750931123751835,
      "tokens_seen": 2714828800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017915747241725174,
      "loss": 1.3502,
      "theoretical_loss": 3.375086774179809,
      "tokens_seen": 2714894336
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 986641,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2018595933914185,
      "objective/train/theoretical_loss": 3.375080436180273,
      "objective/train/tokens_used": 1085418976,
      "theoretical_loss": 3.375080436180273,
      "tokens_seen": 2714959872
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017913741223671013,
      "loss": 1.2518,
      "theoretical_loss": 3.375080436180273,
      "tokens_seen": 2714959872
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001791173520561685,
      "loss": 1.2654,
      "theoretical_loss": 3.3750740983765626,
      "tokens_seen": 2715025408
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017909729187562688,
      "loss": 1.3174,
      "theoretical_loss": 3.375067760768669,
      "tokens_seen": 2715090944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017907723169508524,
      "loss": 1.3371,
      "theoretical_loss": 3.37506142335658,
      "tokens_seen": 2715156480
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017905717151454363,
      "loss": 1.3157,
      "theoretical_loss": 3.3750550861402857,
      "tokens_seen": 2715222016
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017903711133400202,
      "loss": 1.14,
      "theoretical_loss": 3.375048749119775,
      "tokens_seen": 2715287552
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017901705115346038,
      "loss": 1.1996,
      "theoretical_loss": 3.3750424122950373,
      "tokens_seen": 2715353088
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017899699097291877,
      "loss": 1.2348,
      "theoretical_loss": 3.3750360756660616,
      "tokens_seen": 2715418624
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017897693079237714,
      "loss": 1.3652,
      "theoretical_loss": 3.3750297392328372,
      "tokens_seen": 2715484160
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017895687061183553,
      "loss": 1.357,
      "theoretical_loss": 3.375023402995353,
      "tokens_seen": 2715549696
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001789368104312939,
      "loss": 1.2461,
      "theoretical_loss": 3.3750170669535993,
      "tokens_seen": 2715615232
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017891675025075228,
      "loss": 1.1629,
      "theoretical_loss": 3.375010731107564,
      "tokens_seen": 2715680768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017889669007021064,
      "loss": 1.1867,
      "theoretical_loss": 3.375004395457237,
      "tokens_seen": 2715746304
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017887662988966903,
      "loss": 1.2205,
      "theoretical_loss": 3.3749980600026075,
      "tokens_seen": 2715811840
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001788565697091274,
      "loss": 1.3899,
      "theoretical_loss": 3.3749917247436647,
      "tokens_seen": 2715877376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017883650952858578,
      "loss": 1.3079,
      "theoretical_loss": 3.374985389680398,
      "tokens_seen": 2715942912
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017881644934804414,
      "loss": 1.2199,
      "theoretical_loss": 3.3749790548127963,
      "tokens_seen": 2716008448
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001787963891675025,
      "loss": 1.2889,
      "theoretical_loss": 3.3749727201408493,
      "tokens_seen": 2716073984
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001787763289869609,
      "loss": 1.2586,
      "theoretical_loss": 3.3749663856645453,
      "tokens_seen": 2716139520
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017875626880641926,
      "loss": 1.3,
      "theoretical_loss": 3.3749600513838747,
      "tokens_seen": 2716205056
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017873620862587765,
      "loss": 1.2731,
      "theoretical_loss": 3.374953717298826,
      "tokens_seen": 2716270592
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000178716148445336,
      "loss": 1.2522,
      "theoretical_loss": 3.374947383409389,
      "tokens_seen": 2716336128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001786960882647944,
      "loss": 1.1864,
      "theoretical_loss": 3.3749410497155523,
      "tokens_seen": 2716401664
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017867602808425276,
      "loss": 1.3336,
      "theoretical_loss": 3.374934716217305,
      "tokens_seen": 2716467200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017865596790371115,
      "loss": 1.2186,
      "theoretical_loss": 3.3749283829146375,
      "tokens_seen": 2716532736
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001786359077231695,
      "loss": 1.2264,
      "theoretical_loss": 3.374922049807538,
      "tokens_seen": 2716598272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017861584754262787,
      "loss": 1.3095,
      "theoretical_loss": 3.374915716895996,
      "tokens_seen": 2716663808
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017859578736208626,
      "loss": 1.2308,
      "theoretical_loss": 3.374909384180001,
      "tokens_seen": 2716729344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017857572718154462,
      "loss": 1.2784,
      "theoretical_loss": 3.374903051659542,
      "tokens_seen": 2716794880
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017855566700100301,
      "loss": 1.2729,
      "theoretical_loss": 3.374896719334608,
      "tokens_seen": 2716860416
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017853560682046138,
      "loss": 1.3034,
      "theoretical_loss": 3.374890387205189,
      "tokens_seen": 2716925952
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017851554663991977,
      "loss": 1.2636,
      "theoretical_loss": 3.3748840552712736,
      "tokens_seen": 2716991488
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017849548645937813,
      "loss": 1.3777,
      "theoretical_loss": 3.374877723532851,
      "tokens_seen": 2717057024
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017847542627883652,
      "loss": 1.3283,
      "theoretical_loss": 3.374871391989911,
      "tokens_seen": 2717122560
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017845536609829488,
      "loss": 1.2638,
      "theoretical_loss": 3.3748650606424424,
      "tokens_seen": 2717188096
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017843530591775324,
      "loss": 1.3265,
      "theoretical_loss": 3.3748587294904344,
      "tokens_seen": 2717253632
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017841524573721163,
      "loss": 1.2737,
      "theoretical_loss": 3.3748523985338768,
      "tokens_seen": 2717319168
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017839518555667,
      "loss": 1.2463,
      "theoretical_loss": 3.3748460677727583,
      "tokens_seen": 2717384704
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017837512537612838,
      "loss": 1.3187,
      "theoretical_loss": 3.3748397372070684,
      "tokens_seen": 2717450240
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017835506519558675,
      "loss": 1.2911,
      "theoretical_loss": 3.374833406836796,
      "tokens_seen": 2717515776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017833500501504513,
      "loss": 1.339,
      "theoretical_loss": 3.3748270766619313,
      "tokens_seen": 2717581312
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001783149448345035,
      "loss": 1.2923,
      "theoretical_loss": 3.3748207466824622,
      "tokens_seen": 2717646848
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001782948846539619,
      "loss": 1.279,
      "theoretical_loss": 3.374814416898379,
      "tokens_seen": 2717712384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017827482447342028,
      "loss": 1.1514,
      "theoretical_loss": 3.374808087309671,
      "tokens_seen": 2717777920
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017825476429287866,
      "loss": 1.2629,
      "theoretical_loss": 3.3748017579163267,
      "tokens_seen": 2717843456
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017823470411233703,
      "loss": 1.3609,
      "theoretical_loss": 3.3747954287183357,
      "tokens_seen": 2717908992
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001782146439317954,
      "loss": 1.3436,
      "theoretical_loss": 3.374789099715687,
      "tokens_seen": 2717974528
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017819458375125378,
      "loss": 1.2496,
      "theoretical_loss": 3.3747827709083706,
      "tokens_seen": 2718040064
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017817452357071214,
      "loss": 1.2843,
      "theoretical_loss": 3.3747764422963753,
      "tokens_seen": 2718105600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017815446339017053,
      "loss": 1.2061,
      "theoretical_loss": 3.3747701138796904,
      "tokens_seen": 2718171136
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 987833,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4549930095672607,
      "objective/train/theoretical_loss": 3.3747637856583053,
      "objective/train/tokens_used": 1088695776,
      "theoretical_loss": 3.3747637856583053,
      "tokens_seen": 2718236672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001781344032096289,
      "loss": 1.3678,
      "theoretical_loss": 3.3747637856583053,
      "tokens_seen": 2718236672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017811434302908728,
      "loss": 1.2925,
      "theoretical_loss": 3.374757457632209,
      "tokens_seen": 2718302208
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017809428284854564,
      "loss": 1.3231,
      "theoretical_loss": 3.3747511298013904,
      "tokens_seen": 2718367744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017807422266800403,
      "loss": 1.3365,
      "theoretical_loss": 3.3747448021658397,
      "tokens_seen": 2718433280
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001780541624874624,
      "loss": 1.2816,
      "theoretical_loss": 3.374738474725546,
      "tokens_seen": 2718498816
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017803410230692076,
      "loss": 1.2356,
      "theoretical_loss": 3.3747321474804974,
      "tokens_seen": 2718564352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017801404212637915,
      "loss": 1.1705,
      "theoretical_loss": 3.374725820430685,
      "tokens_seen": 2718629888
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001779939819458375,
      "loss": 1.288,
      "theoretical_loss": 3.3747194935760962,
      "tokens_seen": 2718695424
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001779739217652959,
      "loss": 1.2488,
      "theoretical_loss": 3.374713166916722,
      "tokens_seen": 2718760960
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017795386158475426,
      "loss": 1.2265,
      "theoretical_loss": 3.3747068404525504,
      "tokens_seen": 2718826496
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017793380140421265,
      "loss": 1.2195,
      "theoretical_loss": 3.3747005141835715,
      "tokens_seen": 2718892032
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000177913741223671,
      "loss": 1.3235,
      "theoretical_loss": 3.374694188109774,
      "tokens_seen": 2718957568
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001778936810431294,
      "loss": 1.3751,
      "theoretical_loss": 3.3746878622311476,
      "tokens_seen": 2719023104
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017787362086258776,
      "loss": 1.3078,
      "theoretical_loss": 3.3746815365476808,
      "tokens_seen": 2719088640
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017785356068204613,
      "loss": 1.2639,
      "theoretical_loss": 3.3746752110593636,
      "tokens_seen": 2719154176
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017783350050150452,
      "loss": 1.2618,
      "theoretical_loss": 3.3746688857661855,
      "tokens_seen": 2719219712
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017781344032096288,
      "loss": 1.258,
      "theoretical_loss": 3.3746625606681353,
      "tokens_seen": 2719285248
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017779338014042127,
      "loss": 1.3238,
      "theoretical_loss": 3.374656235765202,
      "tokens_seen": 2719350784
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017777331995987963,
      "loss": 1.2617,
      "theoretical_loss": 3.3746499110573756,
      "tokens_seen": 2719416320
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017775325977933802,
      "loss": 1.2599,
      "theoretical_loss": 3.3746435865446447,
      "tokens_seen": 2719481856
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017773319959879638,
      "loss": 1.2598,
      "theoretical_loss": 3.374637262226999,
      "tokens_seen": 2719547392
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017771313941825477,
      "loss": 1.3305,
      "theoretical_loss": 3.374630938104428,
      "tokens_seen": 2719612928
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017769307923771313,
      "loss": 1.3416,
      "theoretical_loss": 3.3746246141769207,
      "tokens_seen": 2719678464
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017767301905717152,
      "loss": 1.329,
      "theoretical_loss": 3.3746182904444657,
      "tokens_seen": 2719744000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017765295887662988,
      "loss": 1.3004,
      "theoretical_loss": 3.374611966907054,
      "tokens_seen": 2719809536
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017763289869608825,
      "loss": 1.2819,
      "theoretical_loss": 3.3746056435646725,
      "tokens_seen": 2719875072
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017761283851554664,
      "loss": 1.192,
      "theoretical_loss": 3.3745993204173126,
      "tokens_seen": 2719940608
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000177592778335005,
      "loss": 1.2836,
      "theoretical_loss": 3.374592997464963,
      "tokens_seen": 2720006144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001775727181544634,
      "loss": 1.3179,
      "theoretical_loss": 3.3745866747076123,
      "tokens_seen": 2720071680
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017755265797392175,
      "loss": 1.3395,
      "theoretical_loss": 3.37458035214525,
      "tokens_seen": 2720137216
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017753259779338017,
      "loss": 1.3639,
      "theoretical_loss": 3.3745740297778664,
      "tokens_seen": 2720202752
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017751253761283853,
      "loss": 1.3331,
      "theoretical_loss": 3.3745677076054497,
      "tokens_seen": 2720268288
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017749247743229692,
      "loss": 1.3166,
      "theoretical_loss": 3.3745613856279895,
      "tokens_seen": 2720333824
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017747241725175528,
      "loss": 1.3033,
      "theoretical_loss": 3.374555063845475,
      "tokens_seen": 2720399360
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017745235707121364,
      "loss": 1.2054,
      "theoretical_loss": 3.374548742257896,
      "tokens_seen": 2720464896
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017743229689067203,
      "loss": 1.3122,
      "theoretical_loss": 3.3745424208652413,
      "tokens_seen": 2720530432
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001774122367101304,
      "loss": 1.2983,
      "theoretical_loss": 3.3745360996675,
      "tokens_seen": 2720595968
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017739217652958878,
      "loss": 1.344,
      "theoretical_loss": 3.374529778664662,
      "tokens_seen": 2720661504
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017737211634904715,
      "loss": 1.1931,
      "theoretical_loss": 3.374523457856716,
      "tokens_seen": 2720727040
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017735205616850553,
      "loss": 1.308,
      "theoretical_loss": 3.374517137243652,
      "tokens_seen": 2720792576
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001773319959879639,
      "loss": 1.2847,
      "theoretical_loss": 3.3745108168254587,
      "tokens_seen": 2720858112
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017731193580742229,
      "loss": 1.4261,
      "theoretical_loss": 3.3745044966021256,
      "tokens_seen": 2720923648
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017729187562688065,
      "loss": 1.266,
      "theoretical_loss": 3.374498176573642,
      "tokens_seen": 2720989184
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000177271815446339,
      "loss": 1.4059,
      "theoretical_loss": 3.374491856739997,
      "tokens_seen": 2721054720
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001772517552657974,
      "loss": 1.1772,
      "theoretical_loss": 3.3744855371011804,
      "tokens_seen": 2721120256
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017723169508525576,
      "loss": 1.2322,
      "theoretical_loss": 3.3744792176571807,
      "tokens_seen": 2721185792
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017721163490471415,
      "loss": 1.2207,
      "theoretical_loss": 3.3744728984079884,
      "tokens_seen": 2721251328
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017719157472417251,
      "loss": 1.3456,
      "theoretical_loss": 3.3744665793535917,
      "tokens_seen": 2721316864
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001771715145436309,
      "loss": 1.368,
      "theoretical_loss": 3.37446026049398,
      "tokens_seen": 2721382400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017715145436308927,
      "loss": 1.283,
      "theoretical_loss": 3.3744539418291435,
      "tokens_seen": 2721447936
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 988960,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1977958679199219,
      "objective/train/theoretical_loss": 3.3744476233590706,
      "objective/train/tokens_used": 1091972576,
      "theoretical_loss": 3.3744476233590706,
      "tokens_seen": 2721513472
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017713139418254765,
      "loss": 1.1915,
      "theoretical_loss": 3.3744476233590706,
      "tokens_seen": 2721513472
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017711133400200602,
      "loss": 1.3639,
      "theoretical_loss": 3.374441305083751,
      "tokens_seen": 2721579008
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001770912738214644,
      "loss": 1.2568,
      "theoretical_loss": 3.3744349870031742,
      "tokens_seen": 2721644544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017707121364092277,
      "loss": 1.3649,
      "theoretical_loss": 3.3744286691173286,
      "tokens_seen": 2721710080
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017705115346038113,
      "loss": 1.321,
      "theoretical_loss": 3.3744223514262046,
      "tokens_seen": 2721775616
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017703109327983952,
      "loss": 1.3016,
      "theoretical_loss": 3.374416033929791,
      "tokens_seen": 2721841152
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017701103309929788,
      "loss": 1.2809,
      "theoretical_loss": 3.3744097166280773,
      "tokens_seen": 2721906688
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017699097291875627,
      "loss": 1.2408,
      "theoretical_loss": 3.3744033995210523,
      "tokens_seen": 2721972224
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017697091273821463,
      "loss": 1.285,
      "theoretical_loss": 3.374397082608706,
      "tokens_seen": 2722037760
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017695085255767302,
      "loss": 1.3707,
      "theoretical_loss": 3.3743907658910275,
      "tokens_seen": 2722103296
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017693079237713139,
      "loss": 1.3192,
      "theoretical_loss": 3.374384449368006,
      "tokens_seen": 2722168832
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017691073219658978,
      "loss": 1.3878,
      "theoretical_loss": 3.3743781330396305,
      "tokens_seen": 2722234368
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017689067201604814,
      "loss": 1.301,
      "theoretical_loss": 3.374371816905891,
      "tokens_seen": 2722299904
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001768706118355065,
      "loss": 1.2924,
      "theoretical_loss": 3.374365500966776,
      "tokens_seen": 2722365440
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001768505516549649,
      "loss": 1.1991,
      "theoretical_loss": 3.374359185222276,
      "tokens_seen": 2722430976
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017683049147442325,
      "loss": 1.3,
      "theoretical_loss": 3.374352869672379,
      "tokens_seen": 2722496512
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017681043129388164,
      "loss": 1.2906,
      "theoretical_loss": 3.374346554317075,
      "tokens_seen": 2722562048
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017679037111334003,
      "loss": 1.2052,
      "theoretical_loss": 3.3743402391563535,
      "tokens_seen": 2722627584
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017677031093279842,
      "loss": 1.2511,
      "theoretical_loss": 3.3743339241902035,
      "tokens_seen": 2722693120
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017675025075225678,
      "loss": 1.4108,
      "theoretical_loss": 3.3743276094186143,
      "tokens_seen": 2722758656
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017673019057171517,
      "loss": 1.4201,
      "theoretical_loss": 3.3743212948415753,
      "tokens_seen": 2722824192
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017671013039117353,
      "loss": 1.2491,
      "theoretical_loss": 3.3743149804590757,
      "tokens_seen": 2722889728
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017669007021063192,
      "loss": 1.2665,
      "theoretical_loss": 3.374308666271105,
      "tokens_seen": 2722955264
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017667001003009028,
      "loss": 1.2304,
      "theoretical_loss": 3.374302352277653,
      "tokens_seen": 2723020800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017664994984954865,
      "loss": 1.1812,
      "theoretical_loss": 3.374296038478708,
      "tokens_seen": 2723086336
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017662988966900704,
      "loss": 1.3866,
      "theoretical_loss": 3.37428972487426,
      "tokens_seen": 2723151872
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001766098294884654,
      "loss": 1.2546,
      "theoretical_loss": 3.3742834114642983,
      "tokens_seen": 2723217408
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001765897693079238,
      "loss": 1.2385,
      "theoretical_loss": 3.374277098248812,
      "tokens_seen": 2723282944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017656970912738215,
      "loss": 1.283,
      "theoretical_loss": 3.3742707852277904,
      "tokens_seen": 2723348480
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017654964894684054,
      "loss": 1.2118,
      "theoretical_loss": 3.3742644724012227,
      "tokens_seen": 2723414016
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001765295887662989,
      "loss": 1.1496,
      "theoretical_loss": 3.3742581597690986,
      "tokens_seen": 2723479552
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001765095285857573,
      "loss": 1.1907,
      "theoretical_loss": 3.374251847331408,
      "tokens_seen": 2723545088
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017648946840521565,
      "loss": 1.3318,
      "theoretical_loss": 3.374245535088139,
      "tokens_seen": 2723610624
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017646940822467402,
      "loss": 1.3166,
      "theoretical_loss": 3.3742392230392815,
      "tokens_seen": 2723676160
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001764493480441324,
      "loss": 1.3096,
      "theoretical_loss": 3.3742329111848246,
      "tokens_seen": 2723741696
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017642928786359077,
      "loss": 1.3086,
      "theoretical_loss": 3.3742265995247585,
      "tokens_seen": 2723807232
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017640922768304916,
      "loss": 1.2076,
      "theoretical_loss": 3.3742202880590715,
      "tokens_seen": 2723872768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017638916750250752,
      "loss": 1.3116,
      "theoretical_loss": 3.3742139767877535,
      "tokens_seen": 2723938304
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001763691073219659,
      "loss": 1.2382,
      "theoretical_loss": 3.3742076657107933,
      "tokens_seen": 2724003840
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017634904714142427,
      "loss": 1.3277,
      "theoretical_loss": 3.374201354828181,
      "tokens_seen": 2724069376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017632898696088266,
      "loss": 1.3191,
      "theoretical_loss": 3.3741950441399053,
      "tokens_seen": 2724134912
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017630892678034102,
      "loss": 1.348,
      "theoretical_loss": 3.3741887336459557,
      "tokens_seen": 2724200448
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017628886659979938,
      "loss": 1.3347,
      "theoretical_loss": 3.374182423346322,
      "tokens_seen": 2724265984
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017626880641925777,
      "loss": 1.3784,
      "theoretical_loss": 3.374176113240993,
      "tokens_seen": 2724331520
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017624874623871614,
      "loss": 1.3103,
      "theoretical_loss": 3.3741698033299583,
      "tokens_seen": 2724397056
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017622868605817452,
      "loss": 1.3243,
      "theoretical_loss": 3.374163493613207,
      "tokens_seen": 2724462592
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001762086258776329,
      "loss": 1.226,
      "theoretical_loss": 3.3741571840907287,
      "tokens_seen": 2724528128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017618856569709128,
      "loss": 1.2321,
      "theoretical_loss": 3.3741508747625124,
      "tokens_seen": 2724593664
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017616850551654964,
      "loss": 1.288,
      "theoretical_loss": 3.374144565628548,
      "tokens_seen": 2724659200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017614844533600803,
      "loss": 1.3012,
      "theoretical_loss": 3.3741382566888243,
      "tokens_seen": 2724724736
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 990115,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3105227947235107,
      "objective/train/theoretical_loss": 3.374131947943331,
      "objective/train/tokens_used": 1095249376,
      "theoretical_loss": 3.374131947943331,
      "tokens_seen": 2724790272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001761283851554664,
      "loss": 1.33,
      "theoretical_loss": 3.374131947943331,
      "tokens_seen": 2724790272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017610832497492478,
      "loss": 1.2385,
      "theoretical_loss": 3.3741256393920573,
      "tokens_seen": 2724855808
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017608826479438314,
      "loss": 1.2984,
      "theoretical_loss": 3.3741193310349926,
      "tokens_seen": 2724921344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001760682046138415,
      "loss": 1.1929,
      "theoretical_loss": 3.3741130228721263,
      "tokens_seen": 2724986880
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017604814443329992,
      "loss": 1.2643,
      "theoretical_loss": 3.374106714903448,
      "tokens_seen": 2725052416
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017602808425275828,
      "loss": 1.2704,
      "theoretical_loss": 3.3741004071289464,
      "tokens_seen": 2725117952
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017600802407221667,
      "loss": 1.3131,
      "theoretical_loss": 3.374094099548611,
      "tokens_seen": 2725183488
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017598796389167503,
      "loss": 1.2973,
      "theoretical_loss": 3.374087792162432,
      "tokens_seen": 2725249024
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017596790371113342,
      "loss": 1.2333,
      "theoretical_loss": 3.3740814849703975,
      "tokens_seen": 2725314560
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017594784353059179,
      "loss": 1.2635,
      "theoretical_loss": 3.3740751779724976,
      "tokens_seen": 2725380096
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017592778335005018,
      "loss": 1.316,
      "theoretical_loss": 3.3740688711687215,
      "tokens_seen": 2725445632
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017590772316950854,
      "loss": 1.3542,
      "theoretical_loss": 3.3740625645590585,
      "tokens_seen": 2725511168
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001758876629889669,
      "loss": 1.3193,
      "theoretical_loss": 3.374056258143498,
      "tokens_seen": 2725576704
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001758676028084253,
      "loss": 1.2329,
      "theoretical_loss": 3.37404995192203,
      "tokens_seen": 2725642240
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017584754262788365,
      "loss": 1.2795,
      "theoretical_loss": 3.3740436458946426,
      "tokens_seen": 2725707776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017582748244734204,
      "loss": 1.3081,
      "theoretical_loss": 3.374037340061326,
      "tokens_seen": 2725773312
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001758074222668004,
      "loss": 1.333,
      "theoretical_loss": 3.3740310344220696,
      "tokens_seen": 2725838848
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001757873620862588,
      "loss": 1.1846,
      "theoretical_loss": 3.3740247289768623,
      "tokens_seen": 2725904384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017576730190571715,
      "loss": 1.224,
      "theoretical_loss": 3.3740184237256936,
      "tokens_seen": 2725969920
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017574724172517554,
      "loss": 1.3963,
      "theoretical_loss": 3.374012118668553,
      "tokens_seen": 2726035456
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001757271815446339,
      "loss": 1.29,
      "theoretical_loss": 3.37400581380543,
      "tokens_seen": 2726100992
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017570712136409227,
      "loss": 1.2437,
      "theoretical_loss": 3.373999509136314,
      "tokens_seen": 2726166528
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017568706118355066,
      "loss": 1.3018,
      "theoretical_loss": 3.373993204661194,
      "tokens_seen": 2726232064
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017566700100300902,
      "loss": 1.2156,
      "theoretical_loss": 3.373986900380059,
      "tokens_seen": 2726297600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001756469408224674,
      "loss": 1.177,
      "theoretical_loss": 3.3739805962928995,
      "tokens_seen": 2726363136
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017562688064192577,
      "loss": 1.3073,
      "theoretical_loss": 3.373974292399704,
      "tokens_seen": 2726428672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017560682046138416,
      "loss": 1.3015,
      "theoretical_loss": 3.3739679887004623,
      "tokens_seen": 2726494208
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017558676028084252,
      "loss": 1.2945,
      "theoretical_loss": 3.3739616851951633,
      "tokens_seen": 2726559744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001755667001003009,
      "loss": 1.2696,
      "theoretical_loss": 3.373955381883797,
      "tokens_seen": 2726625280
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017554663991975927,
      "loss": 1.321,
      "theoretical_loss": 3.373949078766352,
      "tokens_seen": 2726690816
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017552657973921766,
      "loss": 1.3923,
      "theoretical_loss": 3.373942775842819,
      "tokens_seen": 2726756352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017550651955867603,
      "loss": 1.2874,
      "theoretical_loss": 3.3739364731131856,
      "tokens_seen": 2726821888
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001754864593781344,
      "loss": 1.2264,
      "theoretical_loss": 3.3739301705774425,
      "tokens_seen": 2726887424
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017546639919759278,
      "loss": 1.2742,
      "theoretical_loss": 3.373923868235579,
      "tokens_seen": 2726952960
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017544633901705114,
      "loss": 1.2955,
      "theoretical_loss": 3.3739175660875835,
      "tokens_seen": 2727018496
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017542627883650953,
      "loss": 1.2603,
      "theoretical_loss": 3.3739112641334463,
      "tokens_seen": 2727084032
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001754062186559679,
      "loss": 1.2998,
      "theoretical_loss": 3.3739049623731563,
      "tokens_seen": 2727149568
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017538615847542628,
      "loss": 1.1594,
      "theoretical_loss": 3.373898660806703,
      "tokens_seen": 2727215104
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017536609829488464,
      "loss": 1.236,
      "theoretical_loss": 3.373892359434076,
      "tokens_seen": 2727280640
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017534603811434303,
      "loss": 1.3297,
      "theoretical_loss": 3.3738860582552648,
      "tokens_seen": 2727346176
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001753259779338014,
      "loss": 1.238,
      "theoretical_loss": 3.373879757270258,
      "tokens_seen": 2727411712
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017530591775325976,
      "loss": 1.3286,
      "theoretical_loss": 3.373873456479046,
      "tokens_seen": 2727477248
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017528585757271817,
      "loss": 1.2257,
      "theoretical_loss": 3.373867155881617,
      "tokens_seen": 2727542784
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017526579739217654,
      "loss": 1.3354,
      "theoretical_loss": 3.3738608554779614,
      "tokens_seen": 2727608320
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017524573721163493,
      "loss": 1.2808,
      "theoretical_loss": 3.3738545552680685,
      "tokens_seen": 2727673856
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001752256770310933,
      "loss": 1.1986,
      "theoretical_loss": 3.373848255251927,
      "tokens_seen": 2727739392
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017520561685055168,
      "loss": 1.31,
      "theoretical_loss": 3.373841955429527,
      "tokens_seen": 2727804928
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017518555667001004,
      "loss": 1.2716,
      "theoretical_loss": 3.3738356558008578,
      "tokens_seen": 2727870464
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017516549648946843,
      "loss": 1.2646,
      "theoretical_loss": 3.373829356365908,
      "tokens_seen": 2727936000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001751454363089268,
      "loss": 1.2781,
      "theoretical_loss": 3.373823057124668,
      "tokens_seen": 2728001536
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 991320,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.324868083000183,
      "objective/train/theoretical_loss": 3.3738167580771266,
      "objective/train/tokens_used": 1098526176,
      "theoretical_loss": 3.3738167580771266,
      "tokens_seen": 2728067072
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017512537612838518,
      "loss": 1.3126,
      "theoretical_loss": 3.3738167580771266,
      "tokens_seen": 2728067072
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017510531594784354,
      "loss": 1.289,
      "theoretical_loss": 3.3738104592232734,
      "tokens_seen": 2728132608
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001750852557673019,
      "loss": 1.2124,
      "theoretical_loss": 3.3738041605630977,
      "tokens_seen": 2728198144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001750651955867603,
      "loss": 1.2613,
      "theoretical_loss": 3.373797862096589,
      "tokens_seen": 2728263680
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017504513540621866,
      "loss": 1.272,
      "theoretical_loss": 3.3737915638237364,
      "tokens_seen": 2728329216
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017502507522567705,
      "loss": 1.2761,
      "theoretical_loss": 3.3737852657445297,
      "tokens_seen": 2728394752
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001750050150451354,
      "loss": 1.4161,
      "theoretical_loss": 3.373778967858958,
      "tokens_seen": 2728460288
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001749849548645938,
      "loss": 1.2137,
      "theoretical_loss": 3.3737726701670114,
      "tokens_seen": 2728525824
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017496489468405216,
      "loss": 1.3148,
      "theoretical_loss": 3.373766372668678,
      "tokens_seen": 2728591360
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017494483450351055,
      "loss": 1.2861,
      "theoretical_loss": 3.3737600753639483,
      "tokens_seen": 2728656896
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001749247743229689,
      "loss": 1.2961,
      "theoretical_loss": 3.3737537782528113,
      "tokens_seen": 2728722432
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017490471414242727,
      "loss": 1.4394,
      "theoretical_loss": 3.3737474813352564,
      "tokens_seen": 2728787968
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017488465396188566,
      "loss": 1.2621,
      "theoretical_loss": 3.373741184611273,
      "tokens_seen": 2728853504
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017486459378134402,
      "loss": 1.3395,
      "theoretical_loss": 3.3737348880808504,
      "tokens_seen": 2728919040
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017484453360080241,
      "loss": 1.2651,
      "theoretical_loss": 3.373728591743978,
      "tokens_seen": 2728984576
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017482447342026078,
      "loss": 1.3592,
      "theoretical_loss": 3.3737222956006456,
      "tokens_seen": 2729050112
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017480441323971917,
      "loss": 1.3493,
      "theoretical_loss": 3.373715999650842,
      "tokens_seen": 2729115648
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017478435305917753,
      "loss": 1.3399,
      "theoretical_loss": 3.3737097038945576,
      "tokens_seen": 2729181184
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017476429287863592,
      "loss": 1.1831,
      "theoretical_loss": 3.3737034083317803,
      "tokens_seen": 2729246720
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017474423269809428,
      "loss": 1.2721,
      "theoretical_loss": 3.3736971129625006,
      "tokens_seen": 2729312256
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017472417251755264,
      "loss": 1.3091,
      "theoretical_loss": 3.373690817786708,
      "tokens_seen": 2729377792
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017470411233701103,
      "loss": 1.3027,
      "theoretical_loss": 3.3736845228043912,
      "tokens_seen": 2729443328
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001746840521564694,
      "loss": 1.2578,
      "theoretical_loss": 3.37367822801554,
      "tokens_seen": 2729508864
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017466399197592778,
      "loss": 1.2958,
      "theoretical_loss": 3.3736719334201437,
      "tokens_seen": 2729574400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017464393179538614,
      "loss": 1.2738,
      "theoretical_loss": 3.373665639018192,
      "tokens_seen": 2729639936
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017462387161484453,
      "loss": 1.2351,
      "theoretical_loss": 3.373659344809674,
      "tokens_seen": 2729705472
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001746038114343029,
      "loss": 1.2604,
      "theoretical_loss": 3.3736530507945792,
      "tokens_seen": 2729771008
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017458375125376129,
      "loss": 1.3026,
      "theoretical_loss": 3.373646756972897,
      "tokens_seen": 2729836544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017456369107321965,
      "loss": 1.2746,
      "theoretical_loss": 3.373640463344617,
      "tokens_seen": 2729902080
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017454363089267806,
      "loss": 1.2176,
      "theoretical_loss": 3.373634169909728,
      "tokens_seen": 2729967616
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017452357071213643,
      "loss": 1.3601,
      "theoretical_loss": 3.37362787666822,
      "tokens_seen": 2730033152
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001745035105315948,
      "loss": 1.2489,
      "theoretical_loss": 3.3736215836200825,
      "tokens_seen": 2730098688
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017448345035105318,
      "loss": 1.3858,
      "theoretical_loss": 3.3736152907653043,
      "tokens_seen": 2730164224
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017446339017051154,
      "loss": 1.3124,
      "theoretical_loss": 3.373608998103876,
      "tokens_seen": 2730229760
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017444332998996993,
      "loss": 1.2893,
      "theoretical_loss": 3.3736027056357853,
      "tokens_seen": 2730295296
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001744232698094283,
      "loss": 1.2237,
      "theoretical_loss": 3.3735964133610232,
      "tokens_seen": 2730360832
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017440320962888668,
      "loss": 1.3181,
      "theoretical_loss": 3.373590121279578,
      "tokens_seen": 2730426368
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017438314944834504,
      "loss": 1.3088,
      "theoretical_loss": 3.37358382939144,
      "tokens_seen": 2730491904
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017436308926780343,
      "loss": 1.3373,
      "theoretical_loss": 3.373577537696598,
      "tokens_seen": 2730557440
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001743430290872618,
      "loss": 1.2452,
      "theoretical_loss": 3.3735712461950413,
      "tokens_seen": 2730622976
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017432296890672016,
      "loss": 1.2236,
      "theoretical_loss": 3.37356495488676,
      "tokens_seen": 2730688512
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017430290872617855,
      "loss": 1.1967,
      "theoretical_loss": 3.3735586637717434,
      "tokens_seen": 2730754048
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001742828485456369,
      "loss": 1.352,
      "theoretical_loss": 3.3735523728499803,
      "tokens_seen": 2730819584
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001742627883650953,
      "loss": 1.2489,
      "theoretical_loss": 3.373546082121461,
      "tokens_seen": 2730885120
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017424272818455366,
      "loss": 1.2995,
      "theoretical_loss": 3.373539791586174,
      "tokens_seen": 2730950656
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017422266800401205,
      "loss": 1.1804,
      "theoretical_loss": 3.373533501244109,
      "tokens_seen": 2731016192
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001742026078234704,
      "loss": 1.3571,
      "theoretical_loss": 3.373527211095256,
      "tokens_seen": 2731081728
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001741825476429288,
      "loss": 1.283,
      "theoretical_loss": 3.373520921139604,
      "tokens_seen": 2731147264
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017416248746238716,
      "loss": 1.2738,
      "theoretical_loss": 3.3735146313771427,
      "tokens_seen": 2731212800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017414242728184553,
      "loss": 1.2861,
      "theoretical_loss": 3.373508341807861,
      "tokens_seen": 2731278336
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 992345,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4288108348846436,
      "objective/train/theoretical_loss": 3.3735020524317485,
      "objective/train/tokens_used": 1101802976,
      "theoretical_loss": 3.3735020524317485,
      "tokens_seen": 2731343872
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017412236710130392,
      "loss": 1.3499,
      "theoretical_loss": 3.3735020524317485,
      "tokens_seen": 2731343872
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017410230692076228,
      "loss": 1.2654,
      "theoretical_loss": 3.373495763248795,
      "tokens_seen": 2731409408
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017408224674022067,
      "loss": 1.3258,
      "theoretical_loss": 3.3734894742589896,
      "tokens_seen": 2731474944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017406218655967903,
      "loss": 1.28,
      "theoretical_loss": 3.3734831854623217,
      "tokens_seen": 2731540480
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017404212637913742,
      "loss": 1.26,
      "theoretical_loss": 3.373476896858781,
      "tokens_seen": 2731606016
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017402206619859578,
      "loss": 1.3133,
      "theoretical_loss": 3.3734706084483568,
      "tokens_seen": 2731671552
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017400200601805417,
      "loss": 1.3426,
      "theoretical_loss": 3.3734643202310384,
      "tokens_seen": 2731737088
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017398194583751253,
      "loss": 1.1334,
      "theoretical_loss": 3.3734580322068157,
      "tokens_seen": 2731802624
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017396188565697092,
      "loss": 1.3165,
      "theoretical_loss": 3.3734517443756777,
      "tokens_seen": 2731868160
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017394182547642928,
      "loss": 1.247,
      "theoretical_loss": 3.3734454567376138,
      "tokens_seen": 2731933696
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017392176529588765,
      "loss": 1.2913,
      "theoretical_loss": 3.373439169292614,
      "tokens_seen": 2731999232
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017390170511534604,
      "loss": 1.3227,
      "theoretical_loss": 3.373432882040667,
      "tokens_seen": 2732064768
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001738816449348044,
      "loss": 1.3313,
      "theoretical_loss": 3.3734265949817623,
      "tokens_seen": 2732130304
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001738615847542628,
      "loss": 1.3121,
      "theoretical_loss": 3.37342030811589,
      "tokens_seen": 2732195840
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017384152457372115,
      "loss": 1.27,
      "theoretical_loss": 3.3734140214430393,
      "tokens_seen": 2732261376
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017382146439317954,
      "loss": 1.3096,
      "theoretical_loss": 3.373407734963199,
      "tokens_seen": 2732326912
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001738014042126379,
      "loss": 1.2898,
      "theoretical_loss": 3.3734014486763595,
      "tokens_seen": 2732392448
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017378134403209632,
      "loss": 1.2323,
      "theoretical_loss": 3.3733951625825096,
      "tokens_seen": 2732457984
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017376128385155468,
      "loss": 1.3225,
      "theoretical_loss": 3.373388876681639,
      "tokens_seen": 2732523520
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017374122367101304,
      "loss": 1.3805,
      "theoretical_loss": 3.373382590973737,
      "tokens_seen": 2732589056
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017372116349047143,
      "loss": 1.2703,
      "theoretical_loss": 3.3733763054587933,
      "tokens_seen": 2732654592
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001737011033099298,
      "loss": 1.3053,
      "theoretical_loss": 3.373370020136797,
      "tokens_seen": 2732720128
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017368104312938818,
      "loss": 1.3391,
      "theoretical_loss": 3.373363735007738,
      "tokens_seen": 2732785664
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017366098294884655,
      "loss": 1.2892,
      "theoretical_loss": 3.3733574500716053,
      "tokens_seen": 2732851200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017364092276830493,
      "loss": 1.2695,
      "theoretical_loss": 3.3733511653283887,
      "tokens_seen": 2732916736
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001736208625877633,
      "loss": 1.2421,
      "theoretical_loss": 3.3733448807780775,
      "tokens_seen": 2732982272
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017360080240722169,
      "loss": 1.4435,
      "theoretical_loss": 3.373338596420661,
      "tokens_seen": 2733047808
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017358074222668005,
      "loss": 1.2732,
      "theoretical_loss": 3.373332312256129,
      "tokens_seen": 2733113344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001735606820461384,
      "loss": 1.2349,
      "theoretical_loss": 3.3733260282844704,
      "tokens_seen": 2733178880
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001735406218655968,
      "loss": 1.3886,
      "theoretical_loss": 3.3733197445056753,
      "tokens_seen": 2733244416
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017352056168505516,
      "loss": 1.2251,
      "theoretical_loss": 3.3733134609197326,
      "tokens_seen": 2733309952
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017350050150451355,
      "loss": 1.2146,
      "theoretical_loss": 3.373307177526632,
      "tokens_seen": 2733375488
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017348044132397191,
      "loss": 1.1172,
      "theoretical_loss": 3.3733008943263636,
      "tokens_seen": 2733441024
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001734603811434303,
      "loss": 1.3583,
      "theoretical_loss": 3.3732946113189155,
      "tokens_seen": 2733506560
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017344032096288867,
      "loss": 1.3257,
      "theoretical_loss": 3.3732883285042785,
      "tokens_seen": 2733572096
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017342026078234705,
      "loss": 1.3769,
      "theoretical_loss": 3.373282045882441,
      "tokens_seen": 2733637632
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017340020060180542,
      "loss": 1.2443,
      "theoretical_loss": 3.3732757634533934,
      "tokens_seen": 2733703168
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001733801404212638,
      "loss": 1.2918,
      "theoretical_loss": 3.373269481217124,
      "tokens_seen": 2733768704
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017336008024072217,
      "loss": 1.2832,
      "theoretical_loss": 3.3732631991736235,
      "tokens_seen": 2733834240
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017334002006018053,
      "loss": 1.3632,
      "theoretical_loss": 3.373256917322881,
      "tokens_seen": 2733899776
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017331995987963892,
      "loss": 1.2299,
      "theoretical_loss": 3.3732506356648853,
      "tokens_seen": 2733965312
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017329989969909728,
      "loss": 1.3592,
      "theoretical_loss": 3.3732443541996266,
      "tokens_seen": 2734030848
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017327983951855567,
      "loss": 1.3288,
      "theoretical_loss": 3.373238072927094,
      "tokens_seen": 2734096384
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017325977933801403,
      "loss": 1.2295,
      "theoretical_loss": 3.3732317918472767,
      "tokens_seen": 2734161920
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017323971915747242,
      "loss": 1.2117,
      "theoretical_loss": 3.3732255109601654,
      "tokens_seen": 2734227456
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017321965897693079,
      "loss": 1.2221,
      "theoretical_loss": 3.373219230265748,
      "tokens_seen": 2734292992
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017319959879638917,
      "loss": 1.3648,
      "theoretical_loss": 3.373212949764015,
      "tokens_seen": 2734358528
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017317953861584754,
      "loss": 1.3264,
      "theoretical_loss": 3.3732066694549556,
      "tokens_seen": 2734424064
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001731594784353059,
      "loss": 1.3121,
      "theoretical_loss": 3.373200389338559,
      "tokens_seen": 2734489600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001731394182547643,
      "loss": 1.372,
      "theoretical_loss": 3.3731941094148152,
      "tokens_seen": 2734555136
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 993449,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1959069967269897,
      "objective/train/theoretical_loss": 3.373187829683713,
      "objective/train/tokens_used": 1105079776,
      "theoretical_loss": 3.373187829683713,
      "tokens_seen": 2734620672
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017311935807422265,
      "loss": 1.2498,
      "theoretical_loss": 3.373187829683713,
      "tokens_seen": 2734620672
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017309929789368104,
      "loss": 1.2493,
      "theoretical_loss": 3.373181550145243,
      "tokens_seen": 2734686208
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001730792377131394,
      "loss": 1.2668,
      "theoretical_loss": 3.373175270799393,
      "tokens_seen": 2734751744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001730591775325978,
      "loss": 1.1627,
      "theoretical_loss": 3.373168991646154,
      "tokens_seen": 2734817280
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017303911735205618,
      "loss": 1.2443,
      "theoretical_loss": 3.3731627126855144,
      "tokens_seen": 2734882816
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017301905717151457,
      "loss": 1.3004,
      "theoretical_loss": 3.3731564339174644,
      "tokens_seen": 2734948352
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017299899699097293,
      "loss": 1.2801,
      "theoretical_loss": 3.3731501553419934,
      "tokens_seen": 2735013888
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017297893681043132,
      "loss": 1.2171,
      "theoretical_loss": 3.3731438769590905,
      "tokens_seen": 2735079424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017295887662988968,
      "loss": 1.1735,
      "theoretical_loss": 3.3731375987687455,
      "tokens_seen": 2735144960
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017293881644934805,
      "loss": 1.3391,
      "theoretical_loss": 3.373131320770948,
      "tokens_seen": 2735210496
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017291875626880644,
      "loss": 1.1408,
      "theoretical_loss": 3.373125042965687,
      "tokens_seen": 2735276032
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001728986960882648,
      "loss": 1.3259,
      "theoretical_loss": 3.373118765352952,
      "tokens_seen": 2735341568
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001728786359077232,
      "loss": 1.2743,
      "theoretical_loss": 3.373112487932733,
      "tokens_seen": 2735407104
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017285857572718155,
      "loss": 1.2686,
      "theoretical_loss": 3.373106210705019,
      "tokens_seen": 2735472640
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017283851554663994,
      "loss": 1.2688,
      "theoretical_loss": 3.3730999336698,
      "tokens_seen": 2735538176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001728184553660983,
      "loss": 1.2706,
      "theoretical_loss": 3.373093656827065,
      "tokens_seen": 2735603712
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001727983951855567,
      "loss": 1.2556,
      "theoretical_loss": 3.3730873801768038,
      "tokens_seen": 2735669248
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017277833500501505,
      "loss": 1.2062,
      "theoretical_loss": 3.3730811037190054,
      "tokens_seen": 2735734784
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017275827482447342,
      "loss": 1.2474,
      "theoretical_loss": 3.37307482745366,
      "tokens_seen": 2735800320
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001727382146439318,
      "loss": 1.4083,
      "theoretical_loss": 3.3730685513807566,
      "tokens_seen": 2735865856
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017271815446339017,
      "loss": 1.2722,
      "theoretical_loss": 3.373062275500285,
      "tokens_seen": 2735931392
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017269809428284856,
      "loss": 1.1815,
      "theoretical_loss": 3.3730559998122343,
      "tokens_seen": 2735996928
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017267803410230692,
      "loss": 1.2276,
      "theoretical_loss": 3.3730497243165942,
      "tokens_seen": 2736062464
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001726579739217653,
      "loss": 1.2762,
      "theoretical_loss": 3.373043449013354,
      "tokens_seen": 2736128000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017263791374122367,
      "loss": 1.3627,
      "theoretical_loss": 3.3730371739025036,
      "tokens_seen": 2736193536
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017261785356068206,
      "loss": 1.2273,
      "theoretical_loss": 3.373030898984032,
      "tokens_seen": 2736259072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017259779338014042,
      "loss": 1.2698,
      "theoretical_loss": 3.3730246242579294,
      "tokens_seen": 2736324608
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017257773319959878,
      "loss": 1.3509,
      "theoretical_loss": 3.373018349724185,
      "tokens_seen": 2736390144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017255767301905717,
      "loss": 1.1038,
      "theoretical_loss": 3.3730120753827877,
      "tokens_seen": 2736455680
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017253761283851554,
      "loss": 1.2128,
      "theoretical_loss": 3.3730058012337274,
      "tokens_seen": 2736521216
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017251755265797392,
      "loss": 1.3022,
      "theoretical_loss": 3.3729995272769937,
      "tokens_seen": 2736586752
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001724974924774323,
      "loss": 1.2019,
      "theoretical_loss": 3.3729932535125764,
      "tokens_seen": 2736652288
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017247743229689068,
      "loss": 1.2342,
      "theoretical_loss": 3.3729869799404644,
      "tokens_seen": 2736717824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017245737211634904,
      "loss": 1.2242,
      "theoretical_loss": 3.372980706560648,
      "tokens_seen": 2736783360
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017243731193580743,
      "loss": 1.3011,
      "theoretical_loss": 3.3729744333731153,
      "tokens_seen": 2736848896
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001724172517552658,
      "loss": 1.3484,
      "theoretical_loss": 3.372968160377857,
      "tokens_seen": 2736914432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017239719157472418,
      "loss": 1.3555,
      "theoretical_loss": 3.3729618875748626,
      "tokens_seen": 2736979968
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017237713139418254,
      "loss": 1.2507,
      "theoretical_loss": 3.3729556149641207,
      "tokens_seen": 2737045504
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001723570712136409,
      "loss": 1.3588,
      "theoretical_loss": 3.372949342545622,
      "tokens_seen": 2737111040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001723370110330993,
      "loss": 1.2777,
      "theoretical_loss": 3.372943070319355,
      "tokens_seen": 2737176576
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017231695085255766,
      "loss": 1.3673,
      "theoretical_loss": 3.3729367982853096,
      "tokens_seen": 2737242112
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017229689067201607,
      "loss": 1.2261,
      "theoretical_loss": 3.3729305264434752,
      "tokens_seen": 2737307648
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017227683049147443,
      "loss": 1.3322,
      "theoretical_loss": 3.3729242547938414,
      "tokens_seen": 2737373184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017225677031093282,
      "loss": 1.4461,
      "theoretical_loss": 3.372917983336398,
      "tokens_seen": 2737438720
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017223671013039119,
      "loss": 1.388,
      "theoretical_loss": 3.372911712071134,
      "tokens_seen": 2737504256
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017221664994984958,
      "loss": 1.2688,
      "theoretical_loss": 3.372905440998039,
      "tokens_seen": 2737569792
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017219658976930794,
      "loss": 1.2697,
      "theoretical_loss": 3.3728991701171025,
      "tokens_seen": 2737635328
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001721765295887663,
      "loss": 1.2351,
      "theoretical_loss": 3.3728928994283147,
      "tokens_seen": 2737700864
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001721564694082247,
      "loss": 1.2055,
      "theoretical_loss": 3.3728866289316644,
      "tokens_seen": 2737766400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017213640922768305,
      "loss": 1.2692,
      "theoretical_loss": 3.3728803586271408,
      "tokens_seen": 2737831936
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 994721,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2009656429290771,
      "objective/train/theoretical_loss": 3.3728740885147346,
      "objective/train/tokens_used": 1108356576,
      "theoretical_loss": 3.3728740885147346,
      "tokens_seen": 2737897472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017211634904714144,
      "loss": 1.1759,
      "theoretical_loss": 3.3728740885147346,
      "tokens_seen": 2737897472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001720962888665998,
      "loss": 1.2683,
      "theoretical_loss": 3.3728678185944343,
      "tokens_seen": 2737963008
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001720762286860582,
      "loss": 1.311,
      "theoretical_loss": 3.3728615488662292,
      "tokens_seen": 2738028544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017205616850551655,
      "loss": 1.2569,
      "theoretical_loss": 3.37285527933011,
      "tokens_seen": 2738094080
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017203610832497494,
      "loss": 1.2878,
      "theoretical_loss": 3.3728490099860653,
      "tokens_seen": 2738159616
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001720160481444333,
      "loss": 1.2336,
      "theoretical_loss": 3.372842740834085,
      "tokens_seen": 2738225152
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017199598796389167,
      "loss": 1.3294,
      "theoretical_loss": 3.3728364718741584,
      "tokens_seen": 2738290688
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017197592778335006,
      "loss": 1.1971,
      "theoretical_loss": 3.3728302031062753,
      "tokens_seen": 2738356224
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017195586760280842,
      "loss": 1.2158,
      "theoretical_loss": 3.372823934530425,
      "tokens_seen": 2738421760
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001719358074222668,
      "loss": 1.2721,
      "theoretical_loss": 3.372817666146597,
      "tokens_seen": 2738487296
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017191574724172517,
      "loss": 1.2244,
      "theoretical_loss": 3.3728113979547807,
      "tokens_seen": 2738552832
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017189568706118356,
      "loss": 1.24,
      "theoretical_loss": 3.3728051299549655,
      "tokens_seen": 2738618368
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017187562688064192,
      "loss": 1.1594,
      "theoretical_loss": 3.3727988621471416,
      "tokens_seen": 2738683904
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001718555667001003,
      "loss": 1.3453,
      "theoretical_loss": 3.3727925945312984,
      "tokens_seen": 2738749440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017183550651955867,
      "loss": 1.3149,
      "theoretical_loss": 3.372786327107425,
      "tokens_seen": 2738814976
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017181544633901706,
      "loss": 1.2239,
      "theoretical_loss": 3.3727800598755113,
      "tokens_seen": 2738880512
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017179538615847543,
      "loss": 1.3059,
      "theoretical_loss": 3.372773792835546,
      "tokens_seen": 2738946048
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001717753259779338,
      "loss": 1.2291,
      "theoretical_loss": 3.37276752598752,
      "tokens_seen": 2739011584
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017175526579739218,
      "loss": 1.1973,
      "theoretical_loss": 3.372761259331422,
      "tokens_seen": 2739077120
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017173520561685054,
      "loss": 1.3296,
      "theoretical_loss": 3.3727549928672413,
      "tokens_seen": 2739142656
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017171514543630893,
      "loss": 1.3692,
      "theoretical_loss": 3.372748726594968,
      "tokens_seen": 2739208192
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001716950852557673,
      "loss": 1.2796,
      "theoretical_loss": 3.372742460514591,
      "tokens_seen": 2739273728
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017167502507522568,
      "loss": 1.3916,
      "theoretical_loss": 3.3727361946261003,
      "tokens_seen": 2739339264
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017165496489468404,
      "loss": 1.3163,
      "theoretical_loss": 3.372729928929486,
      "tokens_seen": 2739404800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017163490471414243,
      "loss": 1.2901,
      "theoretical_loss": 3.372723663424736,
      "tokens_seen": 2739470336
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001716148445336008,
      "loss": 1.2249,
      "theoretical_loss": 3.3727173981118415,
      "tokens_seen": 2739535872
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017159478435305916,
      "loss": 1.2136,
      "theoretical_loss": 3.372711132990791,
      "tokens_seen": 2739601408
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017157472417251755,
      "loss": 1.3136,
      "theoretical_loss": 3.3727048680615743,
      "tokens_seen": 2739666944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001715546639919759,
      "loss": 1.2873,
      "theoretical_loss": 3.3726986033241815,
      "tokens_seen": 2739732480
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017153460381143433,
      "loss": 1.2386,
      "theoretical_loss": 3.3726923387786014,
      "tokens_seen": 2739798016
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001715145436308927,
      "loss": 1.2803,
      "theoretical_loss": 3.372686074424824,
      "tokens_seen": 2739863552
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017149448345035108,
      "loss": 1.272,
      "theoretical_loss": 3.372679810262839,
      "tokens_seen": 2739929088
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017147442326980944,
      "loss": 1.3538,
      "theoretical_loss": 3.3726735462926345,
      "tokens_seen": 2739994624
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017145436308926783,
      "loss": 1.2966,
      "theoretical_loss": 3.3726672825142017,
      "tokens_seen": 2740060160
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001714343029087262,
      "loss": 1.3439,
      "theoretical_loss": 3.3726610189275297,
      "tokens_seen": 2740125696
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017141424272818455,
      "loss": 1.3288,
      "theoretical_loss": 3.372654755532608,
      "tokens_seen": 2740191232
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017139418254764294,
      "loss": 1.4029,
      "theoretical_loss": 3.3726484923294255,
      "tokens_seen": 2740256768
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001713741223671013,
      "loss": 1.2156,
      "theoretical_loss": 3.3726422293179725,
      "tokens_seen": 2740322304
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001713540621865597,
      "loss": 1.2599,
      "theoretical_loss": 3.3726359664982386,
      "tokens_seen": 2740387840
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017133400200601806,
      "loss": 1.237,
      "theoretical_loss": 3.3726297038702135,
      "tokens_seen": 2740453376
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017131394182547645,
      "loss": 1.2351,
      "theoretical_loss": 3.372623441433886,
      "tokens_seen": 2740518912
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001712938816449348,
      "loss": 1.2672,
      "theoretical_loss": 3.3726171791892456,
      "tokens_seen": 2740584448
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001712738214643932,
      "loss": 1.2791,
      "theoretical_loss": 3.3726109171362824,
      "tokens_seen": 2740649984
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017125376128385156,
      "loss": 1.3374,
      "theoretical_loss": 3.3726046552749858,
      "tokens_seen": 2740715520
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017123370110330995,
      "loss": 1.2464,
      "theoretical_loss": 3.3725983936053456,
      "tokens_seen": 2740781056
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001712136409227683,
      "loss": 1.2517,
      "theoretical_loss": 3.372592132127351,
      "tokens_seen": 2740846592
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017119358074222667,
      "loss": 1.1275,
      "theoretical_loss": 3.3725858708409913,
      "tokens_seen": 2740912128
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017117352056168506,
      "loss": 1.2055,
      "theoretical_loss": 3.3725796097462566,
      "tokens_seen": 2740977664
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017115346038114342,
      "loss": 1.1924,
      "theoretical_loss": 3.3725733488431366,
      "tokens_seen": 2741043200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017113340020060181,
      "loss": 1.3295,
      "theoretical_loss": 3.37256708813162,
      "tokens_seen": 2741108736
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 996102,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3912932872772217,
      "objective/train/theoretical_loss": 3.372560827611697,
      "objective/train/tokens_used": 1111633376,
      "theoretical_loss": 3.372560827611697,
      "tokens_seen": 2741174272
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017111334002006018,
      "loss": 1.3562,
      "theoretical_loss": 3.372560827611697,
      "tokens_seen": 2741174272
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017109327983951857,
      "loss": 1.2341,
      "theoretical_loss": 3.3725545672833572,
      "tokens_seen": 2741239808
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017107321965897693,
      "loss": 1.3187,
      "theoretical_loss": 3.37254830714659,
      "tokens_seen": 2741305344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017105315947843532,
      "loss": 1.3165,
      "theoretical_loss": 3.3725420472013843,
      "tokens_seen": 2741370880
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017103309929789368,
      "loss": 1.2562,
      "theoretical_loss": 3.372535787447731,
      "tokens_seen": 2741436416
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017101303911735204,
      "loss": 1.3679,
      "theoretical_loss": 3.372529527885619,
      "tokens_seen": 2741501952
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017099297893681043,
      "loss": 1.1858,
      "theoretical_loss": 3.372523268515037,
      "tokens_seen": 2741567488
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001709729187562688,
      "loss": 1.3546,
      "theoretical_loss": 3.372517009335976,
      "tokens_seen": 2741633024
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017095285857572718,
      "loss": 1.2919,
      "theoretical_loss": 3.3725107503484253,
      "tokens_seen": 2741698560
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017093279839518554,
      "loss": 1.3078,
      "theoretical_loss": 3.372504491552373,
      "tokens_seen": 2741764096
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017091273821464393,
      "loss": 1.1745,
      "theoretical_loss": 3.372498232947811,
      "tokens_seen": 2741829632
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001708926780341023,
      "loss": 1.3325,
      "theoretical_loss": 3.372491974534727,
      "tokens_seen": 2741895168
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017087261785356069,
      "loss": 1.2115,
      "theoretical_loss": 3.3724857163131112,
      "tokens_seen": 2741960704
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017085255767301905,
      "loss": 1.2321,
      "theoretical_loss": 3.372479458282953,
      "tokens_seen": 2742026240
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001708324974924774,
      "loss": 1.1983,
      "theoretical_loss": 3.3724732004442424,
      "tokens_seen": 2742091776
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001708124373119358,
      "loss": 1.198,
      "theoretical_loss": 3.3724669427969687,
      "tokens_seen": 2742157312
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001707923771313942,
      "loss": 1.3119,
      "theoretical_loss": 3.3724606853411214,
      "tokens_seen": 2742222848
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017077231695085258,
      "loss": 1.2454,
      "theoretical_loss": 3.37245442807669,
      "tokens_seen": 2742288384
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017075225677031094,
      "loss": 1.2738,
      "theoretical_loss": 3.3724481710036645,
      "tokens_seen": 2742353920
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017073219658976933,
      "loss": 1.2454,
      "theoretical_loss": 3.372441914122034,
      "tokens_seen": 2742419456
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001707121364092277,
      "loss": 1.264,
      "theoretical_loss": 3.372435657431788,
      "tokens_seen": 2742484992
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017069207622868608,
      "loss": 1.2531,
      "theoretical_loss": 3.3724294009329165,
      "tokens_seen": 2742550528
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017067201604814444,
      "loss": 1.2194,
      "theoretical_loss": 3.372423144625409,
      "tokens_seen": 2742616064
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017065195586760283,
      "loss": 1.2529,
      "theoretical_loss": 3.372416888509255,
      "tokens_seen": 2742681600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001706318956870612,
      "loss": 1.258,
      "theoretical_loss": 3.372410632584444,
      "tokens_seen": 2742747136
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017061183550651956,
      "loss": 1.258,
      "theoretical_loss": 3.3724043768509655,
      "tokens_seen": 2742812672
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017059177532597795,
      "loss": 1.2722,
      "theoretical_loss": 3.3723981213088092,
      "tokens_seen": 2742878208
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001705717151454363,
      "loss": 1.3189,
      "theoretical_loss": 3.372391865957965,
      "tokens_seen": 2742943744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001705516549648947,
      "loss": 1.2494,
      "theoretical_loss": 3.3723856107984216,
      "tokens_seen": 2743009280
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017053159478435306,
      "loss": 1.195,
      "theoretical_loss": 3.3723793558301693,
      "tokens_seen": 2743074816
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017051153460381145,
      "loss": 1.1838,
      "theoretical_loss": 3.3723731010531974,
      "tokens_seen": 2743140352
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001704914744232698,
      "loss": 1.329,
      "theoretical_loss": 3.372366846467496,
      "tokens_seen": 2743205888
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001704714142427282,
      "loss": 1.2309,
      "theoretical_loss": 3.372360592073054,
      "tokens_seen": 2743271424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017045135406218656,
      "loss": 1.2451,
      "theoretical_loss": 3.3723543378698615,
      "tokens_seen": 2743336960
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017043129388164493,
      "loss": 1.3045,
      "theoretical_loss": 3.3723480838579074,
      "tokens_seen": 2743402496
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017041123370110332,
      "loss": 1.2715,
      "theoretical_loss": 3.372341830037182,
      "tokens_seen": 2743468032
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017039117352056168,
      "loss": 1.2904,
      "theoretical_loss": 3.3723355764076746,
      "tokens_seen": 2743533568
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017037111334002007,
      "loss": 1.1532,
      "theoretical_loss": 3.372329322969375,
      "tokens_seen": 2743599104
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017035105315947843,
      "loss": 1.1719,
      "theoretical_loss": 3.372323069722272,
      "tokens_seen": 2743664640
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017033099297893682,
      "loss": 1.2843,
      "theoretical_loss": 3.3723168166663564,
      "tokens_seen": 2743730176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017031093279839518,
      "loss": 1.2821,
      "theoretical_loss": 3.3723105638016166,
      "tokens_seen": 2743795712
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017029087261785357,
      "loss": 1.2686,
      "theoretical_loss": 3.372304311128043,
      "tokens_seen": 2743861248
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017027081243731193,
      "loss": 1.226,
      "theoretical_loss": 3.372298058645625,
      "tokens_seen": 2743926784
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017025075225677032,
      "loss": 1.2063,
      "theoretical_loss": 3.372291806354352,
      "tokens_seen": 2743992320
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017023069207622868,
      "loss": 1.177,
      "theoretical_loss": 3.372285554254214,
      "tokens_seen": 2744057856
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017021063189568705,
      "loss": 1.2426,
      "theoretical_loss": 3.3722793023452,
      "tokens_seen": 2744123392
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017019057171514544,
      "loss": 1.278,
      "theoretical_loss": 3.3722730506273,
      "tokens_seen": 2744188928
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001701705115346038,
      "loss": 1.1527,
      "theoretical_loss": 3.3722667991005033,
      "tokens_seen": 2744254464
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001701504513540622,
      "loss": 1.2925,
      "theoretical_loss": 3.3722605477648,
      "tokens_seen": 2744320000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017013039117352055,
      "loss": 1.3233,
      "theoretical_loss": 3.372254296620179,
      "tokens_seen": 2744385536
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 997363,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2676897048950195,
      "objective/train/theoretical_loss": 3.372248045666631,
      "objective/train/tokens_used": 1114910176,
      "theoretical_loss": 3.372248045666631,
      "tokens_seen": 2744451072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017011033099297894,
      "loss": 1.3492,
      "theoretical_loss": 3.372248045666631,
      "tokens_seen": 2744451072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001700902708124373,
      "loss": 1.1548,
      "theoretical_loss": 3.372241794904144,
      "tokens_seen": 2744516608
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001700702106318957,
      "loss": 1.3303,
      "theoretical_loss": 3.372235544332709,
      "tokens_seen": 2744582144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017005015045135405,
      "loss": 1.2691,
      "theoretical_loss": 3.372229293952315,
      "tokens_seen": 2744647680
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017003009027081244,
      "loss": 1.2684,
      "theoretical_loss": 3.3722230437629515,
      "tokens_seen": 2744713216
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017001003009027083,
      "loss": 1.2141,
      "theoretical_loss": 3.3722167937646086,
      "tokens_seen": 2744778752
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001699899699097292,
      "loss": 1.2941,
      "theoretical_loss": 3.3722105439572756,
      "tokens_seen": 2744844288
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016996990972918758,
      "loss": 1.1786,
      "theoretical_loss": 3.372204294340942,
      "tokens_seen": 2744909824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016994984954864594,
      "loss": 1.2765,
      "theoretical_loss": 3.3721980449155966,
      "tokens_seen": 2744975360
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016992978936810433,
      "loss": 1.3603,
      "theoretical_loss": 3.372191795681231,
      "tokens_seen": 2745040896
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001699097291875627,
      "loss": 1.2026,
      "theoretical_loss": 3.372185546637833,
      "tokens_seen": 2745106432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016988966900702109,
      "loss": 1.3169,
      "theoretical_loss": 3.3721792977853933,
      "tokens_seen": 2745171968
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016986960882647945,
      "loss": 1.3251,
      "theoretical_loss": 3.372173049123901,
      "tokens_seen": 2745237504
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001698495486459378,
      "loss": 1.3857,
      "theoretical_loss": 3.372166800653346,
      "tokens_seen": 2745303040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001698294884653962,
      "loss": 1.2291,
      "theoretical_loss": 3.3721605523737175,
      "tokens_seen": 2745368576
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016980942828485456,
      "loss": 1.343,
      "theoretical_loss": 3.3721543042850053,
      "tokens_seen": 2745434112
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016978936810431295,
      "loss": 1.2625,
      "theoretical_loss": 3.372148056387199,
      "tokens_seen": 2745499648
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001697693079237713,
      "loss": 1.3129,
      "theoretical_loss": 3.3721418086802886,
      "tokens_seen": 2745565184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001697492477432297,
      "loss": 1.2101,
      "theoretical_loss": 3.3721355611642627,
      "tokens_seen": 2745630720
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016972918756268807,
      "loss": 1.3214,
      "theoretical_loss": 3.372129313839112,
      "tokens_seen": 2745696256
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016970912738214645,
      "loss": 1.3275,
      "theoretical_loss": 3.372123066704826,
      "tokens_seen": 2745761792
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016968906720160482,
      "loss": 1.1856,
      "theoretical_loss": 3.3721168197613935,
      "tokens_seen": 2745827328
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001696690070210632,
      "loss": 1.2177,
      "theoretical_loss": 3.372110573008805,
      "tokens_seen": 2745892864
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016964894684052157,
      "loss": 1.2518,
      "theoretical_loss": 3.372104326447049,
      "tokens_seen": 2745958400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016962888665997993,
      "loss": 1.3004,
      "theoretical_loss": 3.3720980800761167,
      "tokens_seen": 2746023936
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016960882647943832,
      "loss": 1.234,
      "theoretical_loss": 3.372091833895996,
      "tokens_seen": 2746089472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016958876629889668,
      "loss": 1.1374,
      "theoretical_loss": 3.3720855879066782,
      "tokens_seen": 2746155008
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016956870611835507,
      "loss": 1.2979,
      "theoretical_loss": 3.372079342108152,
      "tokens_seen": 2746220544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016954864593781343,
      "loss": 1.2945,
      "theoretical_loss": 3.372073096500406,
      "tokens_seen": 2746286080
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016952858575727182,
      "loss": 1.2146,
      "theoretical_loss": 3.3720668510834324,
      "tokens_seen": 2746351616
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016950852557673019,
      "loss": 1.3769,
      "theoretical_loss": 3.3720606058572185,
      "tokens_seen": 2746417152
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016948846539618857,
      "loss": 1.2489,
      "theoretical_loss": 3.372054360821755,
      "tokens_seen": 2746482688
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016946840521564694,
      "loss": 1.2534,
      "theoretical_loss": 3.372048115977031,
      "tokens_seen": 2746548224
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001694483450351053,
      "loss": 1.1976,
      "theoretical_loss": 3.372041871323037,
      "tokens_seen": 2746613760
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001694282848545637,
      "loss": 1.3825,
      "theoretical_loss": 3.3720356268597618,
      "tokens_seen": 2746679296
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016940822467402205,
      "loss": 1.2015,
      "theoretical_loss": 3.372029382587195,
      "tokens_seen": 2746744832
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016938816449348044,
      "loss": 1.2684,
      "theoretical_loss": 3.3720231385053268,
      "tokens_seen": 2746810368
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001693681043129388,
      "loss": 1.3052,
      "theoretical_loss": 3.3720168946141467,
      "tokens_seen": 2746875904
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001693480441323972,
      "loss": 1.2937,
      "theoretical_loss": 3.3720106509136434,
      "tokens_seen": 2746941440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016932798395185555,
      "loss": 1.2156,
      "theoretical_loss": 3.372004407403808,
      "tokens_seen": 2747006976
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016930792377131394,
      "loss": 1.2643,
      "theoretical_loss": 3.371998164084629,
      "tokens_seen": 2747072512
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016928786359077233,
      "loss": 1.3483,
      "theoretical_loss": 3.3719919209560967,
      "tokens_seen": 2747138048
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016926780341023072,
      "loss": 1.2874,
      "theoretical_loss": 3.3719856780182003,
      "tokens_seen": 2747203584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016924774322968908,
      "loss": 1.2107,
      "theoretical_loss": 3.3719794352709296,
      "tokens_seen": 2747269120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016922768304914745,
      "loss": 1.2744,
      "theoretical_loss": 3.3719731927142744,
      "tokens_seen": 2747334656
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016920762286860584,
      "loss": 1.1012,
      "theoretical_loss": 3.3719669503482237,
      "tokens_seen": 2747400192
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001691875626880642,
      "loss": 1.2836,
      "theoretical_loss": 3.371960708172768,
      "tokens_seen": 2747465728
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001691675025075226,
      "loss": 1.3203,
      "theoretical_loss": 3.3719544661878964,
      "tokens_seen": 2747531264
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016914744232698095,
      "loss": 1.185,
      "theoretical_loss": 3.3719482243935985,
      "tokens_seen": 2747596800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016912738214643934,
      "loss": 1.3491,
      "theoretical_loss": 3.3719419827898647,
      "tokens_seen": 2747662336
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 998534,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.0819231271743774,
      "objective/train/theoretical_loss": 3.3719357413766833,
      "objective/train/tokens_used": 1118186976,
      "theoretical_loss": 3.3719357413766833,
      "tokens_seen": 2747727872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001691073219658977,
      "loss": 1.1673,
      "theoretical_loss": 3.3719357413766833,
      "tokens_seen": 2747727872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001690872617853561,
      "loss": 1.2326,
      "theoretical_loss": 3.371929500154045,
      "tokens_seen": 2747793408
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016906720160481445,
      "loss": 1.1831,
      "theoretical_loss": 3.3719232591219392,
      "tokens_seen": 2747858944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016904714142427281,
      "loss": 1.3285,
      "theoretical_loss": 3.3719170182803557,
      "tokens_seen": 2747924480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001690270812437312,
      "loss": 1.2737,
      "theoretical_loss": 3.3719107776292834,
      "tokens_seen": 2747990016
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016900702106318957,
      "loss": 1.375,
      "theoretical_loss": 3.371904537168713,
      "tokens_seen": 2748055552
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016898696088264796,
      "loss": 1.2779,
      "theoretical_loss": 3.3718982968986326,
      "tokens_seen": 2748121088
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016896690070210632,
      "loss": 1.2167,
      "theoretical_loss": 3.371892056819034,
      "tokens_seen": 2748186624
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001689468405215647,
      "loss": 1.3189,
      "theoretical_loss": 3.371885816929905,
      "tokens_seen": 2748252160
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016892678034102307,
      "loss": 1.2824,
      "theoretical_loss": 3.3718795772312355,
      "tokens_seen": 2748317696
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016890672016048146,
      "loss": 1.2821,
      "theoretical_loss": 3.371873337723016,
      "tokens_seen": 2748383232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016888665997993982,
      "loss": 1.2036,
      "theoretical_loss": 3.371867098405236,
      "tokens_seen": 2748448768
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016886659979939818,
      "loss": 1.2633,
      "theoretical_loss": 3.3718608592778847,
      "tokens_seen": 2748514304
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016884653961885657,
      "loss": 1.3038,
      "theoretical_loss": 3.371854620340952,
      "tokens_seen": 2748579840
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016882647943831494,
      "loss": 1.225,
      "theoretical_loss": 3.3718483815944267,
      "tokens_seen": 2748645376
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016880641925777332,
      "loss": 1.2285,
      "theoretical_loss": 3.3718421430382994,
      "tokens_seen": 2748710912
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001687863590772317,
      "loss": 1.2754,
      "theoretical_loss": 3.3718359046725603,
      "tokens_seen": 2748776448
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016876629889669008,
      "loss": 1.3156,
      "theoretical_loss": 3.371829666497198,
      "tokens_seen": 2748841984
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016874623871614844,
      "loss": 1.1659,
      "theoretical_loss": 3.3718234285122017,
      "tokens_seen": 2748907520
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016872617853560683,
      "loss": 1.2346,
      "theoretical_loss": 3.3718171907175627,
      "tokens_seen": 2748973056
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001687061183550652,
      "loss": 1.3037,
      "theoretical_loss": 3.371810953113269,
      "tokens_seen": 2749038592
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016868605817452355,
      "loss": 1.2666,
      "theoretical_loss": 3.371804715699312,
      "tokens_seen": 2749104128
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016866599799398194,
      "loss": 1.2406,
      "theoretical_loss": 3.3717984784756796,
      "tokens_seen": 2749169664
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001686459378134403,
      "loss": 1.3559,
      "theoretical_loss": 3.3717922414423622,
      "tokens_seen": 2749235200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001686258776328987,
      "loss": 1.2628,
      "theoretical_loss": 3.3717860045993495,
      "tokens_seen": 2749300736
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016860581745235706,
      "loss": 1.2768,
      "theoretical_loss": 3.3717797679466317,
      "tokens_seen": 2749366272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016858575727181544,
      "loss": 1.215,
      "theoretical_loss": 3.3717735314841972,
      "tokens_seen": 2749431808
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001685656970912738,
      "loss": 1.2588,
      "theoretical_loss": 3.3717672952120368,
      "tokens_seen": 2749497344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001685456369107322,
      "loss": 1.3582,
      "theoretical_loss": 3.3717610591301392,
      "tokens_seen": 2749562880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016852557673019059,
      "loss": 1.2727,
      "theoretical_loss": 3.371754823238495,
      "tokens_seen": 2749628416
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016850551654964897,
      "loss": 1.2612,
      "theoretical_loss": 3.3717485875370934,
      "tokens_seen": 2749693952
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016848545636910734,
      "loss": 1.3237,
      "theoretical_loss": 3.3717423520259238,
      "tokens_seen": 2749759488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001684653961885657,
      "loss": 1.263,
      "theoretical_loss": 3.371736116704976,
      "tokens_seen": 2749825024
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001684453360080241,
      "loss": 1.2925,
      "theoretical_loss": 3.3717298815742405,
      "tokens_seen": 2749890560
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016842527582748245,
      "loss": 1.3057,
      "theoretical_loss": 3.371723646633706,
      "tokens_seen": 2749956096
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016840521564694084,
      "loss": 1.2431,
      "theoretical_loss": 3.3717174118833624,
      "tokens_seen": 2750021632
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001683851554663992,
      "loss": 1.3273,
      "theoretical_loss": 3.371711177323199,
      "tokens_seen": 2750087168
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001683650952858576,
      "loss": 1.3489,
      "theoretical_loss": 3.3717049429532064,
      "tokens_seen": 2750152704
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016834503510531595,
      "loss": 1.3423,
      "theoretical_loss": 3.3716987087733736,
      "tokens_seen": 2750218240
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016832497492477434,
      "loss": 1.2318,
      "theoretical_loss": 3.3716924747836905,
      "tokens_seen": 2750283776
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001683049147442327,
      "loss": 1.2561,
      "theoretical_loss": 3.371686240984147,
      "tokens_seen": 2750349312
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016828485456369107,
      "loss": 1.3194,
      "theoretical_loss": 3.3716800073747315,
      "tokens_seen": 2750414848
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016826479438314946,
      "loss": 1.3696,
      "theoretical_loss": 3.371673773955435,
      "tokens_seen": 2750480384
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016824473420260782,
      "loss": 1.348,
      "theoretical_loss": 3.3716675407262473,
      "tokens_seen": 2750545920
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001682246740220662,
      "loss": 1.3902,
      "theoretical_loss": 3.371661307687157,
      "tokens_seen": 2750611456
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016820461384152457,
      "loss": 1.4415,
      "theoretical_loss": 3.371655074838155,
      "tokens_seen": 2750676992
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016818455366098296,
      "loss": 1.3532,
      "theoretical_loss": 3.3716488421792294,
      "tokens_seen": 2750742528
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016816449348044132,
      "loss": 1.1759,
      "theoretical_loss": 3.3716426097103716,
      "tokens_seen": 2750808064
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001681444332998997,
      "loss": 1.217,
      "theoretical_loss": 3.37163637743157,
      "tokens_seen": 2750873600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016812437311935807,
      "loss": 1.3303,
      "theoretical_loss": 3.371630145342815,
      "tokens_seen": 2750939136
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 999260,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1894795894622803,
      "objective/train/theoretical_loss": 3.3716239134440955,
      "objective/train/tokens_used": 1121463776,
      "theoretical_loss": 3.3716239134440955,
      "tokens_seen": 2751004672
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016810431293881646,
      "loss": 1.2375,
      "theoretical_loss": 3.3716239134440955,
      "tokens_seen": 2751004672
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016808425275827483,
      "loss": 1.2298,
      "theoretical_loss": 3.3716176817354024,
      "tokens_seen": 2751070208
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001680641925777332,
      "loss": 1.2025,
      "theoretical_loss": 3.3716114502167245,
      "tokens_seen": 2751135744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016804413239719158,
      "loss": 1.1846,
      "theoretical_loss": 3.3716052188880514,
      "tokens_seen": 2751201280
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016802407221664994,
      "loss": 1.256,
      "theoretical_loss": 3.371598987749373,
      "tokens_seen": 2751266816
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016800401203610833,
      "loss": 1.2829,
      "theoretical_loss": 3.371592756800679,
      "tokens_seen": 2751332352
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001679839518555667,
      "loss": 1.2404,
      "theoretical_loss": 3.3715865260419595,
      "tokens_seen": 2751397888
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016796389167502508,
      "loss": 1.3218,
      "theoretical_loss": 3.3715802954732035,
      "tokens_seen": 2751463424
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016794383149448344,
      "loss": 1.1767,
      "theoretical_loss": 3.3715740650944013,
      "tokens_seen": 2751528960
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016792377131394183,
      "loss": 1.2308,
      "theoretical_loss": 3.371567834905542,
      "tokens_seen": 2751594496
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001679037111334002,
      "loss": 1.2704,
      "theoretical_loss": 3.3715616049066153,
      "tokens_seen": 2751660032
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016788365095285856,
      "loss": 1.1846,
      "theoretical_loss": 3.3715553750976115,
      "tokens_seen": 2751725568
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016786359077231695,
      "loss": 1.2159,
      "theoretical_loss": 3.37154914547852,
      "tokens_seen": 2751791104
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001678435305917753,
      "loss": 1.1793,
      "theoretical_loss": 3.37154291604933,
      "tokens_seen": 2751856640
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001678234704112337,
      "loss": 1.3151,
      "theoretical_loss": 3.371536686810032,
      "tokens_seen": 2751922176
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016780341023069206,
      "loss": 1.3483,
      "theoretical_loss": 3.3715304577606147,
      "tokens_seen": 2751987712
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016778335005015048,
      "loss": 1.2376,
      "theoretical_loss": 3.3715242289010687,
      "tokens_seen": 2752053248
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016776328986960884,
      "loss": 1.343,
      "theoretical_loss": 3.3715180002313834,
      "tokens_seen": 2752118784
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016774322968906723,
      "loss": 1.3001,
      "theoretical_loss": 3.3715117717515484,
      "tokens_seen": 2752184320
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001677231695085256,
      "loss": 1.2814,
      "theoretical_loss": 3.3715055434615535,
      "tokens_seen": 2752249856
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016770310932798395,
      "loss": 1.2549,
      "theoretical_loss": 3.371499315361388,
      "tokens_seen": 2752315392
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016768304914744234,
      "loss": 1.2115,
      "theoretical_loss": 3.3714930874510425,
      "tokens_seen": 2752380928
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001676629889669007,
      "loss": 1.2776,
      "theoretical_loss": 3.371486859730506,
      "tokens_seen": 2752446464
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001676429287863591,
      "loss": 1.2282,
      "theoretical_loss": 3.371480632199768,
      "tokens_seen": 2752512000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016762286860581746,
      "loss": 1.2643,
      "theoretical_loss": 3.3714744048588186,
      "tokens_seen": 2752577536
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016760280842527584,
      "loss": 1.4276,
      "theoretical_loss": 3.371468177707648,
      "tokens_seen": 2752643072
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001675827482447342,
      "loss": 1.3795,
      "theoretical_loss": 3.3714619507462444,
      "tokens_seen": 2752708608
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001675626880641926,
      "loss": 1.4465,
      "theoretical_loss": 3.371455723974599,
      "tokens_seen": 2752774144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016754262788365096,
      "loss": 1.2546,
      "theoretical_loss": 3.371449497392701,
      "tokens_seen": 2752839680
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016752256770310935,
      "loss": 1.2639,
      "theoretical_loss": 3.3714432710005395,
      "tokens_seen": 2752905216
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001675025075225677,
      "loss": 1.1964,
      "theoretical_loss": 3.371437044798105,
      "tokens_seen": 2752970752
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016748244734202607,
      "loss": 1.2831,
      "theoretical_loss": 3.371430818785387,
      "tokens_seen": 2753036288
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016746238716148446,
      "loss": 1.3247,
      "theoretical_loss": 3.371424592962375,
      "tokens_seen": 2753101824
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016744232698094282,
      "loss": 1.2814,
      "theoretical_loss": 3.371418367329059,
      "tokens_seen": 2753167360
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001674222668004012,
      "loss": 1.2264,
      "theoretical_loss": 3.371412141885428,
      "tokens_seen": 2753232896
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016740220661985958,
      "loss": 1.3073,
      "theoretical_loss": 3.371405916631473,
      "tokens_seen": 2753298432
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016738214643931797,
      "loss": 1.2968,
      "theoretical_loss": 3.3713996915671824,
      "tokens_seen": 2753363968
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016736208625877633,
      "loss": 1.1687,
      "theoretical_loss": 3.371393466692547,
      "tokens_seen": 2753429504
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016734202607823472,
      "loss": 1.2647,
      "theoretical_loss": 3.3713872420075557,
      "tokens_seen": 2753495040
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016732196589769308,
      "loss": 1.337,
      "theoretical_loss": 3.3713810175121983,
      "tokens_seen": 2753560576
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016730190571715144,
      "loss": 1.4151,
      "theoretical_loss": 3.371374793206465,
      "tokens_seen": 2753626112
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016728184553660983,
      "loss": 1.3095,
      "theoretical_loss": 3.3713685690903445,
      "tokens_seen": 2753691648
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001672617853560682,
      "loss": 1.3294,
      "theoretical_loss": 3.3713623451638277,
      "tokens_seen": 2753757184
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016724172517552658,
      "loss": 1.3321,
      "theoretical_loss": 3.3713561214269037,
      "tokens_seen": 2753822720
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016722166499498494,
      "loss": 1.2653,
      "theoretical_loss": 3.3713498978795626,
      "tokens_seen": 2753888256
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016720160481444333,
      "loss": 1.2522,
      "theoretical_loss": 3.3713436745217935,
      "tokens_seen": 2753953792
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001671815446339017,
      "loss": 1.214,
      "theoretical_loss": 3.371337451353587,
      "tokens_seen": 2754019328
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016716148445336009,
      "loss": 1.26,
      "theoretical_loss": 3.3713312283749315,
      "tokens_seen": 2754084864
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016714142427281845,
      "loss": 1.2422,
      "theoretical_loss": 3.3713250055858177,
      "tokens_seen": 2754150400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001671213640922768,
      "loss": 1.3388,
      "theoretical_loss": 3.3713187829862354,
      "tokens_seen": 2754215936
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1000409,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.376333475112915,
      "objective/train/theoretical_loss": 3.371312560576174,
      "objective/train/tokens_used": 1124740576,
      "theoretical_loss": 3.371312560576174,
      "tokens_seen": 2754281472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001671013039117352,
      "loss": 1.2714,
      "theoretical_loss": 3.371312560576174,
      "tokens_seen": 2754281472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016708124373119356,
      "loss": 1.3123,
      "theoretical_loss": 3.371306338355623,
      "tokens_seen": 2754347008
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016706118355065195,
      "loss": 1.1889,
      "theoretical_loss": 3.3713001163245724,
      "tokens_seen": 2754412544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016704112337011034,
      "loss": 1.2765,
      "theoretical_loss": 3.3712938944830118,
      "tokens_seen": 2754478080
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016702106318956873,
      "loss": 1.2297,
      "theoretical_loss": 3.3712876728309316,
      "tokens_seen": 2754543616
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001670010030090271,
      "loss": 1.1897,
      "theoretical_loss": 3.3712814513683202,
      "tokens_seen": 2754609152
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016698094282848548,
      "loss": 1.2181,
      "theoretical_loss": 3.3712752300951685,
      "tokens_seen": 2754674688
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016696088264794384,
      "loss": 1.2743,
      "theoretical_loss": 3.371269009011465,
      "tokens_seen": 2754740224
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016694082246740223,
      "loss": 1.3309,
      "theoretical_loss": 3.371262788117201,
      "tokens_seen": 2754805760
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001669207622868606,
      "loss": 1.3581,
      "theoretical_loss": 3.3712565674123653,
      "tokens_seen": 2754871296
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016690070210631896,
      "loss": 1.2551,
      "theoretical_loss": 3.3712503468969475,
      "tokens_seen": 2754936832
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016688064192577735,
      "loss": 1.2573,
      "theoretical_loss": 3.3712441265709376,
      "tokens_seen": 2755002368
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001668605817452357,
      "loss": 1.2758,
      "theoretical_loss": 3.3712379064343256,
      "tokens_seen": 2755067904
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001668405215646941,
      "loss": 1.22,
      "theoretical_loss": 3.371231686487101,
      "tokens_seen": 2755133440
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016682046138415246,
      "loss": 1.3182,
      "theoretical_loss": 3.371225466729253,
      "tokens_seen": 2755198976
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016680040120361085,
      "loss": 1.355,
      "theoretical_loss": 3.371219247160772,
      "tokens_seen": 2755264512
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001667803410230692,
      "loss": 1.3263,
      "theoretical_loss": 3.371213027781647,
      "tokens_seen": 2755330048
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001667602808425276,
      "loss": 1.3499,
      "theoretical_loss": 3.371206808591869,
      "tokens_seen": 2755395584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016674022066198596,
      "loss": 1.3696,
      "theoretical_loss": 3.3712005895914263,
      "tokens_seen": 2755461120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016672016048144433,
      "loss": 1.3301,
      "theoretical_loss": 3.3711943707803096,
      "tokens_seen": 2755526656
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016670010030090271,
      "loss": 1.3457,
      "theoretical_loss": 3.371188152158509,
      "tokens_seen": 2755592192
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016668004012036108,
      "loss": 1.2773,
      "theoretical_loss": 3.371181933726013,
      "tokens_seen": 2755657728
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016665997993981947,
      "loss": 1.3321,
      "theoretical_loss": 3.3711757154828117,
      "tokens_seen": 2755723264
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016663991975927783,
      "loss": 1.2237,
      "theoretical_loss": 3.371169497428895,
      "tokens_seen": 2755788800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016661985957873622,
      "loss": 1.2962,
      "theoretical_loss": 3.371163279564253,
      "tokens_seen": 2755854336
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016659979939819458,
      "loss": 1.2718,
      "theoretical_loss": 3.371157061888875,
      "tokens_seen": 2755919872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016657973921765297,
      "loss": 1.2753,
      "theoretical_loss": 3.371150844402751,
      "tokens_seen": 2755985408
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016655967903711133,
      "loss": 1.2329,
      "theoretical_loss": 3.3711446271058705,
      "tokens_seen": 2756050944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016653961885656972,
      "loss": 1.2437,
      "theoretical_loss": 3.3711384099982236,
      "tokens_seen": 2756116480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016651955867602808,
      "loss": 1.3679,
      "theoretical_loss": 3.3711321930797995,
      "tokens_seen": 2756182016
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016649949849548645,
      "loss": 1.2775,
      "theoretical_loss": 3.371125976350588,
      "tokens_seen": 2756247552
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016647943831494484,
      "loss": 1.2927,
      "theoretical_loss": 3.371119759810579,
      "tokens_seen": 2756313088
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001664593781344032,
      "loss": 1.1863,
      "theoretical_loss": 3.3711135434597628,
      "tokens_seen": 2756378624
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001664393179538616,
      "loss": 1.3011,
      "theoretical_loss": 3.371107327298129,
      "tokens_seen": 2756444160
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016641925777331995,
      "loss": 1.2203,
      "theoretical_loss": 3.3711011113256664,
      "tokens_seen": 2756509696
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016639919759277834,
      "loss": 1.2104,
      "theoretical_loss": 3.371094895542365,
      "tokens_seen": 2756575232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001663791374122367,
      "loss": 1.3844,
      "theoretical_loss": 3.3710886799482154,
      "tokens_seen": 2756640768
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001663590772316951,
      "loss": 1.3288,
      "theoretical_loss": 3.371082464543207,
      "tokens_seen": 2756706304
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016633901705115345,
      "loss": 1.2656,
      "theoretical_loss": 3.371076249327329,
      "tokens_seen": 2756771840
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016631895687061181,
      "loss": 1.2298,
      "theoretical_loss": 3.3710700343005717,
      "tokens_seen": 2756837376
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001662988966900702,
      "loss": 1.3095,
      "theoretical_loss": 3.3710638194629245,
      "tokens_seen": 2756902912
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001662788365095286,
      "loss": 1.2716,
      "theoretical_loss": 3.3710576048143777,
      "tokens_seen": 2756968448
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016625877632898698,
      "loss": 1.2727,
      "theoretical_loss": 3.3710513903549204,
      "tokens_seen": 2757033984
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016623871614844534,
      "loss": 1.1839,
      "theoretical_loss": 3.371045176084543,
      "tokens_seen": 2757099520
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016621865596790373,
      "loss": 1.3455,
      "theoretical_loss": 3.3710389620032344,
      "tokens_seen": 2757165056
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001661985957873621,
      "loss": 1.2874,
      "theoretical_loss": 3.371032748110985,
      "tokens_seen": 2757230592
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016617853560682049,
      "loss": 1.3689,
      "theoretical_loss": 3.3710265344077843,
      "tokens_seen": 2757296128
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016615847542627885,
      "loss": 1.376,
      "theoretical_loss": 3.3710203208936225,
      "tokens_seen": 2757361664
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001661384152457372,
      "loss": 1.331,
      "theoretical_loss": 3.371014107568489,
      "tokens_seen": 2757427200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001661183550651956,
      "loss": 1.2802,
      "theoretical_loss": 3.371007894432373,
      "tokens_seen": 2757492736
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1001599,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3787121772766113,
      "objective/train/theoretical_loss": 3.3710016814852652,
      "objective/train/tokens_used": 1128017376,
      "theoretical_loss": 3.3710016814852652,
      "tokens_seen": 2757558272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016609829488465396,
      "loss": 1.4272,
      "theoretical_loss": 3.3710016814852652,
      "tokens_seen": 2757558272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016607823470411235,
      "loss": 1.2472,
      "theoretical_loss": 3.3709954687271555,
      "tokens_seen": 2757623808
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001660581745235707,
      "loss": 1.3416,
      "theoretical_loss": 3.3709892561580324,
      "tokens_seen": 2757689344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001660381143430291,
      "loss": 1.3559,
      "theoretical_loss": 3.3709830437778865,
      "tokens_seen": 2757754880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016601805416248746,
      "loss": 1.213,
      "theoretical_loss": 3.3709768315867077,
      "tokens_seen": 2757820416
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016599799398194585,
      "loss": 1.3054,
      "theoretical_loss": 3.3709706195844857,
      "tokens_seen": 2757885952
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016597793380140422,
      "loss": 1.2594,
      "theoretical_loss": 3.3709644077712095,
      "tokens_seen": 2757951488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001659578736208626,
      "loss": 1.1767,
      "theoretical_loss": 3.3709581961468698,
      "tokens_seen": 2758017024
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016593781344032097,
      "loss": 1.2584,
      "theoretical_loss": 3.370951984711456,
      "tokens_seen": 2758082560
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016591775325977933,
      "loss": 1.2527,
      "theoretical_loss": 3.3709457734649577,
      "tokens_seen": 2758148096
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016589769307923772,
      "loss": 1.2631,
      "theoretical_loss": 3.370939562407365,
      "tokens_seen": 2758213632
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016587763289869608,
      "loss": 1.2045,
      "theoretical_loss": 3.370933351538668,
      "tokens_seen": 2758279168
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016585757271815447,
      "loss": 1.2822,
      "theoretical_loss": 3.3709271408588553,
      "tokens_seen": 2758344704
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016583751253761283,
      "loss": 1.2555,
      "theoretical_loss": 3.3709209303679177,
      "tokens_seen": 2758410240
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016581745235707122,
      "loss": 1.271,
      "theoretical_loss": 3.3709147200658443,
      "tokens_seen": 2758475776
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016579739217652959,
      "loss": 1.2972,
      "theoretical_loss": 3.3709085099526255,
      "tokens_seen": 2758541312
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016577733199598797,
      "loss": 1.2692,
      "theoretical_loss": 3.3709023000282503,
      "tokens_seen": 2758606848
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016575727181544634,
      "loss": 1.2596,
      "theoretical_loss": 3.3708960902927094,
      "tokens_seen": 2758672384
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001657372116349047,
      "loss": 1.4131,
      "theoretical_loss": 3.370889880745992,
      "tokens_seen": 2758737920
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001657171514543631,
      "loss": 1.3679,
      "theoretical_loss": 3.370883671388088,
      "tokens_seen": 2758803456
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016569709127382145,
      "loss": 1.2896,
      "theoretical_loss": 3.3708774622189868,
      "tokens_seen": 2758868992
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016567703109327984,
      "loss": 1.3551,
      "theoretical_loss": 3.370871253238679,
      "tokens_seen": 2758934528
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001656569709127382,
      "loss": 1.3649,
      "theoretical_loss": 3.3708650444471537,
      "tokens_seen": 2759000064
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001656369107321966,
      "loss": 1.2963,
      "theoretical_loss": 3.370858835844401,
      "tokens_seen": 2759065600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016561685055165495,
      "loss": 1.3216,
      "theoretical_loss": 3.37085262743041,
      "tokens_seen": 2759131136
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016559679037111334,
      "loss": 1.2045,
      "theoretical_loss": 3.3708464192051717,
      "tokens_seen": 2759196672
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001655767301905717,
      "loss": 1.2393,
      "theoretical_loss": 3.370840211168675,
      "tokens_seen": 2759262208
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016555667001003007,
      "loss": 1.2352,
      "theoretical_loss": 3.3708340033209097,
      "tokens_seen": 2759327744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016553660982948848,
      "loss": 1.2635,
      "theoretical_loss": 3.370827795661866,
      "tokens_seen": 2759393280
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016551654964894685,
      "loss": 1.2759,
      "theoretical_loss": 3.370821588191533,
      "tokens_seen": 2759458816
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016549648946840524,
      "loss": 1.324,
      "theoretical_loss": 3.3708153809099013,
      "tokens_seen": 2759524352
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001654764292878636,
      "loss": 1.1488,
      "theoretical_loss": 3.3708091738169603,
      "tokens_seen": 2759589888
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000165456369107322,
      "loss": 1.2354,
      "theoretical_loss": 3.3708029669127,
      "tokens_seen": 2759655424
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016543630892678035,
      "loss": 1.2702,
      "theoretical_loss": 3.37079676019711,
      "tokens_seen": 2759720960
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016541624874623874,
      "loss": 1.272,
      "theoretical_loss": 3.3707905536701794,
      "tokens_seen": 2759786496
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001653961885656971,
      "loss": 1.2732,
      "theoretical_loss": 3.3707843473318992,
      "tokens_seen": 2759852032
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001653761283851555,
      "loss": 1.3023,
      "theoretical_loss": 3.3707781411822584,
      "tokens_seen": 2759917568
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016535606820461385,
      "loss": 1.2426,
      "theoretical_loss": 3.3707719352212475,
      "tokens_seen": 2759983104
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016533600802407221,
      "loss": 1.2797,
      "theoretical_loss": 3.370765729448855,
      "tokens_seen": 2760048640
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001653159478435306,
      "loss": 1.2381,
      "theoretical_loss": 3.3707595238650723,
      "tokens_seen": 2760114176
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016529588766298897,
      "loss": 1.2533,
      "theoretical_loss": 3.370753318469888,
      "tokens_seen": 2760179712
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016527582748244736,
      "loss": 1.3088,
      "theoretical_loss": 3.3707471132632922,
      "tokens_seen": 2760245248
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016525576730190572,
      "loss": 1.1988,
      "theoretical_loss": 3.3707409082452746,
      "tokens_seen": 2760310784
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001652357071213641,
      "loss": 1.3281,
      "theoretical_loss": 3.3707347034158257,
      "tokens_seen": 2760376320
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016521564694082247,
      "loss": 1.251,
      "theoretical_loss": 3.3707284987749344,
      "tokens_seen": 2760441856
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016519558676028086,
      "loss": 1.3619,
      "theoretical_loss": 3.370722294322591,
      "tokens_seen": 2760507392
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016517552657973922,
      "loss": 1.1338,
      "theoretical_loss": 3.370716090058785,
      "tokens_seen": 2760572928
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016515546639919758,
      "loss": 1.2411,
      "theoretical_loss": 3.3707098859835063,
      "tokens_seen": 2760638464
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016513540621865597,
      "loss": 1.2467,
      "theoretical_loss": 3.370703682096745,
      "tokens_seen": 2760704000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016511534603811433,
      "loss": 1.3318,
      "theoretical_loss": 3.3706974783984904,
      "tokens_seen": 2760769536
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1002769,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3618190288543701,
      "objective/train/theoretical_loss": 3.3706912748887325,
      "objective/train/tokens_used": 1131294176,
      "theoretical_loss": 3.3706912748887325,
      "tokens_seen": 2760835072
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016509528585757272,
      "loss": 1.3901,
      "theoretical_loss": 3.3706912748887325,
      "tokens_seen": 2760835072
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001650752256770311,
      "loss": 1.3447,
      "theoretical_loss": 3.370685071567461,
      "tokens_seen": 2760900608
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016505516549648948,
      "loss": 1.2878,
      "theoretical_loss": 3.3706788684346662,
      "tokens_seen": 2760966144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016503510531594784,
      "loss": 1.2157,
      "theoretical_loss": 3.3706726654903374,
      "tokens_seen": 2761031680
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016501504513540623,
      "loss": 1.2263,
      "theoretical_loss": 3.3706664627344645,
      "tokens_seen": 2761097216
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001649949849548646,
      "loss": 1.2767,
      "theoretical_loss": 3.3706602601670372,
      "tokens_seen": 2761162752
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016497492477432295,
      "loss": 1.2408,
      "theoretical_loss": 3.3706540577880455,
      "tokens_seen": 2761228288
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016495486459378134,
      "loss": 1.2725,
      "theoretical_loss": 3.3706478555974795,
      "tokens_seen": 2761293824
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001649348044132397,
      "loss": 1.2328,
      "theoretical_loss": 3.370641653595328,
      "tokens_seen": 2761359360
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001649147442326981,
      "loss": 1.342,
      "theoretical_loss": 3.370635451781582,
      "tokens_seen": 2761424896
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016489468405215646,
      "loss": 1.3471,
      "theoretical_loss": 3.3706292501562305,
      "tokens_seen": 2761490432
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016487462387161484,
      "loss": 1.2588,
      "theoretical_loss": 3.3706230487192634,
      "tokens_seen": 2761555968
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001648545636910732,
      "loss": 1.405,
      "theoretical_loss": 3.3706168474706706,
      "tokens_seen": 2761621504
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001648345035105316,
      "loss": 1.2964,
      "theoretical_loss": 3.3706106464104426,
      "tokens_seen": 2761687040
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016481444332998996,
      "loss": 1.3046,
      "theoretical_loss": 3.370604445538568,
      "tokens_seen": 2761752576
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016479438314944835,
      "loss": 1.2115,
      "theoretical_loss": 3.3705982448550373,
      "tokens_seen": 2761818112
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016477432296890674,
      "loss": 1.3285,
      "theoretical_loss": 3.3705920443598405,
      "tokens_seen": 2761883648
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001647542627883651,
      "loss": 1.2677,
      "theoretical_loss": 3.3705858440529664,
      "tokens_seen": 2761949184
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001647342026078235,
      "loss": 1.3254,
      "theoretical_loss": 3.370579643934406,
      "tokens_seen": 2762014720
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016471414242728185,
      "loss": 1.2949,
      "theoretical_loss": 3.370573444004149,
      "tokens_seen": 2762080256
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016469408224674024,
      "loss": 1.2499,
      "theoretical_loss": 3.370567244262184,
      "tokens_seen": 2762145792
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001646740220661986,
      "loss": 1.2353,
      "theoretical_loss": 3.370561044708502,
      "tokens_seen": 2762211328
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000164653961885657,
      "loss": 1.2908,
      "theoretical_loss": 3.3705548453430927,
      "tokens_seen": 2762276864
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016463390170511535,
      "loss": 1.3118,
      "theoretical_loss": 3.370548646165946,
      "tokens_seen": 2762342400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016461384152457374,
      "loss": 1.2551,
      "theoretical_loss": 3.3705424471770504,
      "tokens_seen": 2762407936
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001645937813440321,
      "loss": 1.2791,
      "theoretical_loss": 3.3705362483763976,
      "tokens_seen": 2762473472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016457372116349047,
      "loss": 1.295,
      "theoretical_loss": 3.370530049763976,
      "tokens_seen": 2762539008
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016455366098294886,
      "loss": 1.2676,
      "theoretical_loss": 3.3705238513397764,
      "tokens_seen": 2762604544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016453360080240722,
      "loss": 1.2614,
      "theoretical_loss": 3.370517653103788,
      "tokens_seen": 2762670080
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001645135406218656,
      "loss": 1.2387,
      "theoretical_loss": 3.3705114550560005,
      "tokens_seen": 2762735616
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016449348044132397,
      "loss": 1.2367,
      "theoretical_loss": 3.3705052571964043,
      "tokens_seen": 2762801152
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016447342026078236,
      "loss": 1.3554,
      "theoretical_loss": 3.370499059524989,
      "tokens_seen": 2762866688
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016445336008024072,
      "loss": 1.3788,
      "theoretical_loss": 3.3704928620417443,
      "tokens_seen": 2762932224
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001644332998996991,
      "loss": 1.2971,
      "theoretical_loss": 3.37048666474666,
      "tokens_seen": 2762997760
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016441323971915747,
      "loss": 1.398,
      "theoretical_loss": 3.370480467639726,
      "tokens_seen": 2763063296
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016439317953861586,
      "loss": 1.2954,
      "theoretical_loss": 3.3704742707209325,
      "tokens_seen": 2763128832
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016437311935807423,
      "loss": 1.3137,
      "theoretical_loss": 3.370468073990269,
      "tokens_seen": 2763194368
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001643530591775326,
      "loss": 1.1932,
      "theoretical_loss": 3.370461877447725,
      "tokens_seen": 2763259904
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016433299899699098,
      "loss": 1.2378,
      "theoretical_loss": 3.3704556810932904,
      "tokens_seen": 2763325440
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016431293881644934,
      "loss": 1.2293,
      "theoretical_loss": 3.3704494849269553,
      "tokens_seen": 2763390976
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016429287863590773,
      "loss": 1.1782,
      "theoretical_loss": 3.3704432889487097,
      "tokens_seen": 2763456512
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001642728184553661,
      "loss": 1.2989,
      "theoretical_loss": 3.3704370931585435,
      "tokens_seen": 2763522048
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016425275827482448,
      "loss": 1.1369,
      "theoretical_loss": 3.370430897556446,
      "tokens_seen": 2763587584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016423269809428284,
      "loss": 1.1962,
      "theoretical_loss": 3.370424702142407,
      "tokens_seen": 2763653120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016421263791374123,
      "loss": 1.2644,
      "theoretical_loss": 3.3704185069164168,
      "tokens_seen": 2763718656
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001641925777331996,
      "loss": 1.3315,
      "theoretical_loss": 3.370412311878465,
      "tokens_seen": 2763784192
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016417251755265796,
      "loss": 1.1844,
      "theoretical_loss": 3.3704061170285415,
      "tokens_seen": 2763849728
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016415245737211635,
      "loss": 1.2556,
      "theoretical_loss": 3.370399922366636,
      "tokens_seen": 2763915264
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001641323971915747,
      "loss": 1.2724,
      "theoretical_loss": 3.3703937278927385,
      "tokens_seen": 2763980800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001641123370110331,
      "loss": 1.3688,
      "theoretical_loss": 3.3703875336068387,
      "tokens_seen": 2764046336
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1003856,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1871845722198486,
      "objective/train/theoretical_loss": 3.370381339508927,
      "objective/train/tokens_used": 1134570976,
      "theoretical_loss": 3.370381339508927,
      "tokens_seen": 2764111872
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016409227683049146,
      "loss": 1.2468,
      "theoretical_loss": 3.370381339508927,
      "tokens_seen": 2764111872
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016407221664994985,
      "loss": 1.3024,
      "theoretical_loss": 3.370375145598992,
      "tokens_seen": 2764177408
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001640521564694082,
      "loss": 1.2677,
      "theoretical_loss": 3.370368951877025,
      "tokens_seen": 2764242944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016403209628886663,
      "loss": 1.2023,
      "theoretical_loss": 3.3703627583430147,
      "tokens_seen": 2764308480
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000164012036108325,
      "loss": 1.2856,
      "theoretical_loss": 3.3703565649969516,
      "tokens_seen": 2764374016
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016399197592778335,
      "loss": 1.3119,
      "theoretical_loss": 3.370350371838825,
      "tokens_seen": 2764439552
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016397191574724174,
      "loss": 1.3218,
      "theoretical_loss": 3.370344178868625,
      "tokens_seen": 2764505088
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001639518555667001,
      "loss": 1.293,
      "theoretical_loss": 3.370337986086342,
      "tokens_seen": 2764570624
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001639317953861585,
      "loss": 1.4112,
      "theoretical_loss": 3.3703317934919648,
      "tokens_seen": 2764636160
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016391173520561686,
      "loss": 1.3238,
      "theoretical_loss": 3.3703256010854843,
      "tokens_seen": 2764701696
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016389167502507524,
      "loss": 1.2879,
      "theoretical_loss": 3.3703194088668895,
      "tokens_seen": 2764767232
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001638716148445336,
      "loss": 1.2413,
      "theoretical_loss": 3.3703132168361707,
      "tokens_seen": 2764832768
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000163851554663992,
      "loss": 1.2491,
      "theoretical_loss": 3.3703070249933176,
      "tokens_seen": 2764898304
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016383149448345036,
      "loss": 1.2682,
      "theoretical_loss": 3.3703008333383204,
      "tokens_seen": 2764963840
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016381143430290875,
      "loss": 1.3225,
      "theoretical_loss": 3.370294641871168,
      "tokens_seen": 2765029376
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001637913741223671,
      "loss": 1.3117,
      "theoretical_loss": 3.3702884505918513,
      "tokens_seen": 2765094912
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016377131394182547,
      "loss": 1.2282,
      "theoretical_loss": 3.3702822595003594,
      "tokens_seen": 2765160448
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016375125376128386,
      "loss": 1.3778,
      "theoretical_loss": 3.3702760685966826,
      "tokens_seen": 2765225984
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016373119358074222,
      "loss": 1.327,
      "theoretical_loss": 3.3702698778808107,
      "tokens_seen": 2765291520
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001637111334002006,
      "loss": 1.3243,
      "theoretical_loss": 3.3702636873527334,
      "tokens_seen": 2765357056
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016369107321965898,
      "loss": 1.3506,
      "theoretical_loss": 3.3702574970124406,
      "tokens_seen": 2765422592
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016367101303911736,
      "loss": 1.3616,
      "theoretical_loss": 3.370251306859922,
      "tokens_seen": 2765488128
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016365095285857573,
      "loss": 1.2166,
      "theoretical_loss": 3.370245116895168,
      "tokens_seen": 2765553664
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016363089267803412,
      "loss": 1.427,
      "theoretical_loss": 3.370238927118168,
      "tokens_seen": 2765619200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016361083249749248,
      "loss": 1.2037,
      "theoretical_loss": 3.370232737528912,
      "tokens_seen": 2765684736
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016359077231695084,
      "loss": 1.2782,
      "theoretical_loss": 3.3702265481273894,
      "tokens_seen": 2765750272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016357071213640923,
      "loss": 1.2883,
      "theoretical_loss": 3.370220358913591,
      "tokens_seen": 2765815808
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001635506519558676,
      "loss": 1.2342,
      "theoretical_loss": 3.3702141698875057,
      "tokens_seen": 2765881344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016353059177532598,
      "loss": 1.3131,
      "theoretical_loss": 3.3702079810491243,
      "tokens_seen": 2765946880
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016351053159478434,
      "loss": 1.2925,
      "theoretical_loss": 3.3702017923984355,
      "tokens_seen": 2766012416
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016349047141424273,
      "loss": 1.394,
      "theoretical_loss": 3.3701956039354304,
      "tokens_seen": 2766077952
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001634704112337011,
      "loss": 1.2878,
      "theoretical_loss": 3.3701894156600978,
      "tokens_seen": 2766143488
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016345035105315949,
      "loss": 1.4733,
      "theoretical_loss": 3.370183227572428,
      "tokens_seen": 2766209024
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016343029087261785,
      "loss": 1.1931,
      "theoretical_loss": 3.370177039672411,
      "tokens_seen": 2766274560
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001634102306920762,
      "loss": 1.23,
      "theoretical_loss": 3.3701708519600366,
      "tokens_seen": 2766340096
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001633901705115346,
      "loss": 1.2989,
      "theoretical_loss": 3.3701646644352943,
      "tokens_seen": 2766405632
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016337011033099296,
      "loss": 1.3169,
      "theoretical_loss": 3.3701584770981747,
      "tokens_seen": 2766471168
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016335005015045135,
      "loss": 1.115,
      "theoretical_loss": 3.370152289948667,
      "tokens_seen": 2766536704
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001633299899699097,
      "loss": 1.3471,
      "theoretical_loss": 3.3701461029867614,
      "tokens_seen": 2766602240
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001633099297893681,
      "loss": 1.3133,
      "theoretical_loss": 3.370139916212448,
      "tokens_seen": 2766667776
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016328986960882646,
      "loss": 1.2611,
      "theoretical_loss": 3.370133729625716,
      "tokens_seen": 2766733312
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016326980942828488,
      "loss": 1.2984,
      "theoretical_loss": 3.3701275432265554,
      "tokens_seen": 2766798848
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016324974924774324,
      "loss": 1.2212,
      "theoretical_loss": 3.370121357014957,
      "tokens_seen": 2766864384
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016322968906720163,
      "loss": 1.2986,
      "theoretical_loss": 3.370115170990909,
      "tokens_seen": 2766929920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016320962888666,
      "loss": 1.2524,
      "theoretical_loss": 3.370108985154403,
      "tokens_seen": 2766995456
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016318956870611836,
      "loss": 1.3288,
      "theoretical_loss": 3.370102799505428,
      "tokens_seen": 2767060992
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016316950852557675,
      "loss": 1.2687,
      "theoretical_loss": 3.3700966140439736,
      "tokens_seen": 2767126528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001631494483450351,
      "loss": 1.1719,
      "theoretical_loss": 3.3700904287700304,
      "tokens_seen": 2767192064
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001631293881644935,
      "loss": 1.2824,
      "theoretical_loss": 3.370084243683588,
      "tokens_seen": 2767257600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016310932798395186,
      "loss": 1.3743,
      "theoretical_loss": 3.3700780587846357,
      "tokens_seen": 2767323136
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1005007,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3620854616165161,
      "objective/train/theoretical_loss": 3.3700718740731643,
      "objective/train/tokens_used": 1137847776,
      "theoretical_loss": 3.3700718740731643,
      "tokens_seen": 2767388672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016308926780341025,
      "loss": 1.217,
      "theoretical_loss": 3.3700718740731643,
      "tokens_seen": 2767388672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001630692076228686,
      "loss": 1.1842,
      "theoretical_loss": 3.370065689549163,
      "tokens_seen": 2767454208
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000163049147442327,
      "loss": 1.3029,
      "theoretical_loss": 3.3700595052126223,
      "tokens_seen": 2767519744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016302908726178536,
      "loss": 1.266,
      "theoretical_loss": 3.3700533210635313,
      "tokens_seen": 2767585280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016300902708124373,
      "loss": 1.2676,
      "theoretical_loss": 3.3700471371018805,
      "tokens_seen": 2767650816
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016298896690070211,
      "loss": 1.3573,
      "theoretical_loss": 3.3700409533276594,
      "tokens_seen": 2767716352
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016296890672016048,
      "loss": 1.252,
      "theoretical_loss": 3.370034769740858,
      "tokens_seen": 2767781888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016294884653961887,
      "loss": 1.4028,
      "theoretical_loss": 3.3700285863414665,
      "tokens_seen": 2767847424
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016292878635907723,
      "loss": 1.1998,
      "theoretical_loss": 3.3700224031294748,
      "tokens_seen": 2767912960
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016290872617853562,
      "loss": 1.3296,
      "theoretical_loss": 3.370016220104872,
      "tokens_seen": 2767978496
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016288866599799398,
      "loss": 1.2872,
      "theoretical_loss": 3.3700100372676487,
      "tokens_seen": 2768044032
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016286860581745237,
      "loss": 1.2329,
      "theoretical_loss": 3.3700038546177944,
      "tokens_seen": 2768109568
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016284854563691073,
      "loss": 1.3292,
      "theoretical_loss": 3.369997672155299,
      "tokens_seen": 2768175104
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016282848545636912,
      "loss": 1.1957,
      "theoretical_loss": 3.369991489880153,
      "tokens_seen": 2768240640
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016280842527582748,
      "loss": 1.34,
      "theoretical_loss": 3.369985307792346,
      "tokens_seen": 2768306176
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016278836509528585,
      "loss": 1.2732,
      "theoretical_loss": 3.3699791258918674,
      "tokens_seen": 2768371712
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016276830491474423,
      "loss": 1.2435,
      "theoretical_loss": 3.3699729441787074,
      "tokens_seen": 2768437248
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001627482447342026,
      "loss": 1.3087,
      "theoretical_loss": 3.3699667626528558,
      "tokens_seen": 2768502784
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000162728184553661,
      "loss": 1.3298,
      "theoretical_loss": 3.3699605813143023,
      "tokens_seen": 2768568320
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016270812437311935,
      "loss": 1.3429,
      "theoretical_loss": 3.3699544001630377,
      "tokens_seen": 2768633856
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016268806419257774,
      "loss": 1.3856,
      "theoretical_loss": 3.369948219199051,
      "tokens_seen": 2768699392
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001626680040120361,
      "loss": 1.146,
      "theoretical_loss": 3.3699420384223324,
      "tokens_seen": 2768764928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001626479438314945,
      "loss": 1.1669,
      "theoretical_loss": 3.3699358578328718,
      "tokens_seen": 2768830464
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016262788365095285,
      "loss": 1.2574,
      "theoretical_loss": 3.369929677430659,
      "tokens_seen": 2768896000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016260782347041121,
      "loss": 1.2857,
      "theoretical_loss": 3.369923497215684,
      "tokens_seen": 2768961536
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001625877632898696,
      "loss": 1.3022,
      "theoretical_loss": 3.3699173171879364,
      "tokens_seen": 2769027072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016256770310932797,
      "loss": 1.3353,
      "theoretical_loss": 3.369911137347407,
      "tokens_seen": 2769092608
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016254764292878636,
      "loss": 1.2525,
      "theoretical_loss": 3.3699049576940845,
      "tokens_seen": 2769158144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016252758274824474,
      "loss": 1.299,
      "theoretical_loss": 3.3698987782279595,
      "tokens_seen": 2769223680
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016250752256770313,
      "loss": 1.2903,
      "theoretical_loss": 3.369892598949021,
      "tokens_seen": 2769289216
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001624874623871615,
      "loss": 1.2865,
      "theoretical_loss": 3.3698864198572607,
      "tokens_seen": 2769354752
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016246740220661989,
      "loss": 1.2789,
      "theoretical_loss": 3.369880240952667,
      "tokens_seen": 2769420288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016244734202607825,
      "loss": 1.2236,
      "theoretical_loss": 3.3698740622352306,
      "tokens_seen": 2769485824
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001624272818455366,
      "loss": 1.315,
      "theoretical_loss": 3.3698678837049405,
      "tokens_seen": 2769551360
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000162407221664995,
      "loss": 1.3916,
      "theoretical_loss": 3.3698617053617874,
      "tokens_seen": 2769616896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016238716148445336,
      "loss": 1.4239,
      "theoretical_loss": 3.369855527205761,
      "tokens_seen": 2769682432
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016236710130391175,
      "loss": 1.2843,
      "theoretical_loss": 3.3698493492368513,
      "tokens_seen": 2769747968
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001623470411233701,
      "loss": 1.2763,
      "theoretical_loss": 3.369843171455048,
      "tokens_seen": 2769813504
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001623269809428285,
      "loss": 1.3035,
      "theoretical_loss": 3.369836993860341,
      "tokens_seen": 2769879040
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016230692076228686,
      "loss": 1.2769,
      "theoretical_loss": 3.36983081645272,
      "tokens_seen": 2769944576
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016228686058174525,
      "loss": 1.2634,
      "theoretical_loss": 3.369824639232175,
      "tokens_seen": 2770010112
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016226680040120362,
      "loss": 1.2869,
      "theoretical_loss": 3.3698184621986966,
      "tokens_seen": 2770075648
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000162246740220662,
      "loss": 1.2722,
      "theoretical_loss": 3.3698122853522743,
      "tokens_seen": 2770141184
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016222668004012037,
      "loss": 1.4375,
      "theoretical_loss": 3.369806108692898,
      "tokens_seen": 2770206720
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016220661985957873,
      "loss": 1.2202,
      "theoretical_loss": 3.369799932220557,
      "tokens_seen": 2770272256
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016218655967903712,
      "loss": 1.2148,
      "theoretical_loss": 3.369793755935242,
      "tokens_seen": 2770337792
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016216649949849548,
      "loss": 1.2942,
      "theoretical_loss": 3.3697875798369425,
      "tokens_seen": 2770403328
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016214643931795387,
      "loss": 1.3072,
      "theoretical_loss": 3.3697814039256486,
      "tokens_seen": 2770468864
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016212637913741223,
      "loss": 1.2976,
      "theoretical_loss": 3.3697752282013504,
      "tokens_seen": 2770534400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016210631895687062,
      "loss": 1.3175,
      "theoretical_loss": 3.3697690526640374,
      "tokens_seen": 2770599936
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1006223,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.198311448097229,
      "objective/train/theoretical_loss": 3.3697628773136996,
      "objective/train/tokens_used": 1141124576,
      "theoretical_loss": 3.3697628773136996,
      "tokens_seen": 2770665472
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016208625877632898,
      "loss": 1.1907,
      "theoretical_loss": 3.3697628773136996,
      "tokens_seen": 2770665472
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016206619859578737,
      "loss": 1.23,
      "theoretical_loss": 3.3697567021503274,
      "tokens_seen": 2770731008
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016204613841524574,
      "loss": 1.3371,
      "theoretical_loss": 3.36975052717391,
      "tokens_seen": 2770796544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001620260782347041,
      "loss": 1.3134,
      "theoretical_loss": 3.369744352384438,
      "tokens_seen": 2770862080
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001620060180541625,
      "loss": 1.4237,
      "theoretical_loss": 3.3697381777819007,
      "tokens_seen": 2770927616
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016198595787362085,
      "loss": 1.3607,
      "theoretical_loss": 3.3697320033662885,
      "tokens_seen": 2770993152
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016196589769307924,
      "loss": 1.2428,
      "theoretical_loss": 3.369725829137591,
      "tokens_seen": 2771058688
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001619458375125376,
      "loss": 1.25,
      "theoretical_loss": 3.369719655095798,
      "tokens_seen": 2771124224
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000161925777331996,
      "loss": 1.2553,
      "theoretical_loss": 3.3697134812409,
      "tokens_seen": 2771189760
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016190571715145435,
      "loss": 1.3469,
      "theoretical_loss": 3.3697073075728863,
      "tokens_seen": 2771255296
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016188565697091274,
      "loss": 1.199,
      "theoretical_loss": 3.3697011340917475,
      "tokens_seen": 2771320832
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001618655967903711,
      "loss": 1.1764,
      "theoretical_loss": 3.3696949607974727,
      "tokens_seen": 2771386368
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016184553660982947,
      "loss": 1.3439,
      "theoretical_loss": 3.3696887876900528,
      "tokens_seen": 2771451904
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016182547642928786,
      "loss": 1.3476,
      "theoretical_loss": 3.3696826147694767,
      "tokens_seen": 2771517440
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016180541624874622,
      "loss": 1.308,
      "theoretical_loss": 3.369676442035735,
      "tokens_seen": 2771582976
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016178535606820464,
      "loss": 1.3253,
      "theoretical_loss": 3.369670269488817,
      "tokens_seen": 2771648512
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000161765295887663,
      "loss": 1.2715,
      "theoretical_loss": 3.3696640971287137,
      "tokens_seen": 2771714048
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001617452357071214,
      "loss": 1.2819,
      "theoretical_loss": 3.369657924955414,
      "tokens_seen": 2771779584
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016172517552657975,
      "loss": 1.2083,
      "theoretical_loss": 3.3696517529689087,
      "tokens_seen": 2771845120
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016170511534603814,
      "loss": 1.2164,
      "theoretical_loss": 3.3696455811691868,
      "tokens_seen": 2771910656
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001616850551654965,
      "loss": 1.2479,
      "theoretical_loss": 3.369639409556239,
      "tokens_seen": 2771976192
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001616649949849549,
      "loss": 1.2754,
      "theoretical_loss": 3.3696332381300547,
      "tokens_seen": 2772041728
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016164493480441325,
      "loss": 1.284,
      "theoretical_loss": 3.369627066890624,
      "tokens_seen": 2772107264
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016162487462387161,
      "loss": 1.3108,
      "theoretical_loss": 3.369620895837937,
      "tokens_seen": 2772172800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016160481444333,
      "loss": 1.2698,
      "theoretical_loss": 3.3696147249719837,
      "tokens_seen": 2772238336
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016158475426278837,
      "loss": 1.2472,
      "theoretical_loss": 3.3696085542927534,
      "tokens_seen": 2772303872
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016156469408224676,
      "loss": 1.2051,
      "theoretical_loss": 3.369602383800237,
      "tokens_seen": 2772369408
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016154463390170512,
      "loss": 1.1932,
      "theoretical_loss": 3.3695962134944235,
      "tokens_seen": 2772434944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001615245737211635,
      "loss": 1.2558,
      "theoretical_loss": 3.3695900433753034,
      "tokens_seen": 2772500480
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016150451354062187,
      "loss": 1.2355,
      "theoretical_loss": 3.369583873442867,
      "tokens_seen": 2772566016
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016148445336008026,
      "loss": 1.2555,
      "theoretical_loss": 3.369577703697103,
      "tokens_seen": 2772631552
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016146439317953862,
      "loss": 1.2524,
      "theoretical_loss": 3.3695715341380024,
      "tokens_seen": 2772697088
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016144433299899698,
      "loss": 1.1769,
      "theoretical_loss": 3.369565364765555,
      "tokens_seen": 2772762624
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016142427281845537,
      "loss": 1.1677,
      "theoretical_loss": 3.3695591955797504,
      "tokens_seen": 2772828160
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016140421263791373,
      "loss": 1.2685,
      "theoretical_loss": 3.3695530265805784,
      "tokens_seen": 2772893696
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016138415245737212,
      "loss": 1.2464,
      "theoretical_loss": 3.3695468577680296,
      "tokens_seen": 2772959232
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016136409227683049,
      "loss": 1.3843,
      "theoretical_loss": 3.3695406891420934,
      "tokens_seen": 2773024768
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016134403209628888,
      "loss": 1.1621,
      "theoretical_loss": 3.3695345207027603,
      "tokens_seen": 2773090304
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016132397191574724,
      "loss": 1.3214,
      "theoretical_loss": 3.3695283524500197,
      "tokens_seen": 2773155840
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016130391173520563,
      "loss": 1.2589,
      "theoretical_loss": 3.369522184383862,
      "tokens_seen": 2773221376
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000161283851554664,
      "loss": 1.2999,
      "theoretical_loss": 3.369516016504276,
      "tokens_seen": 2773286912
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016126379137412235,
      "loss": 1.226,
      "theoretical_loss": 3.369509848811253,
      "tokens_seen": 2773352448
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016124373119358074,
      "loss": 1.3334,
      "theoretical_loss": 3.3695036813047827,
      "tokens_seen": 2773417984
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001612236710130391,
      "loss": 1.2409,
      "theoretical_loss": 3.3694975139848546,
      "tokens_seen": 2773483520
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001612036108324975,
      "loss": 1.302,
      "theoretical_loss": 3.3694913468514596,
      "tokens_seen": 2773549056
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016118355065195585,
      "loss": 1.3112,
      "theoretical_loss": 3.369485179904586,
      "tokens_seen": 2773614592
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016116349047141424,
      "loss": 1.2147,
      "theoretical_loss": 3.369479013144225,
      "tokens_seen": 2773680128
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001611434302908726,
      "loss": 1.2544,
      "theoretical_loss": 3.3694728465703663,
      "tokens_seen": 2773745664
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000161123370110331,
      "loss": 1.3686,
      "theoretical_loss": 3.3694666801829998,
      "tokens_seen": 2773811200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016110330992978936,
      "loss": 1.3114,
      "theoretical_loss": 3.369460513982115,
      "tokens_seen": 2773876736
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1007399,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2168593406677246,
      "objective/train/theoretical_loss": 3.369454347967703,
      "objective/train/tokens_used": 1144401376,
      "theoretical_loss": 3.369454347967703,
      "tokens_seen": 2773942272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016108324974924775,
      "loss": 1.1862,
      "theoretical_loss": 3.369454347967703,
      "tokens_seen": 2773942272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001610631895687061,
      "loss": 1.2611,
      "theoretical_loss": 3.3694481821397524,
      "tokens_seen": 2774007808
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016104312938816447,
      "loss": 1.3657,
      "theoretical_loss": 3.369442016498254,
      "tokens_seen": 2774073344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001610230692076229,
      "loss": 1.2533,
      "theoretical_loss": 3.369435851043198,
      "tokens_seen": 2774138880
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016100300902708125,
      "loss": 1.3528,
      "theoretical_loss": 3.3694296857745734,
      "tokens_seen": 2774204416
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016098294884653964,
      "loss": 1.3286,
      "theoretical_loss": 3.369423520692371,
      "tokens_seen": 2774269952
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000160962888665998,
      "loss": 1.2963,
      "theoretical_loss": 3.36941735579658,
      "tokens_seen": 2774335488
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001609428284854564,
      "loss": 1.2819,
      "theoretical_loss": 3.369411191087191,
      "tokens_seen": 2774401024
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016092276830491475,
      "loss": 1.306,
      "theoretical_loss": 3.369405026564194,
      "tokens_seen": 2774466560
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016090270812437314,
      "loss": 1.2781,
      "theoretical_loss": 3.3693988622275786,
      "tokens_seen": 2774532096
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001608826479438315,
      "loss": 1.2475,
      "theoretical_loss": 3.369392698077335,
      "tokens_seen": 2774597632
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016086258776328987,
      "loss": 1.2191,
      "theoretical_loss": 3.3693865341134526,
      "tokens_seen": 2774663168
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016084252758274826,
      "loss": 1.1975,
      "theoretical_loss": 3.3693803703359224,
      "tokens_seen": 2774728704
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016082246740220662,
      "loss": 1.3225,
      "theoretical_loss": 3.369374206744733,
      "tokens_seen": 2774794240
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000160802407221665,
      "loss": 1.2348,
      "theoretical_loss": 3.369368043339876,
      "tokens_seen": 2774859776
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016078234704112337,
      "loss": 1.2765,
      "theoretical_loss": 3.36936188012134,
      "tokens_seen": 2774925312
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016076228686058176,
      "loss": 1.288,
      "theoretical_loss": 3.3693557170891157,
      "tokens_seen": 2774990848
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016074222668004012,
      "loss": 1.3271,
      "theoretical_loss": 3.3693495542431924,
      "tokens_seen": 2775056384
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001607221664994985,
      "loss": 1.2797,
      "theoretical_loss": 3.369343391583561,
      "tokens_seen": 2775121920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016070210631895687,
      "loss": 1.242,
      "theoretical_loss": 3.369337229110211,
      "tokens_seen": 2775187456
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016068204613841526,
      "loss": 1.1628,
      "theoretical_loss": 3.369331066823132,
      "tokens_seen": 2775252992
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016066198595787363,
      "loss": 1.2762,
      "theoretical_loss": 3.3693249047223146,
      "tokens_seen": 2775318528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000160641925777332,
      "loss": 1.1142,
      "theoretical_loss": 3.3693187428077485,
      "tokens_seen": 2775384064
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016062186559679038,
      "loss": 1.2228,
      "theoretical_loss": 3.3693125810794236,
      "tokens_seen": 2775449600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016060180541624874,
      "loss": 1.2553,
      "theoretical_loss": 3.3693064195373297,
      "tokens_seen": 2775515136
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016058174523570713,
      "loss": 1.2869,
      "theoretical_loss": 3.3693002581814575,
      "tokens_seen": 2775580672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001605616850551655,
      "loss": 1.24,
      "theoretical_loss": 3.369294097011796,
      "tokens_seen": 2775646208
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016054162487462388,
      "loss": 1.2091,
      "theoretical_loss": 3.3692879360283357,
      "tokens_seen": 2775711744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016052156469408224,
      "loss": 1.178,
      "theoretical_loss": 3.369281775231067,
      "tokens_seen": 2775777280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016050150451354063,
      "loss": 1.2619,
      "theoretical_loss": 3.369275614619979,
      "tokens_seen": 2775842816
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000160481444332999,
      "loss": 1.2809,
      "theoretical_loss": 3.3692694541950625,
      "tokens_seen": 2775908352
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016046138415245736,
      "loss": 1.203,
      "theoretical_loss": 3.3692632939563065,
      "tokens_seen": 2775973888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016044132397191575,
      "loss": 1.2157,
      "theoretical_loss": 3.369257133903702,
      "tokens_seen": 2776039424
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001604212637913741,
      "loss": 1.1821,
      "theoretical_loss": 3.3692509740372385,
      "tokens_seen": 2776104960
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001604012036108325,
      "loss": 1.3248,
      "theoretical_loss": 3.369244814356906,
      "tokens_seen": 2776170496
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016038114343029086,
      "loss": 1.2996,
      "theoretical_loss": 3.3692386548626945,
      "tokens_seen": 2776236032
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016036108324974925,
      "loss": 1.2649,
      "theoretical_loss": 3.369232495554594,
      "tokens_seen": 2776301568
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001603410230692076,
      "loss": 1.2395,
      "theoretical_loss": 3.369226336432594,
      "tokens_seen": 2776367104
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000160320962888666,
      "loss": 1.3442,
      "theoretical_loss": 3.3692201774966857,
      "tokens_seen": 2776432640
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016030090270812436,
      "loss": 1.3217,
      "theoretical_loss": 3.3692140187468578,
      "tokens_seen": 2776498176
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016028084252758275,
      "loss": 1.2607,
      "theoretical_loss": 3.369207860183101,
      "tokens_seen": 2776563712
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016026078234704114,
      "loss": 1.3561,
      "theoretical_loss": 3.3692017018054052,
      "tokens_seen": 2776629248
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001602407221664995,
      "loss": 1.2089,
      "theoretical_loss": 3.36919554361376,
      "tokens_seen": 2776694784
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001602206619859579,
      "loss": 1.2782,
      "theoretical_loss": 3.369189385608156,
      "tokens_seen": 2776760320
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016020060180541626,
      "loss": 1.1468,
      "theoretical_loss": 3.369183227788583,
      "tokens_seen": 2776825856
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016018054162487464,
      "loss": 1.2281,
      "theoretical_loss": 3.3691770701550303,
      "tokens_seen": 2776891392
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000160160481444333,
      "loss": 1.2822,
      "theoretical_loss": 3.369170912707489,
      "tokens_seen": 2776956928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001601404212637914,
      "loss": 1.2566,
      "theoretical_loss": 3.369164755445948,
      "tokens_seen": 2777022464
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016012036108324976,
      "loss": 1.3422,
      "theoretical_loss": 3.369158598370398,
      "tokens_seen": 2777088000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016010030090270815,
      "loss": 1.2698,
      "theoretical_loss": 3.3691524414808294,
      "tokens_seen": 2777153536
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1008708,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2562955617904663,
      "objective/train/theoretical_loss": 3.3691462847772313,
      "objective/train/tokens_used": 1147678176,
      "theoretical_loss": 3.3691462847772313,
      "tokens_seen": 2777219072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001600802407221665,
      "loss": 1.2311,
      "theoretical_loss": 3.3691462847772313,
      "tokens_seen": 2777219072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016006018054162487,
      "loss": 1.2917,
      "theoretical_loss": 3.369140128259594,
      "tokens_seen": 2777284608
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016004012036108326,
      "loss": 1.1815,
      "theoretical_loss": 3.3691339719279076,
      "tokens_seen": 2777350144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016002006018054162,
      "loss": 1.2521,
      "theoretical_loss": 3.3691278157821616,
      "tokens_seen": 2777415680
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016,
      "loss": 1.242,
      "theoretical_loss": 3.369121659822347,
      "tokens_seen": 2777481216
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015997993981945838,
      "loss": 1.3357,
      "theoretical_loss": 3.3691155040484526,
      "tokens_seen": 2777546752
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015995987963891676,
      "loss": 1.2895,
      "theoretical_loss": 3.3691093484604693,
      "tokens_seen": 2777612288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015993981945837513,
      "loss": 1.3231,
      "theoretical_loss": 3.3691031930583866,
      "tokens_seen": 2777677824
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015991975927783352,
      "loss": 1.3878,
      "theoretical_loss": 3.369097037842195,
      "tokens_seen": 2777743360
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015989969909729188,
      "loss": 1.2995,
      "theoretical_loss": 3.369090882811884,
      "tokens_seen": 2777808896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015987963891675024,
      "loss": 1.3225,
      "theoretical_loss": 3.369084727967444,
      "tokens_seen": 2777874432
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015985957873620863,
      "loss": 1.182,
      "theoretical_loss": 3.3690785733088644,
      "tokens_seen": 2777939968
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000159839518555667,
      "loss": 1.2515,
      "theoretical_loss": 3.369072418836136,
      "tokens_seen": 2778005504
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015981945837512538,
      "loss": 1.3079,
      "theoretical_loss": 3.3690662645492484,
      "tokens_seen": 2778071040
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015979939819458374,
      "loss": 1.1674,
      "theoretical_loss": 3.3690601104481916,
      "tokens_seen": 2778136576
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015977933801404213,
      "loss": 1.2726,
      "theoretical_loss": 3.3690539565329556,
      "tokens_seen": 2778202112
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001597592778335005,
      "loss": 1.1346,
      "theoretical_loss": 3.3690478028035304,
      "tokens_seen": 2778267648
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015973921765295888,
      "loss": 1.3612,
      "theoretical_loss": 3.3690416492599056,
      "tokens_seen": 2778333184
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015971915747241725,
      "loss": 1.2974,
      "theoretical_loss": 3.369035495902072,
      "tokens_seen": 2778398720
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001596990972918756,
      "loss": 1.2414,
      "theoretical_loss": 3.3690293427300197,
      "tokens_seen": 2778464256
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000159679037111334,
      "loss": 1.1859,
      "theoretical_loss": 3.3690231897437375,
      "tokens_seen": 2778529792
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015965897693079236,
      "loss": 1.2945,
      "theoretical_loss": 3.369017036943217,
      "tokens_seen": 2778595328
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015963891675025075,
      "loss": 1.338,
      "theoretical_loss": 3.3690108843284468,
      "tokens_seen": 2778660864
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001596188565697091,
      "loss": 1.3367,
      "theoretical_loss": 3.3690047318994174,
      "tokens_seen": 2778726400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001595987963891675,
      "loss": 1.2002,
      "theoretical_loss": 3.368998579656119,
      "tokens_seen": 2778791936
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015957873620862586,
      "loss": 1.2901,
      "theoretical_loss": 3.368992427598542,
      "tokens_seen": 2778857472
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015955867602808425,
      "loss": 1.3218,
      "theoretical_loss": 3.3689862757266753,
      "tokens_seen": 2778923008
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015953861584754262,
      "loss": 1.2427,
      "theoretical_loss": 3.36898012404051,
      "tokens_seen": 2778988544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015951855566700103,
      "loss": 1.3137,
      "theoretical_loss": 3.368973972540035,
      "tokens_seen": 2779054080
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001594984954864594,
      "loss": 1.1387,
      "theoretical_loss": 3.3689678212252416,
      "tokens_seen": 2779119616
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015947843530591776,
      "loss": 1.3117,
      "theoretical_loss": 3.368961670096119,
      "tokens_seen": 2779185152
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015945837512537615,
      "loss": 1.2547,
      "theoretical_loss": 3.3689555191526575,
      "tokens_seen": 2779250688
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001594383149448345,
      "loss": 1.3175,
      "theoretical_loss": 3.368949368394847,
      "tokens_seen": 2779316224
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001594182547642929,
      "loss": 1.2757,
      "theoretical_loss": 3.3689432178226775,
      "tokens_seen": 2779381760
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015939819458375126,
      "loss": 1.2866,
      "theoretical_loss": 3.368937067436139,
      "tokens_seen": 2779447296
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015937813440320965,
      "loss": 1.2128,
      "theoretical_loss": 3.3689309172352218,
      "tokens_seen": 2779512832
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000159358074222668,
      "loss": 1.413,
      "theoretical_loss": 3.3689247672199154,
      "tokens_seen": 2779578368
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001593380140421264,
      "loss": 1.2921,
      "theoretical_loss": 3.3689186173902104,
      "tokens_seen": 2779643904
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015931795386158476,
      "loss": 1.2924,
      "theoretical_loss": 3.3689124677460964,
      "tokens_seen": 2779709440
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015929789368104313,
      "loss": 1.2705,
      "theoretical_loss": 3.3689063182875634,
      "tokens_seen": 2779774976
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015927783350050151,
      "loss": 1.1901,
      "theoretical_loss": 3.368900169014602,
      "tokens_seen": 2779840512
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015925777331995988,
      "loss": 1.2478,
      "theoretical_loss": 3.3688940199272017,
      "tokens_seen": 2779906048
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015923771313941827,
      "loss": 1.2087,
      "theoretical_loss": 3.3688878710253523,
      "tokens_seen": 2779971584
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015921765295887663,
      "loss": 1.2486,
      "theoretical_loss": 3.3688817223090446,
      "tokens_seen": 2780037120
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015919759277833502,
      "loss": 1.3353,
      "theoretical_loss": 3.3688755737782685,
      "tokens_seen": 2780102656
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015917753259779338,
      "loss": 1.2924,
      "theoretical_loss": 3.368869425433013,
      "tokens_seen": 2780168192
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015915747241725177,
      "loss": 1.0903,
      "theoretical_loss": 3.3688632772732694,
      "tokens_seen": 2780233728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015913741223671013,
      "loss": 1.3159,
      "theoretical_loss": 3.368857129299027,
      "tokens_seen": 2780299264
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001591173520561685,
      "loss": 1.3081,
      "theoretical_loss": 3.368850981510276,
      "tokens_seen": 2780364800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015909729187562688,
      "loss": 1.2354,
      "theoretical_loss": 3.3688448339070063,
      "tokens_seen": 2780430336
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1010001,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4214930534362793,
      "objective/train/theoretical_loss": 3.3688386864892084,
      "objective/train/tokens_used": 1150954976,
      "theoretical_loss": 3.3688386864892084,
      "tokens_seen": 2780495872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015907723169508525,
      "loss": 1.3234,
      "theoretical_loss": 3.3688386864892084,
      "tokens_seen": 2780495872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015905717151454363,
      "loss": 1.3184,
      "theoretical_loss": 3.368832539256872,
      "tokens_seen": 2780561408
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000159037111334002,
      "loss": 1.3075,
      "theoretical_loss": 3.368826392209987,
      "tokens_seen": 2780626944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015901705115346039,
      "loss": 1.1758,
      "theoretical_loss": 3.3688202453485436,
      "tokens_seen": 2780692480
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015899699097291875,
      "loss": 1.2902,
      "theoretical_loss": 3.368814098672532,
      "tokens_seen": 2780758016
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015897693079237714,
      "loss": 1.2318,
      "theoretical_loss": 3.3688079521819416,
      "tokens_seen": 2780823552
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001589568706118355,
      "loss": 1.2731,
      "theoretical_loss": 3.3688018058767635,
      "tokens_seen": 2780889088
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001589368104312939,
      "loss": 1.2808,
      "theoretical_loss": 3.368795659756987,
      "tokens_seen": 2780954624
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015891675025075225,
      "loss": 1.229,
      "theoretical_loss": 3.368789513822602,
      "tokens_seen": 2781020160
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015889669007021061,
      "loss": 1.2981,
      "theoretical_loss": 3.368783368073599,
      "tokens_seen": 2781085696
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000158876629889669,
      "loss": 1.3202,
      "theoretical_loss": 3.368777222509968,
      "tokens_seen": 2781151232
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015885656970912737,
      "loss": 1.2645,
      "theoretical_loss": 3.3687710771316985,
      "tokens_seen": 2781216768
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015883650952858575,
      "loss": 1.2931,
      "theoretical_loss": 3.3687649319387813,
      "tokens_seen": 2781282304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015881644934804412,
      "loss": 1.1619,
      "theoretical_loss": 3.368758786931206,
      "tokens_seen": 2781347840
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001587963891675025,
      "loss": 1.3567,
      "theoretical_loss": 3.3687526421089626,
      "tokens_seen": 2781413376
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001587763289869609,
      "loss": 1.2666,
      "theoretical_loss": 3.3687464974720416,
      "tokens_seen": 2781478912
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015875626880641929,
      "loss": 1.3107,
      "theoretical_loss": 3.3687403530204323,
      "tokens_seen": 2781544448
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015873620862587765,
      "loss": 1.2805,
      "theoretical_loss": 3.3687342087541254,
      "tokens_seen": 2781609984
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000158716148445336,
      "loss": 1.2343,
      "theoretical_loss": 3.3687280646731104,
      "tokens_seen": 2781675520
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001586960882647944,
      "loss": 1.323,
      "theoretical_loss": 3.3687219207773778,
      "tokens_seen": 2781741056
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015867602808425276,
      "loss": 1.2224,
      "theoretical_loss": 3.368715777066918,
      "tokens_seen": 2781806592
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015865596790371115,
      "loss": 1.2962,
      "theoretical_loss": 3.36870963354172,
      "tokens_seen": 2781872128
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001586359077231695,
      "loss": 1.326,
      "theoretical_loss": 3.3687034902017743,
      "tokens_seen": 2781937664
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001586158475426279,
      "loss": 1.3369,
      "theoretical_loss": 3.368697347047071,
      "tokens_seen": 2782003200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015859578736208626,
      "loss": 1.2518,
      "theoretical_loss": 3.3686912040776007,
      "tokens_seen": 2782068736
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015857572718154465,
      "loss": 1.3001,
      "theoretical_loss": 3.3686850612933528,
      "tokens_seen": 2782134272
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015855566700100302,
      "loss": 1.3071,
      "theoretical_loss": 3.368678918694317,
      "tokens_seen": 2782199808
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001585356068204614,
      "loss": 1.2761,
      "theoretical_loss": 3.3686727762804844,
      "tokens_seen": 2782265344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015851554663991977,
      "loss": 1.2678,
      "theoretical_loss": 3.3686666340518445,
      "tokens_seen": 2782330880
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015849548645937813,
      "loss": 1.328,
      "theoretical_loss": 3.368660492008387,
      "tokens_seen": 2782396416
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015847542627883652,
      "loss": 1.2283,
      "theoretical_loss": 3.368654350150103,
      "tokens_seen": 2782461952
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015845536609829488,
      "loss": 1.2909,
      "theoretical_loss": 3.3686482084769813,
      "tokens_seen": 2782527488
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015843530591775327,
      "loss": 1.3173,
      "theoretical_loss": 3.3686420669890125,
      "tokens_seen": 2782593024
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015841524573721163,
      "loss": 1.2632,
      "theoretical_loss": 3.3686359256861866,
      "tokens_seen": 2782658560
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015839518555667002,
      "loss": 1.2622,
      "theoretical_loss": 3.368629784568494,
      "tokens_seen": 2782724096
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015837512537612838,
      "loss": 1.3249,
      "theoretical_loss": 3.3686236436359245,
      "tokens_seen": 2782789632
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015835506519558677,
      "loss": 1.2992,
      "theoretical_loss": 3.368617502888468,
      "tokens_seen": 2782855168
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015833500501504514,
      "loss": 1.3158,
      "theoretical_loss": 3.3686113623261154,
      "tokens_seen": 2782920704
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001583149448345035,
      "loss": 1.2475,
      "theoretical_loss": 3.3686052219488554,
      "tokens_seen": 2782986240
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001582948846539619,
      "loss": 1.2853,
      "theoretical_loss": 3.368599081756679,
      "tokens_seen": 2783051776
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015827482447342025,
      "loss": 1.315,
      "theoretical_loss": 3.3685929417495757,
      "tokens_seen": 2783117312
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015825476429287864,
      "loss": 1.1589,
      "theoretical_loss": 3.3685868019275365,
      "tokens_seen": 2783182848
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000158234704112337,
      "loss": 1.3116,
      "theoretical_loss": 3.36858066229055,
      "tokens_seen": 2783248384
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001582146439317954,
      "loss": 1.2277,
      "theoretical_loss": 3.3685745228386077,
      "tokens_seen": 2783313920
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015819458375125375,
      "loss": 1.2877,
      "theoretical_loss": 3.3685683835716986,
      "tokens_seen": 2783379456
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015817452357071214,
      "loss": 1.2382,
      "theoretical_loss": 3.368562244489814,
      "tokens_seen": 2783444992
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001581544633901705,
      "loss": 1.3183,
      "theoretical_loss": 3.3685561055929427,
      "tokens_seen": 2783510528
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015813440320962887,
      "loss": 1.3484,
      "theoretical_loss": 3.3685499668810754,
      "tokens_seen": 2783576064
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015811434302908726,
      "loss": 1.4689,
      "theoretical_loss": 3.368543828354202,
      "tokens_seen": 2783641600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015809428284854562,
      "loss": 1.256,
      "theoretical_loss": 3.3685376900123125,
      "tokens_seen": 2783707136
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1011173,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2660366296768188,
      "objective/train/theoretical_loss": 3.3685315518553973,
      "objective/train/tokens_used": 1154231776,
      "theoretical_loss": 3.3685315518553973,
      "tokens_seen": 2783772672
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000158074222668004,
      "loss": 1.2205,
      "theoretical_loss": 3.3685315518553973,
      "tokens_seen": 2783772672
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015805416248746237,
      "loss": 1.2328,
      "theoretical_loss": 3.3685254138834457,
      "tokens_seen": 2783838208
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015803410230692076,
      "loss": 1.3744,
      "theoretical_loss": 3.368519276096449,
      "tokens_seen": 2783903744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015801404212637915,
      "loss": 1.263,
      "theoretical_loss": 3.3685131384943965,
      "tokens_seen": 2783969280
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015799398194583754,
      "loss": 1.2344,
      "theoretical_loss": 3.368507001077278,
      "tokens_seen": 2784034816
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001579739217652959,
      "loss": 1.2693,
      "theoretical_loss": 3.3685008638450844,
      "tokens_seen": 2784100352
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001579538615847543,
      "loss": 1.35,
      "theoretical_loss": 3.368494726797805,
      "tokens_seen": 2784165888
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015793380140421265,
      "loss": 1.1348,
      "theoretical_loss": 3.3684885899354304,
      "tokens_seen": 2784231424
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015791374122367101,
      "loss": 1.301,
      "theoretical_loss": 3.3684824532579505,
      "tokens_seen": 2784296960
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001578936810431294,
      "loss": 1.3771,
      "theoretical_loss": 3.368476316765355,
      "tokens_seen": 2784362496
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015787362086258777,
      "loss": 1.1526,
      "theoretical_loss": 3.368470180457635,
      "tokens_seen": 2784428032
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015785356068204616,
      "loss": 1.3135,
      "theoretical_loss": 3.3684640443347793,
      "tokens_seen": 2784493568
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015783350050150452,
      "loss": 1.2889,
      "theoretical_loss": 3.3684579083967785,
      "tokens_seen": 2784559104
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001578134403209629,
      "loss": 1.2195,
      "theoretical_loss": 3.368451772643623,
      "tokens_seen": 2784624640
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015779338014042127,
      "loss": 1.2382,
      "theoretical_loss": 3.368445637075303,
      "tokens_seen": 2784690176
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015777331995987966,
      "loss": 1.2356,
      "theoretical_loss": 3.368439501691808,
      "tokens_seen": 2784755712
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015775325977933802,
      "loss": 1.2887,
      "theoretical_loss": 3.3684333664931283,
      "tokens_seen": 2784821248
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015773319959879638,
      "loss": 1.2124,
      "theoretical_loss": 3.368427231479254,
      "tokens_seen": 2784886784
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015771313941825477,
      "loss": 1.2323,
      "theoretical_loss": 3.368421096650175,
      "tokens_seen": 2784952320
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015769307923771313,
      "loss": 1.2965,
      "theoretical_loss": 3.3684149620058816,
      "tokens_seen": 2785017856
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015767301905717152,
      "loss": 1.3576,
      "theoretical_loss": 3.3684088275463644,
      "tokens_seen": 2785083392
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015765295887662989,
      "loss": 1.273,
      "theoretical_loss": 3.3684026932716122,
      "tokens_seen": 2785148928
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015763289869608828,
      "loss": 1.3343,
      "theoretical_loss": 3.3683965591816163,
      "tokens_seen": 2785214464
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015761283851554664,
      "loss": 1.2379,
      "theoretical_loss": 3.3683904252763663,
      "tokens_seen": 2785280000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015759277833500503,
      "loss": 1.1643,
      "theoretical_loss": 3.368384291555852,
      "tokens_seen": 2785345536
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001575727181544634,
      "loss": 1.3233,
      "theoretical_loss": 3.3683781580200645,
      "tokens_seen": 2785411072
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015755265797392175,
      "loss": 1.2772,
      "theoretical_loss": 3.3683720246689925,
      "tokens_seen": 2785476608
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015753259779338014,
      "loss": 1.2327,
      "theoretical_loss": 3.368365891502627,
      "tokens_seen": 2785542144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001575125376128385,
      "loss": 1.3708,
      "theoretical_loss": 3.368359758520958,
      "tokens_seen": 2785607680
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001574924774322969,
      "loss": 1.3292,
      "theoretical_loss": 3.3683536257239752,
      "tokens_seen": 2785673216
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015747241725175525,
      "loss": 1.2999,
      "theoretical_loss": 3.368347493111669,
      "tokens_seen": 2785738752
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015745235707121364,
      "loss": 1.2749,
      "theoretical_loss": 3.36834136068403,
      "tokens_seen": 2785804288
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000157432296890672,
      "loss": 1.2429,
      "theoretical_loss": 3.368335228441047,
      "tokens_seen": 2785869824
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001574122367101304,
      "loss": 1.2681,
      "theoretical_loss": 3.3683290963827113,
      "tokens_seen": 2785935360
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015739217652958876,
      "loss": 1.3929,
      "theoretical_loss": 3.3683229645090123,
      "tokens_seen": 2786000896
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015737211634904715,
      "loss": 1.3789,
      "theoretical_loss": 3.36831683281994,
      "tokens_seen": 2786066432
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001573520561685055,
      "loss": 1.2159,
      "theoretical_loss": 3.368310701315486,
      "tokens_seen": 2786131968
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015733199598796387,
      "loss": 1.3449,
      "theoretical_loss": 3.368304569995638,
      "tokens_seen": 2786197504
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015731193580742226,
      "loss": 1.3518,
      "theoretical_loss": 3.3682984388603883,
      "tokens_seen": 2786263040
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015729187562688062,
      "loss": 1.2081,
      "theoretical_loss": 3.3682923079097256,
      "tokens_seen": 2786328576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015727181544633904,
      "loss": 1.2804,
      "theoretical_loss": 3.36828617714364,
      "tokens_seen": 2786394112
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001572517552657974,
      "loss": 1.2806,
      "theoretical_loss": 3.3682800465621225,
      "tokens_seen": 2786459648
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001572316950852558,
      "loss": 1.2719,
      "theoretical_loss": 3.3682739161651627,
      "tokens_seen": 2786525184
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015721163490471415,
      "loss": 1.2245,
      "theoretical_loss": 3.3682677859527512,
      "tokens_seen": 2786590720
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015719157472417254,
      "loss": 1.3396,
      "theoretical_loss": 3.368261655924877,
      "tokens_seen": 2786656256
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001571715145436309,
      "loss": 1.2403,
      "theoretical_loss": 3.3682555260815312,
      "tokens_seen": 2786721792
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015715145436308927,
      "loss": 1.2992,
      "theoretical_loss": 3.368249396422703,
      "tokens_seen": 2786787328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015713139418254766,
      "loss": 1.3695,
      "theoretical_loss": 3.368243266948384,
      "tokens_seen": 2786852864
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015711133400200602,
      "loss": 1.3909,
      "theoretical_loss": 3.368237137658563,
      "tokens_seen": 2786918400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001570912738214644,
      "loss": 1.4028,
      "theoretical_loss": 3.3682310085532303,
      "tokens_seen": 2786983936
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1012303,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2136070728302002,
      "objective/train/theoretical_loss": 3.3682248796323764,
      "objective/train/tokens_used": 1157508576,
      "theoretical_loss": 3.3682248796323764,
      "tokens_seen": 2787049472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015707121364092277,
      "loss": 1.2652,
      "theoretical_loss": 3.3682248796323764,
      "tokens_seen": 2787049472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015705115346038116,
      "loss": 1.298,
      "theoretical_loss": 3.368218750895991,
      "tokens_seen": 2787115008
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015703109327983952,
      "loss": 1.363,
      "theoretical_loss": 3.3682126223440645,
      "tokens_seen": 2787180544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001570110330992979,
      "loss": 1.188,
      "theoretical_loss": 3.368206493976587,
      "tokens_seen": 2787246080
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015699097291875627,
      "loss": 1.3281,
      "theoretical_loss": 3.368200365793548,
      "tokens_seen": 2787311616
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015697091273821466,
      "loss": 1.2751,
      "theoretical_loss": 3.368194237794939,
      "tokens_seen": 2787377152
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015695085255767303,
      "loss": 1.3443,
      "theoretical_loss": 3.368188109980749,
      "tokens_seen": 2787442688
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001569307923771314,
      "loss": 1.2638,
      "theoretical_loss": 3.368181982350968,
      "tokens_seen": 2787508224
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015691073219658978,
      "loss": 1.2809,
      "theoretical_loss": 3.368175854905587,
      "tokens_seen": 2787573760
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015689067201604814,
      "loss": 1.3232,
      "theoretical_loss": 3.368169727644595,
      "tokens_seen": 2787639296
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015687061183550653,
      "loss": 1.2255,
      "theoretical_loss": 3.3681636005679834,
      "tokens_seen": 2787704832
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001568505516549649,
      "loss": 1.2868,
      "theoretical_loss": 3.3681574736757414,
      "tokens_seen": 2787770368
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015683049147442328,
      "loss": 1.2943,
      "theoretical_loss": 3.368151346967859,
      "tokens_seen": 2787835904
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015681043129388164,
      "loss": 1.2692,
      "theoretical_loss": 3.368145220444327,
      "tokens_seen": 2787901440
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015679037111334003,
      "loss": 1.1873,
      "theoretical_loss": 3.3681390941051355,
      "tokens_seen": 2787966976
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001567703109327984,
      "loss": 1.3214,
      "theoretical_loss": 3.368132967950274,
      "tokens_seen": 2788032512
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015675025075225676,
      "loss": 1.2472,
      "theoretical_loss": 3.3681268419797328,
      "tokens_seen": 2788098048
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015673019057171515,
      "loss": 1.3471,
      "theoretical_loss": 3.3681207161935025,
      "tokens_seen": 2788163584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001567101303911735,
      "loss": 1.179,
      "theoretical_loss": 3.368114590591573,
      "tokens_seen": 2788229120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001566900702106319,
      "loss": 1.2836,
      "theoretical_loss": 3.3681084651739335,
      "tokens_seen": 2788294656
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015667001003009026,
      "loss": 1.2969,
      "theoretical_loss": 3.3681023399405756,
      "tokens_seen": 2788360192
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015664994984954865,
      "loss": 1.3795,
      "theoretical_loss": 3.3680962148914886,
      "tokens_seen": 2788425728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000156629889669007,
      "loss": 1.2702,
      "theoretical_loss": 3.368090090026663,
      "tokens_seen": 2788491264
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001566098294884654,
      "loss": 1.4187,
      "theoretical_loss": 3.3680839653460883,
      "tokens_seen": 2788556800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015658976930792376,
      "loss": 1.2504,
      "theoretical_loss": 3.368077840849755,
      "tokens_seen": 2788622336
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015656970912738212,
      "loss": 1.2987,
      "theoretical_loss": 3.3680717165376537,
      "tokens_seen": 2788687872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015654964894684051,
      "loss": 1.2605,
      "theoretical_loss": 3.368065592409774,
      "tokens_seen": 2788753408
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001565295887662989,
      "loss": 1.2618,
      "theoretical_loss": 3.368059468466106,
      "tokens_seen": 2788818944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001565095285857573,
      "loss": 1.3242,
      "theoretical_loss": 3.36805334470664,
      "tokens_seen": 2788884480
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015648946840521565,
      "loss": 1.3215,
      "theoretical_loss": 3.3680472211313663,
      "tokens_seen": 2788950016
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015646940822467404,
      "loss": 1.2712,
      "theoretical_loss": 3.3680410977402744,
      "tokens_seen": 2789015552
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001564493480441324,
      "loss": 1.2236,
      "theoretical_loss": 3.368034974533355,
      "tokens_seen": 2789081088
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001564292878635908,
      "loss": 1.3979,
      "theoretical_loss": 3.3680288515105983,
      "tokens_seen": 2789146624
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015640922768304916,
      "loss": 1.2403,
      "theoretical_loss": 3.368022728671994,
      "tokens_seen": 2789212160
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015638916750250755,
      "loss": 1.2464,
      "theoretical_loss": 3.3680166060175325,
      "tokens_seen": 2789277696
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001563691073219659,
      "loss": 1.2973,
      "theoretical_loss": 3.368010483547204,
      "tokens_seen": 2789343232
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015634904714142427,
      "loss": 1.2604,
      "theoretical_loss": 3.368004361260998,
      "tokens_seen": 2789408768
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015632898696088266,
      "loss": 1.262,
      "theoretical_loss": 3.367998239158906,
      "tokens_seen": 2789474304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015630892678034102,
      "loss": 1.2161,
      "theoretical_loss": 3.3679921172409166,
      "tokens_seen": 2789539840
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001562888665997994,
      "loss": 1.3997,
      "theoretical_loss": 3.367985995507021,
      "tokens_seen": 2789605376
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015626880641925778,
      "loss": 1.3152,
      "theoretical_loss": 3.367979873957209,
      "tokens_seen": 2789670912
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015624874623871616,
      "loss": 1.2825,
      "theoretical_loss": 3.3679737525914706,
      "tokens_seen": 2789736448
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015622868605817453,
      "loss": 1.309,
      "theoretical_loss": 3.3679676314097957,
      "tokens_seen": 2789801984
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015620862587763292,
      "loss": 1.2371,
      "theoretical_loss": 3.3679615104121754,
      "tokens_seen": 2789867520
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015618856569709128,
      "loss": 1.2543,
      "theoretical_loss": 3.367955389598599,
      "tokens_seen": 2789933056
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015616850551654964,
      "loss": 1.2357,
      "theoretical_loss": 3.367949268969057,
      "tokens_seen": 2789998592
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015614844533600803,
      "loss": 1.2691,
      "theoretical_loss": 3.3679431485235396,
      "tokens_seen": 2790064128
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001561283851554664,
      "loss": 1.2776,
      "theoretical_loss": 3.3679370282620362,
      "tokens_seen": 2790129664
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015610832497492478,
      "loss": 1.3179,
      "theoretical_loss": 3.3679309081845377,
      "tokens_seen": 2790195200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015608826479438314,
      "loss": 1.3128,
      "theoretical_loss": 3.3679247882910346,
      "tokens_seen": 2790260736
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1013591,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.435854434967041,
      "objective/train/theoretical_loss": 3.367918668581516,
      "objective/train/tokens_used": 1160785376,
      "theoretical_loss": 3.367918668581516,
      "tokens_seen": 2790326272
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015606820461384153,
      "loss": 1.4045,
      "theoretical_loss": 3.367918668581516,
      "tokens_seen": 2790326272
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001560481444332999,
      "loss": 1.3343,
      "theoretical_loss": 3.3679125490559727,
      "tokens_seen": 2790391808
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015602808425275828,
      "loss": 1.2761,
      "theoretical_loss": 3.3679064297143944,
      "tokens_seen": 2790457344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015600802407221665,
      "loss": 1.3228,
      "theoretical_loss": 3.367900310556772,
      "tokens_seen": 2790522880
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000155987963891675,
      "loss": 1.2184,
      "theoretical_loss": 3.367894191583095,
      "tokens_seen": 2790588416
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001559679037111334,
      "loss": 1.2595,
      "theoretical_loss": 3.367888072793354,
      "tokens_seen": 2790653952
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015594784353059176,
      "loss": 1.239,
      "theoretical_loss": 3.3678819541875384,
      "tokens_seen": 2790719488
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015592778335005015,
      "loss": 1.3567,
      "theoretical_loss": 3.3678758357656395,
      "tokens_seen": 2790785024
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001559077231695085,
      "loss": 1.3545,
      "theoretical_loss": 3.3678697175276464,
      "tokens_seen": 2790850560
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001558876629889669,
      "loss": 1.3513,
      "theoretical_loss": 3.3678635994735497,
      "tokens_seen": 2790916096
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015586760280842526,
      "loss": 1.2292,
      "theoretical_loss": 3.3678574816033393,
      "tokens_seen": 2790981632
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015584754262788365,
      "loss": 1.2174,
      "theoretical_loss": 3.3678513639170062,
      "tokens_seen": 2791047168
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015582748244734202,
      "loss": 1.395,
      "theoretical_loss": 3.3678452464145394,
      "tokens_seen": 2791112704
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001558074222668004,
      "loss": 1.3019,
      "theoretical_loss": 3.36783912909593,
      "tokens_seen": 2791178240
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015578736208625877,
      "loss": 1.3322,
      "theoretical_loss": 3.3678330119611672,
      "tokens_seen": 2791243776
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015576730190571716,
      "loss": 1.2482,
      "theoretical_loss": 3.3678268950102423,
      "tokens_seen": 2791309312
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015574724172517555,
      "loss": 1.2258,
      "theoretical_loss": 3.3678207782431446,
      "tokens_seen": 2791374848
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001557271815446339,
      "loss": 1.2809,
      "theoretical_loss": 3.3678146616598648,
      "tokens_seen": 2791440384
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001557071213640923,
      "loss": 1.1845,
      "theoretical_loss": 3.3678085452603925,
      "tokens_seen": 2791505920
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015568706118355066,
      "loss": 1.2645,
      "theoretical_loss": 3.367802429044718,
      "tokens_seen": 2791571456
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015566700100300905,
      "loss": 1.3305,
      "theoretical_loss": 3.367796313012832,
      "tokens_seen": 2791636992
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001556469408224674,
      "loss": 1.3637,
      "theoretical_loss": 3.367790197164724,
      "tokens_seen": 2791702528
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001556268806419258,
      "loss": 1.2409,
      "theoretical_loss": 3.3677840815003846,
      "tokens_seen": 2791768064
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015560682046138416,
      "loss": 1.3536,
      "theoretical_loss": 3.3677779660198035,
      "tokens_seen": 2791833600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015558676028084252,
      "loss": 1.3528,
      "theoretical_loss": 3.3677718507229715,
      "tokens_seen": 2791899136
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015556670010030091,
      "loss": 1.2712,
      "theoretical_loss": 3.3677657356098782,
      "tokens_seen": 2791964672
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015554663991975928,
      "loss": 1.2894,
      "theoretical_loss": 3.3677596206805145,
      "tokens_seen": 2792030208
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015552657973921767,
      "loss": 1.2682,
      "theoretical_loss": 3.3677535059348696,
      "tokens_seen": 2792095744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015550651955867603,
      "loss": 1.3866,
      "theoretical_loss": 3.367747391372934,
      "tokens_seen": 2792161280
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015548645937813442,
      "loss": 1.3225,
      "theoretical_loss": 3.3677412769946984,
      "tokens_seen": 2792226816
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015546639919759278,
      "loss": 1.3357,
      "theoretical_loss": 3.3677351628001526,
      "tokens_seen": 2792292352
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015544633901705117,
      "loss": 1.2411,
      "theoretical_loss": 3.367729048789286,
      "tokens_seen": 2792357888
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015542627883650953,
      "loss": 1.2915,
      "theoretical_loss": 3.3677229349620905,
      "tokens_seen": 2792423424
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001554062186559679,
      "loss": 1.2711,
      "theoretical_loss": 3.367716821318555,
      "tokens_seen": 2792488960
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015538615847542628,
      "loss": 1.2748,
      "theoretical_loss": 3.3677107078586697,
      "tokens_seen": 2792554496
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015536609829488465,
      "loss": 1.3145,
      "theoretical_loss": 3.367704594582425,
      "tokens_seen": 2792620032
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015534603811434303,
      "loss": 1.2352,
      "theoretical_loss": 3.3676984814898114,
      "tokens_seen": 2792685568
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001553259779338014,
      "loss": 1.273,
      "theoretical_loss": 3.3676923685808187,
      "tokens_seen": 2792751104
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015530591775325979,
      "loss": 1.2772,
      "theoretical_loss": 3.3676862558554372,
      "tokens_seen": 2792816640
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015528585757271815,
      "loss": 1.2598,
      "theoretical_loss": 3.367680143313657,
      "tokens_seen": 2792882176
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015526579739217654,
      "loss": 1.3922,
      "theoretical_loss": 3.367674030955468,
      "tokens_seen": 2792947712
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001552457372116349,
      "loss": 1.3467,
      "theoretical_loss": 3.367667918780861,
      "tokens_seen": 2793013248
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001552256770310933,
      "loss": 1.1551,
      "theoretical_loss": 3.367661806789826,
      "tokens_seen": 2793078784
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015520561685055165,
      "loss": 1.2571,
      "theoretical_loss": 3.3676556949823526,
      "tokens_seen": 2793144320
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015518555667001001,
      "loss": 1.2757,
      "theoretical_loss": 3.367649583358432,
      "tokens_seen": 2793209856
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001551654964894684,
      "loss": 1.2331,
      "theoretical_loss": 3.3676434719180532,
      "tokens_seen": 2793275392
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015514543630892677,
      "loss": 1.2455,
      "theoretical_loss": 3.367637360661208,
      "tokens_seen": 2793340928
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015512537612838515,
      "loss": 1.2727,
      "theoretical_loss": 3.367631249587885,
      "tokens_seen": 2793406464
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015510531594784352,
      "loss": 1.2624,
      "theoretical_loss": 3.3676251386980747,
      "tokens_seen": 2793472000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001550852557673019,
      "loss": 1.3668,
      "theoretical_loss": 3.3676190279917675,
      "tokens_seen": 2793537536
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1014727,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2322803735733032,
      "objective/train/theoretical_loss": 3.3676129174689535,
      "objective/train/tokens_used": 1164062176,
      "theoretical_loss": 3.3676129174689535,
      "tokens_seen": 2793603072
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015506519558676027,
      "loss": 1.2469,
      "theoretical_loss": 3.3676129174689535,
      "tokens_seen": 2793603072
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015504513540621866,
      "loss": 1.2831,
      "theoretical_loss": 3.3676068071296235,
      "tokens_seen": 2793668608
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015502507522567705,
      "loss": 1.1917,
      "theoretical_loss": 3.367600696973767,
      "tokens_seen": 2793734144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001550050150451354,
      "loss": 1.1831,
      "theoretical_loss": 3.3675945870013746,
      "tokens_seen": 2793799680
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001549849548645938,
      "loss": 1.3836,
      "theoretical_loss": 3.3675884772124363,
      "tokens_seen": 2793865216
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015496489468405216,
      "loss": 1.1708,
      "theoretical_loss": 3.367582367606942,
      "tokens_seen": 2793930752
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015494483450351055,
      "loss": 1.3289,
      "theoretical_loss": 3.367576258184882,
      "tokens_seen": 2793996288
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001549247743229689,
      "loss": 1.279,
      "theoretical_loss": 3.367570148946247,
      "tokens_seen": 2794061824
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001549047141424273,
      "loss": 1.2905,
      "theoretical_loss": 3.3675640398910267,
      "tokens_seen": 2794127360
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015488465396188566,
      "loss": 1.2941,
      "theoretical_loss": 3.3675579310192116,
      "tokens_seen": 2794192896
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015486459378134405,
      "loss": 1.2935,
      "theoretical_loss": 3.3675518223307916,
      "tokens_seen": 2794258432
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015484453360080242,
      "loss": 1.1678,
      "theoretical_loss": 3.367545713825757,
      "tokens_seen": 2794323968
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001548244734202608,
      "loss": 1.259,
      "theoretical_loss": 3.367539605504098,
      "tokens_seen": 2794389504
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015480441323971917,
      "loss": 1.3381,
      "theoretical_loss": 3.3675334973658044,
      "tokens_seen": 2794455040
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015478435305917753,
      "loss": 1.2109,
      "theoretical_loss": 3.3675273894108675,
      "tokens_seen": 2794520576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015476429287863592,
      "loss": 1.2585,
      "theoretical_loss": 3.3675212816392763,
      "tokens_seen": 2794586112
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015474423269809428,
      "loss": 1.2944,
      "theoretical_loss": 3.3675151740510216,
      "tokens_seen": 2794651648
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015472417251755267,
      "loss": 1.2783,
      "theoretical_loss": 3.3675090666460936,
      "tokens_seen": 2794717184
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015470411233701103,
      "loss": 1.3082,
      "theoretical_loss": 3.3675029594244825,
      "tokens_seen": 2794782720
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015468405215646942,
      "loss": 1.3727,
      "theoretical_loss": 3.367496852386178,
      "tokens_seen": 2794848256
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015466399197592778,
      "loss": 1.2887,
      "theoretical_loss": 3.367490745531171,
      "tokens_seen": 2794913792
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015464393179538617,
      "loss": 1.3228,
      "theoretical_loss": 3.3674846388594513,
      "tokens_seen": 2794979328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015462387161484454,
      "loss": 1.3273,
      "theoretical_loss": 3.367478532371009,
      "tokens_seen": 2795044864
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001546038114343029,
      "loss": 1.2028,
      "theoretical_loss": 3.367472426065835,
      "tokens_seen": 2795110400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001545837512537613,
      "loss": 1.3059,
      "theoretical_loss": 3.3674663199439183,
      "tokens_seen": 2795175936
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015456369107321965,
      "loss": 1.2816,
      "theoretical_loss": 3.3674602140052503,
      "tokens_seen": 2795241472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015454363089267804,
      "loss": 1.3267,
      "theoretical_loss": 3.3674541082498206,
      "tokens_seen": 2795307008
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001545235707121364,
      "loss": 1.3369,
      "theoretical_loss": 3.3674480026776195,
      "tokens_seen": 2795372544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001545035105315948,
      "loss": 1.2822,
      "theoretical_loss": 3.367441897288637,
      "tokens_seen": 2795438080
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015448345035105315,
      "loss": 1.3787,
      "theoretical_loss": 3.3674357920828637,
      "tokens_seen": 2795503616
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015446339017051154,
      "loss": 1.3471,
      "theoretical_loss": 3.3674296870602896,
      "tokens_seen": 2795569152
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001544433299899699,
      "loss": 1.3749,
      "theoretical_loss": 3.367423582220905,
      "tokens_seen": 2795634688
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015442326980942827,
      "loss": 1.2505,
      "theoretical_loss": 3.3674174775647,
      "tokens_seen": 2795700224
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015440320962888666,
      "loss": 1.232,
      "theoretical_loss": 3.367411373091665,
      "tokens_seen": 2795765760
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015438314944834502,
      "loss": 1.3499,
      "theoretical_loss": 3.3674052688017895,
      "tokens_seen": 2795831296
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001543630892678034,
      "loss": 1.2694,
      "theoretical_loss": 3.367399164695065,
      "tokens_seen": 2795896832
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015434302908726177,
      "loss": 1.342,
      "theoretical_loss": 3.36739306077148,
      "tokens_seen": 2795962368
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015432296890672016,
      "loss": 1.3576,
      "theoretical_loss": 3.3673869570310266,
      "tokens_seen": 2796027904
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015430290872617852,
      "loss": 1.2752,
      "theoretical_loss": 3.367380853473694,
      "tokens_seen": 2796093440
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001542828485456369,
      "loss": 1.3021,
      "theoretical_loss": 3.367374750099472,
      "tokens_seen": 2796158976
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001542627883650953,
      "loss": 1.2407,
      "theoretical_loss": 3.367368646908352,
      "tokens_seen": 2796224512
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001542427281845537,
      "loss": 1.2935,
      "theoretical_loss": 3.3673625439003234,
      "tokens_seen": 2796290048
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015422266800401205,
      "loss": 1.3069,
      "theoretical_loss": 3.3673564410753767,
      "tokens_seen": 2796355584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015420260782347041,
      "loss": 1.3316,
      "theoretical_loss": 3.3673503384335017,
      "tokens_seen": 2796421120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001541825476429288,
      "loss": 1.3127,
      "theoretical_loss": 3.367344235974689,
      "tokens_seen": 2796486656
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015416248746238717,
      "loss": 1.335,
      "theoretical_loss": 3.3673381336989285,
      "tokens_seen": 2796552192
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015414242728184555,
      "loss": 1.2738,
      "theoretical_loss": 3.3673320316062108,
      "tokens_seen": 2796617728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015412236710130392,
      "loss": 1.3322,
      "theoretical_loss": 3.367325929696526,
      "tokens_seen": 2796683264
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001541023069207623,
      "loss": 1.2551,
      "theoretical_loss": 3.3673198279698644,
      "tokens_seen": 2796748800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015408224674022067,
      "loss": 1.3306,
      "theoretical_loss": 3.367313726426216,
      "tokens_seen": 2796814336
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1015946,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2699360847473145,
      "objective/train/theoretical_loss": 3.367307625065571,
      "objective/train/tokens_used": 1167338976,
      "theoretical_loss": 3.367307625065571,
      "tokens_seen": 2796879872
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015406218655967906,
      "loss": 1.2302,
      "theoretical_loss": 3.367307625065571,
      "tokens_seen": 2796879872
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015404212637913742,
      "loss": 1.199,
      "theoretical_loss": 3.36730152388792,
      "tokens_seen": 2796945408
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015402206619859578,
      "loss": 1.3447,
      "theoretical_loss": 3.367295422893253,
      "tokens_seen": 2797010944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015400200601805417,
      "loss": 1.3696,
      "theoretical_loss": 3.36728932208156,
      "tokens_seen": 2797076480
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015398194583751253,
      "loss": 1.3815,
      "theoretical_loss": 3.3672832214528317,
      "tokens_seen": 2797142016
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015396188565697092,
      "loss": 1.2788,
      "theoretical_loss": 3.3672771210070573,
      "tokens_seen": 2797207552
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015394182547642929,
      "loss": 1.2762,
      "theoretical_loss": 3.3672710207442287,
      "tokens_seen": 2797273088
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015392176529588768,
      "loss": 1.3541,
      "theoretical_loss": 3.3672649206643346,
      "tokens_seen": 2797338624
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015390170511534604,
      "loss": 1.2917,
      "theoretical_loss": 3.367258820767366,
      "tokens_seen": 2797404160
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015388164493480443,
      "loss": 1.4599,
      "theoretical_loss": 3.367252721053313,
      "tokens_seen": 2797469696
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001538615847542628,
      "loss": 1.2909,
      "theoretical_loss": 3.367246621522166,
      "tokens_seen": 2797535232
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015384152457372115,
      "loss": 1.305,
      "theoretical_loss": 3.367240522173915,
      "tokens_seen": 2797600768
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015382146439317954,
      "loss": 1.3089,
      "theoretical_loss": 3.3672344230085502,
      "tokens_seen": 2797666304
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001538014042126379,
      "loss": 1.3421,
      "theoretical_loss": 3.367228324026062,
      "tokens_seen": 2797731840
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001537813440320963,
      "loss": 1.2542,
      "theoretical_loss": 3.36722222522644,
      "tokens_seen": 2797797376
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015376128385155465,
      "loss": 1.2337,
      "theoretical_loss": 3.3672161266096756,
      "tokens_seen": 2797862912
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015374122367101304,
      "loss": 1.2396,
      "theoretical_loss": 3.3672100281757578,
      "tokens_seen": 2797928448
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001537211634904714,
      "loss": 1.3445,
      "theoretical_loss": 3.367203929924678,
      "tokens_seen": 2797993984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001537011033099298,
      "loss": 1.2104,
      "theoretical_loss": 3.3671978318564255,
      "tokens_seen": 2798059520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015368104312938816,
      "loss": 1.2544,
      "theoretical_loss": 3.367191733970991,
      "tokens_seen": 2798125056
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015366098294884655,
      "loss": 1.2602,
      "theoretical_loss": 3.3671856362683648,
      "tokens_seen": 2798190592
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001536409227683049,
      "loss": 1.3109,
      "theoretical_loss": 3.3671795387485366,
      "tokens_seen": 2798256128
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015362086258776327,
      "loss": 1.2553,
      "theoretical_loss": 3.3671734414114978,
      "tokens_seen": 2798321664
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015360080240722166,
      "loss": 1.2937,
      "theoretical_loss": 3.367167344257237,
      "tokens_seen": 2798387200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015358074222668002,
      "loss": 1.2817,
      "theoretical_loss": 3.3671612472857455,
      "tokens_seen": 2798452736
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001535606820461384,
      "loss": 1.3663,
      "theoretical_loss": 3.3671551504970134,
      "tokens_seen": 2798518272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015354062186559677,
      "loss": 1.3485,
      "theoretical_loss": 3.367149053891031,
      "tokens_seen": 2798583808
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001535205616850552,
      "loss": 1.2589,
      "theoretical_loss": 3.3671429574677885,
      "tokens_seen": 2798649344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015350050150451355,
      "loss": 1.2109,
      "theoretical_loss": 3.367136861227276,
      "tokens_seen": 2798714880
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015348044132397194,
      "loss": 1.3117,
      "theoretical_loss": 3.3671307651694837,
      "tokens_seen": 2798780416
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001534603811434303,
      "loss": 1.3473,
      "theoretical_loss": 3.3671246692944017,
      "tokens_seen": 2798845952
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015344032096288867,
      "loss": 1.3155,
      "theoretical_loss": 3.367118573602021,
      "tokens_seen": 2798911488
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015342026078234706,
      "loss": 1.2024,
      "theoretical_loss": 3.367112478092331,
      "tokens_seen": 2798977024
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015340020060180542,
      "loss": 1.3188,
      "theoretical_loss": 3.3671063827653227,
      "tokens_seen": 2799042560
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001533801404212638,
      "loss": 1.2507,
      "theoretical_loss": 3.3671002876209855,
      "tokens_seen": 2799108096
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015336008024072217,
      "loss": 1.2107,
      "theoretical_loss": 3.3670941926593105,
      "tokens_seen": 2799173632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015334002006018056,
      "loss": 1.2544,
      "theoretical_loss": 3.3670880978802873,
      "tokens_seen": 2799239168
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015331995987963892,
      "loss": 1.319,
      "theoretical_loss": 3.3670820032839064,
      "tokens_seen": 2799304704
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001532998996990973,
      "loss": 1.2526,
      "theoretical_loss": 3.367075908870158,
      "tokens_seen": 2799370240
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015327983951855567,
      "loss": 1.258,
      "theoretical_loss": 3.3670698146390325,
      "tokens_seen": 2799435776
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015325977933801406,
      "loss": 1.3511,
      "theoretical_loss": 3.36706372059052,
      "tokens_seen": 2799501312
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015323971915747243,
      "loss": 1.2584,
      "theoretical_loss": 3.367057626724611,
      "tokens_seen": 2799566848
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001532196589769308,
      "loss": 1.2545,
      "theoretical_loss": 3.367051533041295,
      "tokens_seen": 2799632384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015319959879638918,
      "loss": 1.276,
      "theoretical_loss": 3.367045439540563,
      "tokens_seen": 2799697920
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015317953861584754,
      "loss": 1.2717,
      "theoretical_loss": 3.3670393462224055,
      "tokens_seen": 2799763456
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015315947843530593,
      "loss": 1.2771,
      "theoretical_loss": 3.367033253086812,
      "tokens_seen": 2799828992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001531394182547643,
      "loss": 1.2478,
      "theoretical_loss": 3.367027160133773,
      "tokens_seen": 2799894528
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015311935807422268,
      "loss": 1.216,
      "theoretical_loss": 3.367021067363279,
      "tokens_seen": 2799960064
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015309929789368104,
      "loss": 1.1972,
      "theoretical_loss": 3.36701497477532,
      "tokens_seen": 2800025600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015307923771313943,
      "loss": 1.279,
      "theoretical_loss": 3.3670088823698867,
      "tokens_seen": 2800091136
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1017141,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3148716688156128,
      "objective/train/theoretical_loss": 3.3670027901469686,
      "objective/train/tokens_used": 1170615776,
      "theoretical_loss": 3.3670027901469686,
      "tokens_seen": 2800156672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001530591775325978,
      "loss": 1.2839,
      "theoretical_loss": 3.3670027901469686,
      "tokens_seen": 2800156672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015303911735205616,
      "loss": 1.3135,
      "theoretical_loss": 3.3669966981065564,
      "tokens_seen": 2800222208
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015301905717151455,
      "loss": 1.204,
      "theoretical_loss": 3.3669906062486405,
      "tokens_seen": 2800287744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001529989969909729,
      "loss": 1.4353,
      "theoretical_loss": 3.366984514573211,
      "tokens_seen": 2800353280
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001529789368104313,
      "loss": 1.1759,
      "theoretical_loss": 3.3669784230802584,
      "tokens_seen": 2800418816
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015295887662988966,
      "loss": 1.2574,
      "theoretical_loss": 3.366972331769772,
      "tokens_seen": 2800484352
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015293881644934805,
      "loss": 1.2388,
      "theoretical_loss": 3.3669662406417435,
      "tokens_seen": 2800549888
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001529187562688064,
      "loss": 1.266,
      "theoretical_loss": 3.3669601496961623,
      "tokens_seen": 2800615424
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001528986960882648,
      "loss": 1.2993,
      "theoretical_loss": 3.3669540589330187,
      "tokens_seen": 2800680960
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015287863590772316,
      "loss": 1.2379,
      "theoretical_loss": 3.366947968352303,
      "tokens_seen": 2800746496
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015285857572718152,
      "loss": 1.2036,
      "theoretical_loss": 3.3669418779540057,
      "tokens_seen": 2800812032
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015283851554663991,
      "loss": 1.1584,
      "theoretical_loss": 3.3669357877381167,
      "tokens_seen": 2800877568
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015281845536609828,
      "loss": 1.294,
      "theoretical_loss": 3.366929697704627,
      "tokens_seen": 2800943104
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015279839518555667,
      "loss": 1.2787,
      "theoretical_loss": 3.3669236078535256,
      "tokens_seen": 2801008640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015277833500501503,
      "loss": 1.3286,
      "theoretical_loss": 3.3669175181848043,
      "tokens_seen": 2801074176
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015275827482447344,
      "loss": 1.3035,
      "theoretical_loss": 3.3669114286984523,
      "tokens_seen": 2801139712
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001527382146439318,
      "loss": 1.2963,
      "theoretical_loss": 3.36690533939446,
      "tokens_seen": 2801205248
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001527181544633902,
      "loss": 1.2752,
      "theoretical_loss": 3.3668992502728177,
      "tokens_seen": 2801270784
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015269809428284856,
      "loss": 1.2226,
      "theoretical_loss": 3.3668931613335165,
      "tokens_seen": 2801336320
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015267803410230695,
      "loss": 1.3268,
      "theoretical_loss": 3.3668870725765454,
      "tokens_seen": 2801401856
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001526579739217653,
      "loss": 1.1744,
      "theoretical_loss": 3.3668809840018956,
      "tokens_seen": 2801467392
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015263791374122367,
      "loss": 1.2586,
      "theoretical_loss": 3.366874895609557,
      "tokens_seen": 2801532928
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015261785356068206,
      "loss": 1.3905,
      "theoretical_loss": 3.36686880739952,
      "tokens_seen": 2801598464
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015259779338014042,
      "loss": 1.2298,
      "theoretical_loss": 3.3668627193717744,
      "tokens_seen": 2801664000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001525777331995988,
      "loss": 1.2986,
      "theoretical_loss": 3.366856631526311,
      "tokens_seen": 2801729536
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015255767301905717,
      "loss": 1.2971,
      "theoretical_loss": 3.36685054386312,
      "tokens_seen": 2801795072
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015253761283851556,
      "loss": 1.3904,
      "theoretical_loss": 3.366844456382192,
      "tokens_seen": 2801860608
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015251755265797393,
      "loss": 1.2592,
      "theoretical_loss": 3.3668383690835166,
      "tokens_seen": 2801926144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015249749247743232,
      "loss": 1.2643,
      "theoretical_loss": 3.3668322819670844,
      "tokens_seen": 2801991680
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015247743229689068,
      "loss": 1.2977,
      "theoretical_loss": 3.3668261950328855,
      "tokens_seen": 2802057216
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015245737211634904,
      "loss": 1.2586,
      "theoretical_loss": 3.3668201082809106,
      "tokens_seen": 2802122752
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015243731193580743,
      "loss": 1.3086,
      "theoretical_loss": 3.36681402171115,
      "tokens_seen": 2802188288
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001524172517552658,
      "loss": 1.2223,
      "theoretical_loss": 3.3668079353235933,
      "tokens_seen": 2802253824
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015239719157472418,
      "loss": 1.2393,
      "theoretical_loss": 3.3668018491182314,
      "tokens_seen": 2802319360
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015237713139418254,
      "loss": 1.289,
      "theoretical_loss": 3.3667957630950545,
      "tokens_seen": 2802384896
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015235707121364093,
      "loss": 1.35,
      "theoretical_loss": 3.3667896772540526,
      "tokens_seen": 2802450432
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001523370110330993,
      "loss": 1.3308,
      "theoretical_loss": 3.3667835915952162,
      "tokens_seen": 2802515968
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015231695085255768,
      "loss": 1.2928,
      "theoretical_loss": 3.3667775061185354,
      "tokens_seen": 2802581504
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015229689067201605,
      "loss": 1.3302,
      "theoretical_loss": 3.366771420824001,
      "tokens_seen": 2802647040
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001522768304914744,
      "loss": 1.2223,
      "theoretical_loss": 3.3667653357116025,
      "tokens_seen": 2802712576
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001522567703109328,
      "loss": 1.3194,
      "theoretical_loss": 3.366759250781331,
      "tokens_seen": 2802778112
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015223671013039116,
      "loss": 1.3334,
      "theoretical_loss": 3.3667531660331766,
      "tokens_seen": 2802843648
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015221664994984955,
      "loss": 1.3514,
      "theoretical_loss": 3.3667470814671288,
      "tokens_seen": 2802909184
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001521965897693079,
      "loss": 1.2199,
      "theoretical_loss": 3.3667409970831788,
      "tokens_seen": 2802974720
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001521765295887663,
      "loss": 1.2347,
      "theoretical_loss": 3.3667349128813164,
      "tokens_seen": 2803040256
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015215646940822466,
      "loss": 1.3234,
      "theoretical_loss": 3.3667288288615325,
      "tokens_seen": 2803105792
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015213640922768305,
      "loss": 1.2668,
      "theoretical_loss": 3.3667227450238166,
      "tokens_seen": 2803171328
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015211634904714142,
      "loss": 1.2806,
      "theoretical_loss": 3.3667166613681596,
      "tokens_seen": 2803236864
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001520962888665998,
      "loss": 1.2325,
      "theoretical_loss": 3.3667105778945516,
      "tokens_seen": 2803302400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015207622868605817,
      "loss": 1.3488,
      "theoretical_loss": 3.366704494602983,
      "tokens_seen": 2803367936
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1018418,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.270719289779663,
      "objective/train/theoretical_loss": 3.3666984114934437,
      "objective/train/tokens_used": 1173892576,
      "theoretical_loss": 3.3666984114934437,
      "tokens_seen": 2803433472
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015205616850551653,
      "loss": 1.2654,
      "theoretical_loss": 3.3666984114934437,
      "tokens_seen": 2803433472
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015203610832497492,
      "loss": 1.1946,
      "theoretical_loss": 3.3666923285659243,
      "tokens_seen": 2803499008
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001520160481444333,
      "loss": 1.2658,
      "theoretical_loss": 3.366686245820415,
      "tokens_seen": 2803564544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001519959879638917,
      "loss": 1.2379,
      "theoretical_loss": 3.3666801632569063,
      "tokens_seen": 2803630080
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015197592778335006,
      "loss": 1.1718,
      "theoretical_loss": 3.3666740808753883,
      "tokens_seen": 2803695616
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015195586760280845,
      "loss": 1.2403,
      "theoretical_loss": 3.3666679986758514,
      "tokens_seen": 2803761152
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001519358074222668,
      "loss": 1.3886,
      "theoretical_loss": 3.366661916658286,
      "tokens_seen": 2803826688
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001519157472417252,
      "loss": 1.3762,
      "theoretical_loss": 3.366655834822682,
      "tokens_seen": 2803892224
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015189568706118356,
      "loss": 1.3187,
      "theoretical_loss": 3.36664975316903,
      "tokens_seen": 2803957760
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015187562688064192,
      "loss": 1.2707,
      "theoretical_loss": 3.3666436716973203,
      "tokens_seen": 2804023296
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015185556670010031,
      "loss": 1.3147,
      "theoretical_loss": 3.366637590407543,
      "tokens_seen": 2804088832
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015183550651955868,
      "loss": 1.3437,
      "theoretical_loss": 3.366631509299689,
      "tokens_seen": 2804154368
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015181544633901707,
      "loss": 1.3232,
      "theoretical_loss": 3.366625428373748,
      "tokens_seen": 2804219904
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015179538615847543,
      "loss": 1.3253,
      "theoretical_loss": 3.366619347629711,
      "tokens_seen": 2804285440
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015177532597793382,
      "loss": 1.2332,
      "theoretical_loss": 3.366613267067567,
      "tokens_seen": 2804350976
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015175526579739218,
      "loss": 1.2617,
      "theoretical_loss": 3.3666071866873075,
      "tokens_seen": 2804416512
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015173520561685057,
      "loss": 1.2785,
      "theoretical_loss": 3.3666011064889223,
      "tokens_seen": 2804482048
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015171514543630893,
      "loss": 1.3013,
      "theoretical_loss": 3.366595026472402,
      "tokens_seen": 2804547584
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001516950852557673,
      "loss": 1.2512,
      "theoretical_loss": 3.3665889466377363,
      "tokens_seen": 2804613120
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015167502507522568,
      "loss": 1.3265,
      "theoretical_loss": 3.3665828669849165,
      "tokens_seen": 2804678656
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015165496489468404,
      "loss": 1.1703,
      "theoretical_loss": 3.366576787513932,
      "tokens_seen": 2804744192
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015163490471414243,
      "loss": 1.2439,
      "theoretical_loss": 3.366570708224774,
      "tokens_seen": 2804809728
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001516148445336008,
      "loss": 1.2812,
      "theoretical_loss": 3.366564629117432,
      "tokens_seen": 2804875264
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015159478435305919,
      "loss": 1.2602,
      "theoretical_loss": 3.366558550191896,
      "tokens_seen": 2804940800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015157472417251755,
      "loss": 1.307,
      "theoretical_loss": 3.3665524714481574,
      "tokens_seen": 2805006336
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015155466399197594,
      "loss": 1.1957,
      "theoretical_loss": 3.3665463928862063,
      "tokens_seen": 2805071872
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001515346038114343,
      "loss": 1.3938,
      "theoretical_loss": 3.3665403145060324,
      "tokens_seen": 2805137408
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001515145436308927,
      "loss": 1.2709,
      "theoretical_loss": 3.3665342363076265,
      "tokens_seen": 2805202944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015149448345035105,
      "loss": 1.331,
      "theoretical_loss": 3.366528158290979,
      "tokens_seen": 2805268480
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001514744232698094,
      "loss": 1.302,
      "theoretical_loss": 3.36652208045608,
      "tokens_seen": 2805334016
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001514543630892678,
      "loss": 1.3629,
      "theoretical_loss": 3.3665160028029195,
      "tokens_seen": 2805399552
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015143430290872617,
      "loss": 1.2956,
      "theoretical_loss": 3.3665099253314885,
      "tokens_seen": 2805465088
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015141424272818455,
      "loss": 1.2546,
      "theoretical_loss": 3.3665038480417766,
      "tokens_seen": 2805530624
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015139418254764292,
      "loss": 1.3184,
      "theoretical_loss": 3.3664977709337744,
      "tokens_seen": 2805596160
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001513741223671013,
      "loss": 1.2493,
      "theoretical_loss": 3.3664916940074727,
      "tokens_seen": 2805661696
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015135406218655967,
      "loss": 1.3128,
      "theoretical_loss": 3.3664856172628617,
      "tokens_seen": 2805727232
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015133400200601806,
      "loss": 1.1913,
      "theoretical_loss": 3.366479540699931,
      "tokens_seen": 2805792768
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015131394182547642,
      "loss": 1.3337,
      "theoretical_loss": 3.3664734643186716,
      "tokens_seen": 2805858304
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015129388164493478,
      "loss": 1.259,
      "theoretical_loss": 3.3664673881190734,
      "tokens_seen": 2805923840
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001512738214643932,
      "loss": 1.2508,
      "theoretical_loss": 3.366461312101127,
      "tokens_seen": 2805989376
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015125376128385156,
      "loss": 1.3314,
      "theoretical_loss": 3.3664552362648226,
      "tokens_seen": 2806054912
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015123370110330995,
      "loss": 1.3179,
      "theoretical_loss": 3.3664491606101508,
      "tokens_seen": 2806120448
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001512136409227683,
      "loss": 1.3026,
      "theoretical_loss": 3.366443085137101,
      "tokens_seen": 2806185984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001511935807422267,
      "loss": 1.3589,
      "theoretical_loss": 3.3664370098456655,
      "tokens_seen": 2806251520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015117352056168506,
      "loss": 1.2968,
      "theoretical_loss": 3.3664309347358325,
      "tokens_seen": 2806317056
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015115346038114345,
      "loss": 1.2514,
      "theoretical_loss": 3.3664248598075934,
      "tokens_seen": 2806382592
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015113340020060182,
      "loss": 1.2259,
      "theoretical_loss": 3.3664187850609384,
      "tokens_seen": 2806448128
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001511133400200602,
      "loss": 1.2635,
      "theoretical_loss": 3.366412710495858,
      "tokens_seen": 2806513664
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015109327983951857,
      "loss": 1.2431,
      "theoretical_loss": 3.366406636112342,
      "tokens_seen": 2806579200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015107321965897693,
      "loss": 1.2988,
      "theoretical_loss": 3.366400561910381,
      "tokens_seen": 2806644736
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1019695,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.300621747970581,
      "objective/train/theoretical_loss": 3.366394487889966,
      "objective/train/tokens_used": 1177169376,
      "theoretical_loss": 3.366394487889966,
      "tokens_seen": 2806710272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015105315947843532,
      "loss": 1.2623,
      "theoretical_loss": 3.366394487889966,
      "tokens_seen": 2806710272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015103309929789368,
      "loss": 1.2336,
      "theoretical_loss": 3.366388414051086,
      "tokens_seen": 2806775808
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015101303911735207,
      "loss": 1.2963,
      "theoretical_loss": 3.366382340393732,
      "tokens_seen": 2806841344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015099297893681043,
      "loss": 1.2037,
      "theoretical_loss": 3.3663762669178947,
      "tokens_seen": 2806906880
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015097291875626882,
      "loss": 1.2767,
      "theoretical_loss": 3.366370193623564,
      "tokens_seen": 2806972416
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015095285857572718,
      "loss": 1.1683,
      "theoretical_loss": 3.3663641205107306,
      "tokens_seen": 2807037952
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015093279839518557,
      "loss": 1.329,
      "theoretical_loss": 3.3663580475793844,
      "tokens_seen": 2807103488
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015091273821464394,
      "loss": 1.211,
      "theoretical_loss": 3.366351974829516,
      "tokens_seen": 2807169024
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001508926780341023,
      "loss": 1.3223,
      "theoretical_loss": 3.3663459022611155,
      "tokens_seen": 2807234560
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001508726178535607,
      "loss": 1.256,
      "theoretical_loss": 3.366339829874174,
      "tokens_seen": 2807300096
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015085255767301905,
      "loss": 1.2948,
      "theoretical_loss": 3.3663337576686807,
      "tokens_seen": 2807365632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015083249749247744,
      "loss": 1.371,
      "theoretical_loss": 3.3663276856446265,
      "tokens_seen": 2807431168
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001508124373119358,
      "loss": 1.2891,
      "theoretical_loss": 3.3663216138020022,
      "tokens_seen": 2807496704
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001507923771313942,
      "loss": 1.3914,
      "theoretical_loss": 3.3663155421407973,
      "tokens_seen": 2807562240
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015077231695085255,
      "loss": 1.3284,
      "theoretical_loss": 3.3663094706610024,
      "tokens_seen": 2807627776
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015075225677031094,
      "loss": 1.2942,
      "theoretical_loss": 3.366303399362608,
      "tokens_seen": 2807693312
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001507321965897693,
      "loss": 1.3384,
      "theoretical_loss": 3.366297328245605,
      "tokens_seen": 2807758848
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015071213640922767,
      "loss": 1.24,
      "theoretical_loss": 3.3662912573099826,
      "tokens_seen": 2807824384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015069207622868606,
      "loss": 1.2714,
      "theoretical_loss": 3.366285186555732,
      "tokens_seen": 2807889920
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015067201604814442,
      "loss": 1.2375,
      "theoretical_loss": 3.3662791159828434,
      "tokens_seen": 2807955456
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001506519558676028,
      "loss": 1.2936,
      "theoretical_loss": 3.366273045591307,
      "tokens_seen": 2808020992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015063189568706117,
      "loss": 1.2367,
      "theoretical_loss": 3.3662669753811127,
      "tokens_seen": 2808086528
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015061183550651956,
      "loss": 1.2781,
      "theoretical_loss": 3.3662609053522514,
      "tokens_seen": 2808152064
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015059177532597792,
      "loss": 1.3548,
      "theoretical_loss": 3.366254835504714,
      "tokens_seen": 2808217600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001505717151454363,
      "loss": 1.238,
      "theoretical_loss": 3.3662487658384896,
      "tokens_seen": 2808283136
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015055165496489467,
      "loss": 1.2755,
      "theoretical_loss": 3.366242696353569,
      "tokens_seen": 2808348672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015053159478435304,
      "loss": 1.2329,
      "theoretical_loss": 3.3662366270499433,
      "tokens_seen": 2808414208
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015051153460381145,
      "loss": 1.2944,
      "theoretical_loss": 3.366230557927602,
      "tokens_seen": 2808479744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015049147442326981,
      "loss": 1.3628,
      "theoretical_loss": 3.366224488986536,
      "tokens_seen": 2808545280
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001504714142427282,
      "loss": 1.334,
      "theoretical_loss": 3.366218420226735,
      "tokens_seen": 2808610816
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015045135406218657,
      "loss": 1.3177,
      "theoretical_loss": 3.3662123516481905,
      "tokens_seen": 2808676352
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015043129388164495,
      "loss": 1.3237,
      "theoretical_loss": 3.366206283250891,
      "tokens_seen": 2808741888
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015041123370110332,
      "loss": 1.2799,
      "theoretical_loss": 3.366200215034829,
      "tokens_seen": 2808807424
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001503911735205617,
      "loss": 1.291,
      "theoretical_loss": 3.3661941469999936,
      "tokens_seen": 2808872960
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015037111334002007,
      "loss": 1.3049,
      "theoretical_loss": 3.366188079146375,
      "tokens_seen": 2808938496
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015035105315947846,
      "loss": 1.2636,
      "theoretical_loss": 3.3661820114739642,
      "tokens_seen": 2809004032
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015033099297893682,
      "loss": 1.2467,
      "theoretical_loss": 3.3661759439827508,
      "tokens_seen": 2809069568
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015031093279839518,
      "loss": 1.2171,
      "theoretical_loss": 3.366169876672726,
      "tokens_seen": 2809135104
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015029087261785357,
      "loss": 1.1793,
      "theoretical_loss": 3.36616380954388,
      "tokens_seen": 2809200640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015027081243731193,
      "loss": 1.1923,
      "theoretical_loss": 3.366157742596203,
      "tokens_seen": 2809266176
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015025075225677032,
      "loss": 1.3145,
      "theoretical_loss": 3.3661516758296854,
      "tokens_seen": 2809331712
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015023069207622869,
      "loss": 1.3295,
      "theoretical_loss": 3.3661456092443176,
      "tokens_seen": 2809397248
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015021063189568707,
      "loss": 1.3785,
      "theoretical_loss": 3.3661395428400898,
      "tokens_seen": 2809462784
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015019057171514544,
      "loss": 1.2697,
      "theoretical_loss": 3.366133476616992,
      "tokens_seen": 2809528320
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015017051153460383,
      "loss": 1.3382,
      "theoretical_loss": 3.3661274105750154,
      "tokens_seen": 2809593856
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001501504513540622,
      "loss": 1.2608,
      "theoretical_loss": 3.3661213447141503,
      "tokens_seen": 2809659392
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015013039117352055,
      "loss": 1.326,
      "theoretical_loss": 3.3661152790343865,
      "tokens_seen": 2809724928
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015011033099297894,
      "loss": 1.1812,
      "theoretical_loss": 3.3661092135357142,
      "tokens_seen": 2809790464
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001500902708124373,
      "loss": 1.3251,
      "theoretical_loss": 3.3661031482181247,
      "tokens_seen": 2809856000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001500702106318957,
      "loss": 1.3088,
      "theoretical_loss": 3.3660970830816077,
      "tokens_seen": 2809921536
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1020904,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3573384284973145,
      "objective/train/theoretical_loss": 3.366091018126154,
      "objective/train/tokens_used": 1180446176,
      "theoretical_loss": 3.366091018126154,
      "tokens_seen": 2809987072
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015005015045135405,
      "loss": 1.3532,
      "theoretical_loss": 3.366091018126154,
      "tokens_seen": 2809987072
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015003009027081244,
      "loss": 1.2438,
      "theoretical_loss": 3.366084953351753,
      "tokens_seen": 2810052608
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001500100300902708,
      "loss": 1.3242,
      "theoretical_loss": 3.3660788887583966,
      "tokens_seen": 2810118144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001499899699097292,
      "loss": 1.2805,
      "theoretical_loss": 3.366072824346074,
      "tokens_seen": 2810183680
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014996990972918756,
      "loss": 1.3058,
      "theoretical_loss": 3.3660667601147756,
      "tokens_seen": 2810249216
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014994984954864595,
      "loss": 1.2142,
      "theoretical_loss": 3.3660606960644928,
      "tokens_seen": 2810314752
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001499297893681043,
      "loss": 1.2758,
      "theoretical_loss": 3.366054632195215,
      "tokens_seen": 2810380288
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014990972918756267,
      "loss": 1.2353,
      "theoretical_loss": 3.3660485685069323,
      "tokens_seen": 2810445824
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014988966900702106,
      "loss": 1.2517,
      "theoretical_loss": 3.3660425049996365,
      "tokens_seen": 2810511360
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014986960882647942,
      "loss": 1.3745,
      "theoretical_loss": 3.366036441673317,
      "tokens_seen": 2810576896
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001498495486459378,
      "loss": 1.2953,
      "theoretical_loss": 3.3660303785279635,
      "tokens_seen": 2810642432
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014982948846539617,
      "loss": 1.2827,
      "theoretical_loss": 3.3660243155635676,
      "tokens_seen": 2810707968
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014980942828485456,
      "loss": 1.3026,
      "theoretical_loss": 3.3660182527801195,
      "tokens_seen": 2810773504
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014978936810431293,
      "loss": 1.2254,
      "theoretical_loss": 3.366012190177609,
      "tokens_seen": 2810839040
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014976930792377134,
      "loss": 1.3262,
      "theoretical_loss": 3.3660061277560276,
      "tokens_seen": 2810904576
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001497492477432297,
      "loss": 1.3479,
      "theoretical_loss": 3.366000065515364,
      "tokens_seen": 2810970112
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014972918756268807,
      "loss": 1.207,
      "theoretical_loss": 3.3659940034556097,
      "tokens_seen": 2811035648
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014970912738214646,
      "loss": 1.3488,
      "theoretical_loss": 3.365987941576755,
      "tokens_seen": 2811101184
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014968906720160482,
      "loss": 1.3395,
      "theoretical_loss": 3.36598187987879,
      "tokens_seen": 2811166720
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001496690070210632,
      "loss": 1.3602,
      "theoretical_loss": 3.3659758183617057,
      "tokens_seen": 2811232256
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014964894684052157,
      "loss": 1.364,
      "theoretical_loss": 3.3659697570254914,
      "tokens_seen": 2811297792
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014962888665997996,
      "loss": 1.3224,
      "theoretical_loss": 3.365963695870138,
      "tokens_seen": 2811363328
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014960882647943832,
      "loss": 1.242,
      "theoretical_loss": 3.365957634895637,
      "tokens_seen": 2811428864
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001495887662988967,
      "loss": 1.4712,
      "theoretical_loss": 3.365951574101977,
      "tokens_seen": 2811494400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014956870611835507,
      "loss": 1.3002,
      "theoretical_loss": 3.365945513489149,
      "tokens_seen": 2811559936
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014954864593781344,
      "loss": 1.3035,
      "theoretical_loss": 3.365939453057144,
      "tokens_seen": 2811625472
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014952858575727182,
      "loss": 1.2132,
      "theoretical_loss": 3.365933392805952,
      "tokens_seen": 2811691008
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001495085255767302,
      "loss": 1.283,
      "theoretical_loss": 3.3659273327355628,
      "tokens_seen": 2811756544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014948846539618858,
      "loss": 1.1784,
      "theoretical_loss": 3.3659212728459678,
      "tokens_seen": 2811822080
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014946840521564694,
      "loss": 1.2441,
      "theoretical_loss": 3.365915213137157,
      "tokens_seen": 2811887616
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014944834503510533,
      "loss": 1.3213,
      "theoretical_loss": 3.3659091536091204,
      "tokens_seen": 2811953152
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001494282848545637,
      "loss": 1.3185,
      "theoretical_loss": 3.365903094261849,
      "tokens_seen": 2812018688
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014940822467402208,
      "loss": 1.2277,
      "theoretical_loss": 3.365897035095333,
      "tokens_seen": 2812084224
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014938816449348044,
      "loss": 1.3234,
      "theoretical_loss": 3.3658909761095623,
      "tokens_seen": 2812149760
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014936810431293883,
      "loss": 1.1999,
      "theoretical_loss": 3.3658849173045278,
      "tokens_seen": 2812215296
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001493480441323972,
      "loss": 1.3286,
      "theoretical_loss": 3.3658788586802197,
      "tokens_seen": 2812280832
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014932798395185556,
      "loss": 1.2499,
      "theoretical_loss": 3.3658728002366285,
      "tokens_seen": 2812346368
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014930792377131394,
      "loss": 1.3928,
      "theoretical_loss": 3.365866741973745,
      "tokens_seen": 2812411904
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001492878635907723,
      "loss": 1.3609,
      "theoretical_loss": 3.365860683891559,
      "tokens_seen": 2812477440
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001492678034102307,
      "loss": 1.2247,
      "theoretical_loss": 3.3658546259900612,
      "tokens_seen": 2812542976
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014924774322968906,
      "loss": 1.2921,
      "theoretical_loss": 3.3658485682692416,
      "tokens_seen": 2812608512
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014922768304914745,
      "loss": 1.3862,
      "theoretical_loss": 3.365842510729091,
      "tokens_seen": 2812674048
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001492076228686058,
      "loss": 1.3808,
      "theoretical_loss": 3.3658364533696,
      "tokens_seen": 2812739584
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001491875626880642,
      "loss": 1.3087,
      "theoretical_loss": 3.3658303961907583,
      "tokens_seen": 2812805120
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014916750250752256,
      "loss": 1.3023,
      "theoretical_loss": 3.3658243391925566,
      "tokens_seen": 2812870656
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014914744232698092,
      "loss": 1.2785,
      "theoretical_loss": 3.3658182823749856,
      "tokens_seen": 2812936192
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001491273821464393,
      "loss": 1.3601,
      "theoretical_loss": 3.365812225738036,
      "tokens_seen": 2813001728
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014910732196589768,
      "loss": 1.2484,
      "theoretical_loss": 3.365806169281697,
      "tokens_seen": 2813067264
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014908726178535607,
      "loss": 1.2908,
      "theoretical_loss": 3.3658001130059603,
      "tokens_seen": 2813132800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014906720160481443,
      "loss": 1.3212,
      "theoretical_loss": 3.3657940569108153,
      "tokens_seen": 2813198336
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1022064,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3037389516830444,
      "objective/train/theoretical_loss": 3.365788000996253,
      "objective/train/tokens_used": 1183722976,
      "theoretical_loss": 3.365788000996253,
      "tokens_seen": 2813263872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014904714142427282,
      "loss": 1.2881,
      "theoretical_loss": 3.365788000996253,
      "tokens_seen": 2813263872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014902708124373118,
      "loss": 1.247,
      "theoretical_loss": 3.3657819452622637,
      "tokens_seen": 2813329408
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001490070210631896,
      "loss": 1.3213,
      "theoretical_loss": 3.3657758897088375,
      "tokens_seen": 2813394944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014898696088264796,
      "loss": 1.3377,
      "theoretical_loss": 3.3657698343359654,
      "tokens_seen": 2813460480
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014896690070210635,
      "loss": 1.347,
      "theoretical_loss": 3.3657637791436374,
      "tokens_seen": 2813526016
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001489468405215647,
      "loss": 1.2703,
      "theoretical_loss": 3.3657577241318437,
      "tokens_seen": 2813591552
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014892678034102307,
      "loss": 1.2967,
      "theoretical_loss": 3.3657516693005753,
      "tokens_seen": 2813657088
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014890672016048146,
      "loss": 1.3306,
      "theoretical_loss": 3.365745614649822,
      "tokens_seen": 2813722624
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014888665997993982,
      "loss": 1.1877,
      "theoretical_loss": 3.365739560179575,
      "tokens_seen": 2813788160
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001488665997993982,
      "loss": 1.2836,
      "theoretical_loss": 3.3657335058898243,
      "tokens_seen": 2813853696
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014884653961885657,
      "loss": 1.2961,
      "theoretical_loss": 3.36572745178056,
      "tokens_seen": 2813919232
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014882647943831496,
      "loss": 1.1976,
      "theoretical_loss": 3.365721397851773,
      "tokens_seen": 2813984768
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014880641925777333,
      "loss": 1.2404,
      "theoretical_loss": 3.365715344103453,
      "tokens_seen": 2814050304
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014878635907723172,
      "loss": 1.351,
      "theoretical_loss": 3.3657092905355914,
      "tokens_seen": 2814115840
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014876629889669008,
      "loss": 1.2728,
      "theoretical_loss": 3.365703237148178,
      "tokens_seen": 2814181376
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014874623871614844,
      "loss": 1.3906,
      "theoretical_loss": 3.365697183941203,
      "tokens_seen": 2814246912
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014872617853560683,
      "loss": 1.2368,
      "theoretical_loss": 3.3656911309146578,
      "tokens_seen": 2814312448
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001487061183550652,
      "loss": 1.2756,
      "theoretical_loss": 3.3656850780685317,
      "tokens_seen": 2814377984
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014868605817452358,
      "loss": 1.2397,
      "theoretical_loss": 3.365679025402816,
      "tokens_seen": 2814443520
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014866599799398194,
      "loss": 1.1949,
      "theoretical_loss": 3.3656729729175003,
      "tokens_seen": 2814509056
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014864593781344033,
      "loss": 1.2854,
      "theoretical_loss": 3.3656669206125756,
      "tokens_seen": 2814574592
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001486258776328987,
      "loss": 1.2674,
      "theoretical_loss": 3.3656608684880323,
      "tokens_seen": 2814640128
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014860581745235708,
      "loss": 1.3069,
      "theoretical_loss": 3.3656548165438607,
      "tokens_seen": 2814705664
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014858575727181545,
      "loss": 1.3918,
      "theoretical_loss": 3.365648764780051,
      "tokens_seen": 2814771200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001485656970912738,
      "loss": 1.2786,
      "theoretical_loss": 3.3656427131965945,
      "tokens_seen": 2814836736
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001485456369107322,
      "loss": 1.3064,
      "theoretical_loss": 3.3656366617934803,
      "tokens_seen": 2814902272
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014852557673019056,
      "loss": 1.2267,
      "theoretical_loss": 3.3656306105706997,
      "tokens_seen": 2814967808
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014850551654964895,
      "loss": 1.2111,
      "theoretical_loss": 3.365624559528243,
      "tokens_seen": 2815033344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001484854563691073,
      "loss": 1.2221,
      "theoretical_loss": 3.3656185086661004,
      "tokens_seen": 2815098880
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001484653961885657,
      "loss": 1.3426,
      "theoretical_loss": 3.3656124579842626,
      "tokens_seen": 2815164416
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014844533600802406,
      "loss": 1.2264,
      "theoretical_loss": 3.36560640748272,
      "tokens_seen": 2815229952
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014842527582748245,
      "loss": 1.2896,
      "theoretical_loss": 3.3656003571614628,
      "tokens_seen": 2815295488
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014840521564694081,
      "loss": 1.3565,
      "theoretical_loss": 3.3655943070204817,
      "tokens_seen": 2815361024
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001483851554663992,
      "loss": 1.4383,
      "theoretical_loss": 3.365588257059767,
      "tokens_seen": 2815426560
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014836509528585757,
      "loss": 1.2021,
      "theoretical_loss": 3.365582207279309,
      "tokens_seen": 2815492096
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014834503510531593,
      "loss": 1.4061,
      "theoretical_loss": 3.3655761576790986,
      "tokens_seen": 2815557632
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014832497492477432,
      "loss": 1.3142,
      "theoretical_loss": 3.365570108259126,
      "tokens_seen": 2815623168
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014830491474423268,
      "loss": 1.2672,
      "theoretical_loss": 3.365564059019381,
      "tokens_seen": 2815688704
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014828485456369107,
      "loss": 1.2992,
      "theoretical_loss": 3.365558009959855,
      "tokens_seen": 2815754240
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014826479438314946,
      "loss": 1.4026,
      "theoretical_loss": 3.365551961080538,
      "tokens_seen": 2815819776
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014824473420260785,
      "loss": 1.3509,
      "theoretical_loss": 3.36554591238142,
      "tokens_seen": 2815885312
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001482246740220662,
      "loss": 1.2679,
      "theoretical_loss": 3.3655398638624927,
      "tokens_seen": 2815950848
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001482046138415246,
      "loss": 1.3021,
      "theoretical_loss": 3.3655338155237455,
      "tokens_seen": 2816016384
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014818455366098296,
      "loss": 1.2219,
      "theoretical_loss": 3.3655277673651685,
      "tokens_seen": 2816081920
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014816449348044132,
      "loss": 1.1718,
      "theoretical_loss": 3.3655217193867535,
      "tokens_seen": 2816147456
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014814443329989971,
      "loss": 1.3033,
      "theoretical_loss": 3.3655156715884895,
      "tokens_seen": 2816212992
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014812437311935808,
      "loss": 1.2286,
      "theoretical_loss": 3.365509623970368,
      "tokens_seen": 2816278528
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014810431293881647,
      "loss": 1.2077,
      "theoretical_loss": 3.365503576532379,
      "tokens_seen": 2816344064
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014808425275827483,
      "loss": 1.2522,
      "theoretical_loss": 3.365497529274513,
      "tokens_seen": 2816409600
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014806419257773322,
      "loss": 1.4091,
      "theoretical_loss": 3.3654914821967603,
      "tokens_seen": 2816475136
    },
    {
      "epoch": 0.71,
      "objective/train/docs_used": 1023297,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.332075834274292,
      "objective/train/theoretical_loss": 3.3654854352991115,
      "objective/train/tokens_used": 1186999776,
      "theoretical_loss": 3.3654854352991115,
      "tokens_seen": 2816540672
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014804413239719158,
      "loss": 1.3627,
      "theoretical_loss": 3.3654854352991115,
      "tokens_seen": 2816540672
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014802407221664997,
      "loss": 1.3526,
      "theoretical_loss": 3.3654793885815573,
      "tokens_seen": 2816606208
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014800401203610833,
      "loss": 1.3273,
      "theoretical_loss": 3.3654733420440874,
      "tokens_seen": 2816671744
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001479839518555667,
      "loss": 1.2575,
      "theoretical_loss": 3.365467295686693,
      "tokens_seen": 2816737280
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014796389167502508,
      "loss": 1.2802,
      "theoretical_loss": 3.365461249509364,
      "tokens_seen": 2816802816
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014794383149448344,
      "loss": 1.2967,
      "theoretical_loss": 3.3654552035120915,
      "tokens_seen": 2816868352
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014792377131394183,
      "loss": 1.3133,
      "theoretical_loss": 3.3654491576948655,
      "tokens_seen": 2816933888
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001479037111334002,
      "loss": 1.3794,
      "theoretical_loss": 3.3654431120576764,
      "tokens_seen": 2816999424
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014788365095285859,
      "loss": 1.2313,
      "theoretical_loss": 3.365437066600515,
      "tokens_seen": 2817064960
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014786359077231695,
      "loss": 1.3032,
      "theoretical_loss": 3.365431021323371,
      "tokens_seen": 2817130496
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014784353059177534,
      "loss": 1.1861,
      "theoretical_loss": 3.365424976226236,
      "tokens_seen": 2817196032
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001478234704112337,
      "loss": 1.3505,
      "theoretical_loss": 3.365418931309099,
      "tokens_seen": 2817261568
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001478034102306921,
      "loss": 1.2523,
      "theoretical_loss": 3.365412886571952,
      "tokens_seen": 2817327104
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014778335005015045,
      "loss": 1.21,
      "theoretical_loss": 3.3654068420147842,
      "tokens_seen": 2817392640
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001477632898696088,
      "loss": 1.3126,
      "theoretical_loss": 3.3654007976375873,
      "tokens_seen": 2817458176
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001477432296890672,
      "loss": 1.3921,
      "theoretical_loss": 3.3653947534403503,
      "tokens_seen": 2817523712
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014772316950852556,
      "loss": 1.3034,
      "theoretical_loss": 3.3653887094230646,
      "tokens_seen": 2817589248
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014770310932798395,
      "loss": 1.26,
      "theoretical_loss": 3.3653826655857206,
      "tokens_seen": 2817654784
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014768304914744232,
      "loss": 1.2252,
      "theoretical_loss": 3.3653766219283083,
      "tokens_seen": 2817720320
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001476629889669007,
      "loss": 1.2532,
      "theoretical_loss": 3.3653705784508188,
      "tokens_seen": 2817785856
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014764292878635907,
      "loss": 1.4185,
      "theoretical_loss": 3.365364535153242,
      "tokens_seen": 2817851392
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014762286860581746,
      "loss": 1.3713,
      "theoretical_loss": 3.365358492035569,
      "tokens_seen": 2817916928
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014760280842527582,
      "loss": 1.3146,
      "theoretical_loss": 3.3653524490977893,
      "tokens_seen": 2817982464
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014758274824473418,
      "loss": 1.2004,
      "theoretical_loss": 3.3653464063398943,
      "tokens_seen": 2818048000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014756268806419257,
      "loss": 1.2827,
      "theoretical_loss": 3.3653403637618737,
      "tokens_seen": 2818113536
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014754262788365093,
      "loss": 1.4074,
      "theoretical_loss": 3.3653343213637186,
      "tokens_seen": 2818179072
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014752256770310932,
      "loss": 1.2958,
      "theoretical_loss": 3.365328279145419,
      "tokens_seen": 2818244608
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001475025075225677,
      "loss": 1.2943,
      "theoretical_loss": 3.365322237106966,
      "tokens_seen": 2818310144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001474824473420261,
      "loss": 1.3692,
      "theoretical_loss": 3.365316195248349,
      "tokens_seen": 2818375680
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014746238716148446,
      "loss": 1.318,
      "theoretical_loss": 3.3653101535695598,
      "tokens_seen": 2818441216
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014744232698094285,
      "loss": 1.201,
      "theoretical_loss": 3.3653041120705875,
      "tokens_seen": 2818506752
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014742226680040122,
      "loss": 1.1917,
      "theoretical_loss": 3.3652980707514235,
      "tokens_seen": 2818572288
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001474022066198596,
      "loss": 1.223,
      "theoretical_loss": 3.365292029612058,
      "tokens_seen": 2818637824
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014738214643931797,
      "loss": 1.2914,
      "theoretical_loss": 3.3652859886524813,
      "tokens_seen": 2818703360
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014736208625877633,
      "loss": 1.3589,
      "theoretical_loss": 3.3652799478726845,
      "tokens_seen": 2818768896
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014734202607823472,
      "loss": 1.3577,
      "theoretical_loss": 3.365273907272657,
      "tokens_seen": 2818834432
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014732196589769308,
      "loss": 1.2074,
      "theoretical_loss": 3.36526786685239,
      "tokens_seen": 2818899968
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014730190571715147,
      "loss": 1.2743,
      "theoretical_loss": 3.3652618266118743,
      "tokens_seen": 2818965504
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014728184553660983,
      "loss": 1.3328,
      "theoretical_loss": 3.3652557865510992,
      "tokens_seen": 2819031040
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014726178535606822,
      "loss": 1.2639,
      "theoretical_loss": 3.3652497466700564,
      "tokens_seen": 2819096576
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014724172517552658,
      "loss": 1.2576,
      "theoretical_loss": 3.3652437069687355,
      "tokens_seen": 2819162112
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014722166499498497,
      "loss": 1.2791,
      "theoretical_loss": 3.365237667447128,
      "tokens_seen": 2819227648
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014720160481444334,
      "loss": 1.3342,
      "theoretical_loss": 3.365231628105223,
      "tokens_seen": 2819293184
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001471815446339017,
      "loss": 1.3723,
      "theoretical_loss": 3.3652255889430123,
      "tokens_seen": 2819358720
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001471614844533601,
      "loss": 1.3382,
      "theoretical_loss": 3.365219549960485,
      "tokens_seen": 2819424256
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014714142427281845,
      "loss": 1.2768,
      "theoretical_loss": 3.365213511157633,
      "tokens_seen": 2819489792
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014712136409227684,
      "loss": 1.2425,
      "theoretical_loss": 3.3652074725344456,
      "tokens_seen": 2819555328
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001471013039117352,
      "loss": 1.3266,
      "theoretical_loss": 3.365201434090914,
      "tokens_seen": 2819620864
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001470812437311936,
      "loss": 1.2581,
      "theoretical_loss": 3.365195395827029,
      "tokens_seen": 2819686400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014706118355065195,
      "loss": 1.2493,
      "theoretical_loss": 3.36518935774278,
      "tokens_seen": 2819751936
    },
    {
      "epoch": 0.71,
      "objective/train/docs_used": 1024583,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.230042815208435,
      "objective/train/theoretical_loss": 3.3651833198381578,
      "objective/train/tokens_used": 1190276576,
      "theoretical_loss": 3.3651833198381578,
      "tokens_seen": 2819817472
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014704112337011034,
      "loss": 1.2822,
      "theoretical_loss": 3.3651833198381578,
      "tokens_seen": 2819817472
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001470210631895687,
      "loss": 1.2733,
      "theoretical_loss": 3.3651772821131534,
      "tokens_seen": 2819883008
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014700100300902707,
      "loss": 1.3366,
      "theoretical_loss": 3.3651712445677573,
      "tokens_seen": 2819948544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014698094282848546,
      "loss": 1.2138,
      "theoretical_loss": 3.3651652072019593,
      "tokens_seen": 2820014080
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014696088264794382,
      "loss": 1.2268,
      "theoretical_loss": 3.3651591700157506,
      "tokens_seen": 2820079616
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001469408224674022,
      "loss": 1.2365,
      "theoretical_loss": 3.365153133009121,
      "tokens_seen": 2820145152
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014692076228686057,
      "loss": 1.3833,
      "theoretical_loss": 3.3651470961820618,
      "tokens_seen": 2820210688
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014690070210631896,
      "loss": 1.2667,
      "theoretical_loss": 3.3651410595345626,
      "tokens_seen": 2820276224
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014688064192577732,
      "loss": 1.3036,
      "theoretical_loss": 3.365135023066614,
      "tokens_seen": 2820341760
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001468605817452357,
      "loss": 1.2481,
      "theoretical_loss": 3.3651289867782075,
      "tokens_seen": 2820407296
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014684052156469407,
      "loss": 1.2153,
      "theoretical_loss": 3.3651229506693325,
      "tokens_seen": 2820472832
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014682046138415243,
      "loss": 1.2854,
      "theoretical_loss": 3.36511691473998,
      "tokens_seen": 2820538368
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014680040120361082,
      "loss": 1.2934,
      "theoretical_loss": 3.3651108789901403,
      "tokens_seen": 2820603904
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001467803410230692,
      "loss": 1.3354,
      "theoretical_loss": 3.365104843419804,
      "tokens_seen": 2820669440
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001467602808425276,
      "loss": 1.3075,
      "theoretical_loss": 3.3650988080289617,
      "tokens_seen": 2820734976
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014674022066198597,
      "loss": 1.2173,
      "theoretical_loss": 3.3650927728176034,
      "tokens_seen": 2820800512
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014672016048144435,
      "loss": 1.3617,
      "theoretical_loss": 3.3650867377857203,
      "tokens_seen": 2820866048
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014670010030090272,
      "loss": 1.2539,
      "theoretical_loss": 3.365080702933302,
      "tokens_seen": 2820931584
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001466800401203611,
      "loss": 1.2408,
      "theoretical_loss": 3.3650746682603403,
      "tokens_seen": 2820997120
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014665997993981947,
      "loss": 1.237,
      "theoretical_loss": 3.3650686337668243,
      "tokens_seen": 2821062656
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014663991975927786,
      "loss": 1.2972,
      "theoretical_loss": 3.365062599452745,
      "tokens_seen": 2821128192
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014661985957873622,
      "loss": 1.2806,
      "theoretical_loss": 3.3650565653180937,
      "tokens_seen": 2821193728
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014659979939819458,
      "loss": 1.3165,
      "theoretical_loss": 3.3650505313628596,
      "tokens_seen": 2821259264
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014657973921765297,
      "loss": 1.1982,
      "theoretical_loss": 3.365044497587034,
      "tokens_seen": 2821324800
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014655967903711133,
      "loss": 1.2624,
      "theoretical_loss": 3.365038463990607,
      "tokens_seen": 2821390336
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014653961885656972,
      "loss": 1.2847,
      "theoretical_loss": 3.36503243057357,
      "tokens_seen": 2821455872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014651955867602809,
      "loss": 1.3052,
      "theoretical_loss": 3.365026397335912,
      "tokens_seen": 2821521408
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014649949849548647,
      "loss": 1.3114,
      "theoretical_loss": 3.3650203642776244,
      "tokens_seen": 2821586944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014647943831494484,
      "loss": 1.3226,
      "theoretical_loss": 3.3650143313986978,
      "tokens_seen": 2821652480
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014645937813440323,
      "loss": 1.2237,
      "theoretical_loss": 3.3650082986991228,
      "tokens_seen": 2821718016
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001464393179538616,
      "loss": 1.2157,
      "theoretical_loss": 3.365002266178889,
      "tokens_seen": 2821783552
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014641925777331995,
      "loss": 1.2968,
      "theoretical_loss": 3.3649962338379877,
      "tokens_seen": 2821849088
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014639919759277834,
      "loss": 1.3028,
      "theoretical_loss": 3.3649902016764095,
      "tokens_seen": 2821914624
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001463791374122367,
      "loss": 1.3197,
      "theoretical_loss": 3.3649841696941443,
      "tokens_seen": 2821980160
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001463590772316951,
      "loss": 1.3784,
      "theoretical_loss": 3.364978137891183,
      "tokens_seen": 2822045696
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014633901705115345,
      "loss": 1.2514,
      "theoretical_loss": 3.3649721062675164,
      "tokens_seen": 2822111232
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014631895687061184,
      "loss": 1.3369,
      "theoretical_loss": 3.3649660748231343,
      "tokens_seen": 2822176768
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001462988966900702,
      "loss": 1.3616,
      "theoretical_loss": 3.3649600435580274,
      "tokens_seen": 2822242304
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001462788365095286,
      "loss": 1.2211,
      "theoretical_loss": 3.3649540124721864,
      "tokens_seen": 2822307840
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014625877632898696,
      "loss": 1.2585,
      "theoretical_loss": 3.364947981565602,
      "tokens_seen": 2822373376
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014623871614844535,
      "loss": 1.3647,
      "theoretical_loss": 3.364941950838264,
      "tokens_seen": 2822438912
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001462186559679037,
      "loss": 1.3117,
      "theoretical_loss": 3.364935920290164,
      "tokens_seen": 2822504448
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014619859578736207,
      "loss": 1.2697,
      "theoretical_loss": 3.3649298899212914,
      "tokens_seen": 2822569984
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014617853560682046,
      "loss": 1.2733,
      "theoretical_loss": 3.3649238597316375,
      "tokens_seen": 2822635520
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014615847542627882,
      "loss": 1.3767,
      "theoretical_loss": 3.3649178297211924,
      "tokens_seen": 2822701056
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001461384152457372,
      "loss": 1.1376,
      "theoretical_loss": 3.3649117998899465,
      "tokens_seen": 2822766592
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014611835506519557,
      "loss": 1.3238,
      "theoretical_loss": 3.364905770237891,
      "tokens_seen": 2822832128
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014609829488465396,
      "loss": 1.2654,
      "theoretical_loss": 3.364899740765016,
      "tokens_seen": 2822897664
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014607823470411233,
      "loss": 1.3341,
      "theoretical_loss": 3.3648937114713116,
      "tokens_seen": 2822963200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014605817452357072,
      "loss": 1.4008,
      "theoretical_loss": 3.364887682356769,
      "tokens_seen": 2823028736
    },
    {
      "epoch": 0.71,
      "objective/train/docs_used": 1025771,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2943767309188843,
      "objective/train/theoretical_loss": 3.364881653421378,
      "objective/train/tokens_used": 1193553376,
      "theoretical_loss": 3.364881653421378,
      "tokens_seen": 2823094272
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014603811434302908,
      "loss": 1.3036,
      "theoretical_loss": 3.364881653421378,
      "tokens_seen": 2823094272
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014601805416248747,
      "loss": 1.4194,
      "theoretical_loss": 3.36487562466513,
      "tokens_seen": 2823159808
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014599799398194586,
      "loss": 1.2723,
      "theoretical_loss": 3.364869596088014,
      "tokens_seen": 2823225344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014597793380140422,
      "loss": 1.29,
      "theoretical_loss": 3.3648635676900227,
      "tokens_seen": 2823290880
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001459578736208626,
      "loss": 1.3217,
      "theoretical_loss": 3.3648575394711453,
      "tokens_seen": 2823356416
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014593781344032097,
      "loss": 1.2406,
      "theoretical_loss": 3.364851511431372,
      "tokens_seen": 2823421952
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014591775325977936,
      "loss": 1.392,
      "theoretical_loss": 3.3648454835706945,
      "tokens_seen": 2823487488
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014589769307923772,
      "loss": 1.333,
      "theoretical_loss": 3.3648394558891024,
      "tokens_seen": 2823553024
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001458776328986961,
      "loss": 1.3446,
      "theoretical_loss": 3.3648334283865866,
      "tokens_seen": 2823618560
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014585757271815447,
      "loss": 1.3201,
      "theoretical_loss": 3.364827401063137,
      "tokens_seen": 2823684096
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014583751253761284,
      "loss": 1.3109,
      "theoretical_loss": 3.364821373918745,
      "tokens_seen": 2823749632
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014581745235707122,
      "loss": 1.1853,
      "theoretical_loss": 3.364815346953401,
      "tokens_seen": 2823815168
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001457973921765296,
      "loss": 1.2411,
      "theoretical_loss": 3.3648093201670948,
      "tokens_seen": 2823880704
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014577733199598798,
      "loss": 1.2904,
      "theoretical_loss": 3.3648032935598176,
      "tokens_seen": 2823946240
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014575727181544634,
      "loss": 1.2688,
      "theoretical_loss": 3.3647972671315594,
      "tokens_seen": 2824011776
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014573721163490473,
      "loss": 1.3251,
      "theoretical_loss": 3.3647912408823113,
      "tokens_seen": 2824077312
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001457171514543631,
      "loss": 1.3927,
      "theoretical_loss": 3.3647852148120636,
      "tokens_seen": 2824142848
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014569709127382148,
      "loss": 1.277,
      "theoretical_loss": 3.3647791889208065,
      "tokens_seen": 2824208384
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014567703109327984,
      "loss": 1.341,
      "theoretical_loss": 3.3647731632085316,
      "tokens_seen": 2824273920
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014565697091273823,
      "loss": 1.2855,
      "theoretical_loss": 3.364767137675228,
      "tokens_seen": 2824339456
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001456369107321966,
      "loss": 1.3014,
      "theoretical_loss": 3.3647611123208874,
      "tokens_seen": 2824404992
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014561685055165496,
      "loss": 1.2117,
      "theoretical_loss": 3.3647550871454994,
      "tokens_seen": 2824470528
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014559679037111334,
      "loss": 1.2314,
      "theoretical_loss": 3.364749062149055,
      "tokens_seen": 2824536064
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001455767301905717,
      "loss": 1.2678,
      "theoretical_loss": 3.364743037331545,
      "tokens_seen": 2824601600
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001455566700100301,
      "loss": 1.3055,
      "theoretical_loss": 3.3647370126929594,
      "tokens_seen": 2824667136
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014553660982948846,
      "loss": 1.1507,
      "theoretical_loss": 3.364730988233289,
      "tokens_seen": 2824732672
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014551654964894685,
      "loss": 1.2466,
      "theoretical_loss": 3.3647249639525243,
      "tokens_seen": 2824798208
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001454964894684052,
      "loss": 1.3325,
      "theoretical_loss": 3.364718939850656,
      "tokens_seen": 2824863744
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001454764292878636,
      "loss": 1.3071,
      "theoretical_loss": 3.3647129159276745,
      "tokens_seen": 2824929280
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014545636910732196,
      "loss": 1.2456,
      "theoretical_loss": 3.36470689218357,
      "tokens_seen": 2824994816
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014543630892678032,
      "loss": 1.3039,
      "theoretical_loss": 3.3647008686183337,
      "tokens_seen": 2825060352
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001454162487462387,
      "loss": 1.2025,
      "theoretical_loss": 3.3646948452319556,
      "tokens_seen": 2825125888
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014539618856569708,
      "loss": 1.2729,
      "theoretical_loss": 3.364688822024427,
      "tokens_seen": 2825191424
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014537612838515546,
      "loss": 1.252,
      "theoretical_loss": 3.3646827989957373,
      "tokens_seen": 2825256960
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014535606820461383,
      "loss": 1.1788,
      "theoretical_loss": 3.3646767761458776,
      "tokens_seen": 2825322496
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014533600802407222,
      "loss": 1.2315,
      "theoretical_loss": 3.3646707534748383,
      "tokens_seen": 2825388032
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014531594784353058,
      "loss": 1.4156,
      "theoretical_loss": 3.3646647309826108,
      "tokens_seen": 2825453568
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014529588766298897,
      "loss": 1.2846,
      "theoretical_loss": 3.364658708669184,
      "tokens_seen": 2825519104
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014527582748244733,
      "loss": 1.2878,
      "theoretical_loss": 3.3646526865345505,
      "tokens_seen": 2825584640
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014525576730190575,
      "loss": 1.3465,
      "theoretical_loss": 3.3646466645786988,
      "tokens_seen": 2825650176
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001452357071213641,
      "loss": 1.3168,
      "theoretical_loss": 3.364640642801621,
      "tokens_seen": 2825715712
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014521564694082247,
      "loss": 1.2529,
      "theoretical_loss": 3.3646346212033067,
      "tokens_seen": 2825781248
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014519558676028086,
      "loss": 1.2339,
      "theoretical_loss": 3.364628599783747,
      "tokens_seen": 2825846784
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014517552657973922,
      "loss": 1.2838,
      "theoretical_loss": 3.364622578542932,
      "tokens_seen": 2825912320
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001451554663991976,
      "loss": 1.3679,
      "theoretical_loss": 3.3646165574808524,
      "tokens_seen": 2825977856
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014513540621865597,
      "loss": 1.201,
      "theoretical_loss": 3.364610536597499,
      "tokens_seen": 2826043392
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014511534603811436,
      "loss": 1.3193,
      "theoretical_loss": 3.3646045158928626,
      "tokens_seen": 2826108928
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014509528585757273,
      "loss": 1.1942,
      "theoretical_loss": 3.364598495366933,
      "tokens_seen": 2826174464
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014507522567703112,
      "loss": 1.2649,
      "theoretical_loss": 3.364592475019701,
      "tokens_seen": 2826240000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014505516549648948,
      "loss": 1.2474,
      "theoretical_loss": 3.364586454851157,
      "tokens_seen": 2826305536
    },
    {
      "epoch": 0.71,
      "objective/train/docs_used": 1027020,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1658639907836914,
      "objective/train/theoretical_loss": 3.364580434861292,
      "objective/train/tokens_used": 1196830176,
      "theoretical_loss": 3.364580434861292,
      "tokens_seen": 2826371072
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014503510531594784,
      "loss": 1.2868,
      "theoretical_loss": 3.364580434861292,
      "tokens_seen": 2826371072
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014501504513540623,
      "loss": 1.1793,
      "theoretical_loss": 3.3645744150500967,
      "tokens_seen": 2826436608
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001449949849548646,
      "loss": 1.3026,
      "theoretical_loss": 3.3645683954175607,
      "tokens_seen": 2826502144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014497492477432298,
      "loss": 1.2663,
      "theoretical_loss": 3.3645623759636756,
      "tokens_seen": 2826567680
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014495486459378134,
      "loss": 1.3111,
      "theoretical_loss": 3.364556356688431,
      "tokens_seen": 2826633216
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014493480441323973,
      "loss": 1.2384,
      "theoretical_loss": 3.3645503375918184,
      "tokens_seen": 2826698752
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001449147442326981,
      "loss": 1.4291,
      "theoretical_loss": 3.3645443186738277,
      "tokens_seen": 2826764288
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014489468405215648,
      "loss": 1.2627,
      "theoretical_loss": 3.36453829993445,
      "tokens_seen": 2826829824
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014487462387161485,
      "loss": 1.2481,
      "theoretical_loss": 3.364532281373675,
      "tokens_seen": 2826895360
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001448545636910732,
      "loss": 1.2689,
      "theoretical_loss": 3.3645262629914945,
      "tokens_seen": 2826960896
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001448345035105316,
      "loss": 1.208,
      "theoretical_loss": 3.364520244787898,
      "tokens_seen": 2827026432
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014481444332998996,
      "loss": 1.2726,
      "theoretical_loss": 3.3645142267628763,
      "tokens_seen": 2827091968
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014479438314944835,
      "loss": 1.2645,
      "theoretical_loss": 3.36450820891642,
      "tokens_seen": 2827157504
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001447743229689067,
      "loss": 1.234,
      "theoretical_loss": 3.36450219124852,
      "tokens_seen": 2827223040
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001447542627883651,
      "loss": 1.3298,
      "theoretical_loss": 3.364496173759167,
      "tokens_seen": 2827288576
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014473420260782346,
      "loss": 1.2756,
      "theoretical_loss": 3.3644901564483503,
      "tokens_seen": 2827354112
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014471414242728185,
      "loss": 1.2192,
      "theoretical_loss": 3.364484139316062,
      "tokens_seen": 2827419648
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014469408224674021,
      "loss": 1.2172,
      "theoretical_loss": 3.3644781223622915,
      "tokens_seen": 2827485184
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001446740220661986,
      "loss": 1.3347,
      "theoretical_loss": 3.3644721055870304,
      "tokens_seen": 2827550720
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014465396188565697,
      "loss": 1.3454,
      "theoretical_loss": 3.3644660889902682,
      "tokens_seen": 2827616256
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014463390170511533,
      "loss": 1.3022,
      "theoretical_loss": 3.364460072571996,
      "tokens_seen": 2827681792
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014461384152457372,
      "loss": 1.2448,
      "theoretical_loss": 3.364454056332205,
      "tokens_seen": 2827747328
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014459378134403208,
      "loss": 1.2399,
      "theoretical_loss": 3.3644480402708847,
      "tokens_seen": 2827812864
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014457372116349047,
      "loss": 1.3387,
      "theoretical_loss": 3.364442024388026,
      "tokens_seen": 2827878400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014455366098294883,
      "loss": 1.3847,
      "theoretical_loss": 3.36443600868362,
      "tokens_seen": 2827943936
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014453360080240722,
      "loss": 1.309,
      "theoretical_loss": 3.3644299931576565,
      "tokens_seen": 2828009472
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001445135406218656,
      "loss": 1.2764,
      "theoretical_loss": 3.3644239778101266,
      "tokens_seen": 2828075008
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000144493480441324,
      "loss": 1.3503,
      "theoretical_loss": 3.3644179626410207,
      "tokens_seen": 2828140544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014447342026078236,
      "loss": 1.2589,
      "theoretical_loss": 3.3644119476503294,
      "tokens_seen": 2828206080
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014445336008024072,
      "loss": 1.249,
      "theoretical_loss": 3.364405932838043,
      "tokens_seen": 2828271616
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014443329989969911,
      "loss": 1.4068,
      "theoretical_loss": 3.3643999182041524,
      "tokens_seen": 2828337152
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014441323971915748,
      "loss": 1.2353,
      "theoretical_loss": 3.364393903748648,
      "tokens_seen": 2828402688
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014439317953861587,
      "loss": 1.2533,
      "theoretical_loss": 3.3643878894715207,
      "tokens_seen": 2828468224
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014437311935807423,
      "loss": 1.2319,
      "theoretical_loss": 3.364381875372761,
      "tokens_seen": 2828533760
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014435305917753262,
      "loss": 1.2365,
      "theoretical_loss": 3.364375861452359,
      "tokens_seen": 2828599296
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014433299899699098,
      "loss": 1.2301,
      "theoretical_loss": 3.3643698477103055,
      "tokens_seen": 2828664832
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014431293881644937,
      "loss": 1.2235,
      "theoretical_loss": 3.3643638341465913,
      "tokens_seen": 2828730368
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014429287863590773,
      "loss": 1.2787,
      "theoretical_loss": 3.364357820761207,
      "tokens_seen": 2828795904
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001442728184553661,
      "loss": 1.2708,
      "theoretical_loss": 3.3643518075541428,
      "tokens_seen": 2828861440
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014425275827482448,
      "loss": 1.3504,
      "theoretical_loss": 3.3643457945253896,
      "tokens_seen": 2828926976
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014423269809428284,
      "loss": 1.2737,
      "theoretical_loss": 3.3643397816749383,
      "tokens_seen": 2828992512
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014421263791374123,
      "loss": 1.3276,
      "theoretical_loss": 3.3643337690027786,
      "tokens_seen": 2829058048
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001441925777331996,
      "loss": 1.3435,
      "theoretical_loss": 3.364327756508902,
      "tokens_seen": 2829123584
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014417251755265799,
      "loss": 1.2439,
      "theoretical_loss": 3.3643217441932984,
      "tokens_seen": 2829189120
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014415245737211635,
      "loss": 1.4018,
      "theoretical_loss": 3.3643157320559585,
      "tokens_seen": 2829254656
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014413239719157474,
      "loss": 1.2549,
      "theoretical_loss": 3.3643097200968732,
      "tokens_seen": 2829320192
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001441123370110331,
      "loss": 1.2949,
      "theoretical_loss": 3.364303708316033,
      "tokens_seen": 2829385728
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001440922768304915,
      "loss": 1.2714,
      "theoretical_loss": 3.364297696713428,
      "tokens_seen": 2829451264
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014407221664994985,
      "loss": 1.2752,
      "theoretical_loss": 3.3642916852890496,
      "tokens_seen": 2829516800
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001440521564694082,
      "loss": 1.3084,
      "theoretical_loss": 3.364285674042888,
      "tokens_seen": 2829582336
    },
    {
      "epoch": 0.71,
      "objective/train/docs_used": 1028281,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2568520307540894,
      "objective/train/theoretical_loss": 3.3642796629749334,
      "objective/train/tokens_used": 1200106976,
      "theoretical_loss": 3.3642796629749334,
      "tokens_seen": 2829647872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001440320962888666,
      "loss": 1.2582,
      "theoretical_loss": 3.3642796629749334,
      "tokens_seen": 2829647872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014401203610832496,
      "loss": 1.2644,
      "theoretical_loss": 3.364273652085177,
      "tokens_seen": 2829713408
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014399197592778335,
      "loss": 1.1729,
      "theoretical_loss": 3.364267641373609,
      "tokens_seen": 2829778944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014397191574724172,
      "loss": 1.3068,
      "theoretical_loss": 3.36426163084022,
      "tokens_seen": 2829844480
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001439518555667001,
      "loss": 1.3008,
      "theoretical_loss": 3.364255620485001,
      "tokens_seen": 2829910016
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014393179538615847,
      "loss": 1.3146,
      "theoretical_loss": 3.364249610307942,
      "tokens_seen": 2829975552
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014391173520561686,
      "loss": 1.3191,
      "theoretical_loss": 3.3642436003090346,
      "tokens_seen": 2830041088
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014389167502507522,
      "loss": 1.3142,
      "theoretical_loss": 3.364237590488268,
      "tokens_seen": 2830106624
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014387161484453358,
      "loss": 1.2097,
      "theoretical_loss": 3.364231580845634,
      "tokens_seen": 2830172160
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014385155466399197,
      "loss": 1.3364,
      "theoretical_loss": 3.364225571381122,
      "tokens_seen": 2830237696
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014383149448345033,
      "loss": 1.281,
      "theoretical_loss": 3.364219562094724,
      "tokens_seen": 2830303232
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014381143430290872,
      "loss": 1.3632,
      "theoretical_loss": 3.3642135529864294,
      "tokens_seen": 2830368768
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014379137412236708,
      "loss": 1.3048,
      "theoretical_loss": 3.36420754405623,
      "tokens_seen": 2830434304
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014377131394182547,
      "loss": 1.2389,
      "theoretical_loss": 3.364201535304115,
      "tokens_seen": 2830499840
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014375125376128386,
      "loss": 1.2113,
      "theoretical_loss": 3.364195526730076,
      "tokens_seen": 2830565376
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014373119358074225,
      "loss": 1.2114,
      "theoretical_loss": 3.364189518334103,
      "tokens_seen": 2830630912
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014371113340020062,
      "loss": 1.2416,
      "theoretical_loss": 3.364183510116187,
      "tokens_seen": 2830696448
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000143691073219659,
      "loss": 1.2644,
      "theoretical_loss": 3.3641775020763185,
      "tokens_seen": 2830761984
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014367101303911737,
      "loss": 1.3088,
      "theoretical_loss": 3.3641714942144882,
      "tokens_seen": 2830827520
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014365095285857573,
      "loss": 1.214,
      "theoretical_loss": 3.3641654865306867,
      "tokens_seen": 2830893056
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014363089267803412,
      "loss": 1.2747,
      "theoretical_loss": 3.364159479024904,
      "tokens_seen": 2830958592
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014361083249749248,
      "loss": 1.2038,
      "theoretical_loss": 3.3641534716971315,
      "tokens_seen": 2831024128
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014359077231695087,
      "loss": 1.2341,
      "theoretical_loss": 3.3641474645473597,
      "tokens_seen": 2831089664
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014357071213640923,
      "loss": 1.2822,
      "theoretical_loss": 3.364141457575579,
      "tokens_seen": 2831155200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014355065195586762,
      "loss": 1.4534,
      "theoretical_loss": 3.3641354507817796,
      "tokens_seen": 2831220736
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014353059177532598,
      "loss": 1.3247,
      "theoretical_loss": 3.3641294441659526,
      "tokens_seen": 2831286272
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014351053159478437,
      "loss": 1.3032,
      "theoretical_loss": 3.3641234377280886,
      "tokens_seen": 2831351808
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014349047141424274,
      "loss": 1.3182,
      "theoretical_loss": 3.3641174314681783,
      "tokens_seen": 2831417344
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001434704112337011,
      "loss": 1.3308,
      "theoretical_loss": 3.364111425386212,
      "tokens_seen": 2831482880
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001434503510531595,
      "loss": 1.3181,
      "theoretical_loss": 3.3641054194821804,
      "tokens_seen": 2831548416
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014343029087261785,
      "loss": 1.2895,
      "theoretical_loss": 3.3640994137560742,
      "tokens_seen": 2831613952
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014341023069207624,
      "loss": 1.3253,
      "theoretical_loss": 3.364093408207884,
      "tokens_seen": 2831679488
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001433901705115346,
      "loss": 1.3764,
      "theoretical_loss": 3.3640874028376007,
      "tokens_seen": 2831745024
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000143370110330993,
      "loss": 1.3683,
      "theoretical_loss": 3.364081397645214,
      "tokens_seen": 2831810560
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014335005015045135,
      "loss": 1.2616,
      "theoretical_loss": 3.3640753926307156,
      "tokens_seen": 2831876096
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014332998996990974,
      "loss": 1.2399,
      "theoretical_loss": 3.3640693877940957,
      "tokens_seen": 2831941632
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001433099297893681,
      "loss": 1.2846,
      "theoretical_loss": 3.3640633831353446,
      "tokens_seen": 2832007168
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014328986960882647,
      "loss": 1.3498,
      "theoretical_loss": 3.364057378654453,
      "tokens_seen": 2832072704
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014326980942828486,
      "loss": 1.3158,
      "theoretical_loss": 3.364051374351412,
      "tokens_seen": 2832138240
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014324974924774322,
      "loss": 1.4228,
      "theoretical_loss": 3.3640453702262114,
      "tokens_seen": 2832203776
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001432296890672016,
      "loss": 1.2375,
      "theoretical_loss": 3.364039366278843,
      "tokens_seen": 2832269312
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014320962888665997,
      "loss": 1.3199,
      "theoretical_loss": 3.3640333625092964,
      "tokens_seen": 2832334848
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014318956870611836,
      "loss": 1.2364,
      "theoretical_loss": 3.3640273589175624,
      "tokens_seen": 2832400384
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014316950852557672,
      "loss": 1.3034,
      "theoretical_loss": 3.364021355503632,
      "tokens_seen": 2832465920
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001431494483450351,
      "loss": 1.2495,
      "theoretical_loss": 3.3640153522674954,
      "tokens_seen": 2832531456
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014312938816449347,
      "loss": 1.255,
      "theoretical_loss": 3.3640093492091436,
      "tokens_seen": 2832596992
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014310932798395183,
      "loss": 1.3292,
      "theoretical_loss": 3.3640033463285666,
      "tokens_seen": 2832662528
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014308926780341022,
      "loss": 1.2612,
      "theoretical_loss": 3.3639973436257558,
      "tokens_seen": 2832728064
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014306920762286859,
      "loss": 1.2908,
      "theoretical_loss": 3.3639913411007014,
      "tokens_seen": 2832793600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014304914744232698,
      "loss": 1.2409,
      "theoretical_loss": 3.3639853387533942,
      "tokens_seen": 2832859136
    },
    {
      "epoch": 0.72,
      "objective/train/docs_used": 1029660,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4082143306732178,
      "objective/train/theoretical_loss": 3.3639793365838244,
      "objective/train/tokens_used": 1203383776,
      "theoretical_loss": 3.3639793365838244,
      "tokens_seen": 2832924672
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014302908726178534,
      "loss": 1.3257,
      "theoretical_loss": 3.3639793365838244,
      "tokens_seen": 2832924672
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014300902708124375,
      "loss": 1.2782,
      "theoretical_loss": 3.363973334591983,
      "tokens_seen": 2832990208
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014298896690070212,
      "loss": 1.3576,
      "theoretical_loss": 3.363967332777861,
      "tokens_seen": 2833055744
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001429689067201605,
      "loss": 1.3947,
      "theoretical_loss": 3.3639613311414482,
      "tokens_seen": 2833121280
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014294884653961887,
      "loss": 1.2626,
      "theoretical_loss": 3.363955329682736,
      "tokens_seen": 2833186816
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014292878635907726,
      "loss": 1.2876,
      "theoretical_loss": 3.3639493284017146,
      "tokens_seen": 2833252352
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014290872617853562,
      "loss": 1.2542,
      "theoretical_loss": 3.363943327298374,
      "tokens_seen": 2833317888
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014288866599799398,
      "loss": 1.3092,
      "theoretical_loss": 3.3639373263727066,
      "tokens_seen": 2833383424
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014286860581745237,
      "loss": 1.2629,
      "theoretical_loss": 3.363931325624701,
      "tokens_seen": 2833448960
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014284854563691073,
      "loss": 1.228,
      "theoretical_loss": 3.3639253250543493,
      "tokens_seen": 2833514496
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014282848545636912,
      "loss": 1.3005,
      "theoretical_loss": 3.3639193246616417,
      "tokens_seen": 2833580032
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014280842527582749,
      "loss": 1.2397,
      "theoretical_loss": 3.3639133244465684,
      "tokens_seen": 2833645568
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014278836509528587,
      "loss": 1.2603,
      "theoretical_loss": 3.36390732440912,
      "tokens_seen": 2833711104
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014276830491474424,
      "loss": 1.3287,
      "theoretical_loss": 3.3639013245492886,
      "tokens_seen": 2833776640
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014274824473420263,
      "loss": 1.3638,
      "theoretical_loss": 3.363895324867063,
      "tokens_seen": 2833842176
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000142728184553661,
      "loss": 1.3131,
      "theoretical_loss": 3.3638893253624347,
      "tokens_seen": 2833907712
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014270812437311935,
      "loss": 1.2498,
      "theoretical_loss": 3.363883326035394,
      "tokens_seen": 2833973248
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014268806419257774,
      "loss": 1.2757,
      "theoretical_loss": 3.363877326885932,
      "tokens_seen": 2834038784
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001426680040120361,
      "loss": 1.3788,
      "theoretical_loss": 3.3638713279140395,
      "tokens_seen": 2834104320
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001426479438314945,
      "loss": 1.2572,
      "theoretical_loss": 3.3638653291197063,
      "tokens_seen": 2834169856
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014262788365095285,
      "loss": 1.3608,
      "theoretical_loss": 3.3638593305029234,
      "tokens_seen": 2834235392
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014260782347041124,
      "loss": 1.3693,
      "theoretical_loss": 3.363853332063681,
      "tokens_seen": 2834300928
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001425877632898696,
      "loss": 1.3047,
      "theoretical_loss": 3.363847333801971,
      "tokens_seen": 2834366464
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000142567703109328,
      "loss": 1.3688,
      "theoretical_loss": 3.363841335717783,
      "tokens_seen": 2834432000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014254764292878636,
      "loss": 1.3609,
      "theoretical_loss": 3.3638353378111083,
      "tokens_seen": 2834497536
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014252758274824475,
      "loss": 1.3068,
      "theoretical_loss": 3.363829340081937,
      "tokens_seen": 2834563072
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001425075225677031,
      "loss": 1.3109,
      "theoretical_loss": 3.3638233425302597,
      "tokens_seen": 2834628608
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014248746238716147,
      "loss": 1.2613,
      "theoretical_loss": 3.3638173451560673,
      "tokens_seen": 2834694144
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014246740220661986,
      "loss": 1.3885,
      "theoretical_loss": 3.3638113479593508,
      "tokens_seen": 2834759680
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014244734202607822,
      "loss": 1.3265,
      "theoretical_loss": 3.3638053509401002,
      "tokens_seen": 2834825216
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001424272818455366,
      "loss": 1.3495,
      "theoretical_loss": 3.363799354098306,
      "tokens_seen": 2834890752
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014240722166499497,
      "loss": 1.321,
      "theoretical_loss": 3.3637933574339596,
      "tokens_seen": 2834956288
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014238716148445336,
      "loss": 1.2629,
      "theoretical_loss": 3.363787360947051,
      "tokens_seen": 2835021824
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014236710130391173,
      "loss": 1.3567,
      "theoretical_loss": 3.3637813646375716,
      "tokens_seen": 2835087360
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014234704112337011,
      "loss": 1.3325,
      "theoretical_loss": 3.3637753685055114,
      "tokens_seen": 2835152896
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014232698094282848,
      "loss": 1.2738,
      "theoretical_loss": 3.3637693725508613,
      "tokens_seen": 2835218432
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014230692076228684,
      "loss": 1.3108,
      "theoretical_loss": 3.363763376773612,
      "tokens_seen": 2835283968
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014228686058174523,
      "loss": 1.3203,
      "theoretical_loss": 3.3637573811737536,
      "tokens_seen": 2835349504
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014226680040120362,
      "loss": 1.3103,
      "theoretical_loss": 3.3637513857512773,
      "tokens_seen": 2835415040
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000142246740220662,
      "loss": 1.3049,
      "theoretical_loss": 3.363745390506174,
      "tokens_seen": 2835480576
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014222668004012037,
      "loss": 1.2377,
      "theoretical_loss": 3.363739395438434,
      "tokens_seen": 2835546112
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014220661985957876,
      "loss": 1.286,
      "theoretical_loss": 3.363733400548047,
      "tokens_seen": 2835611648
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014218655967903712,
      "loss": 1.4105,
      "theoretical_loss": 3.3637274058350055,
      "tokens_seen": 2835677184
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001421664994984955,
      "loss": 1.1648,
      "theoretical_loss": 3.363721411299299,
      "tokens_seen": 2835742720
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014214643931795387,
      "loss": 1.2669,
      "theoretical_loss": 3.3637154169409187,
      "tokens_seen": 2835808256
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014212637913741223,
      "loss": 1.257,
      "theoretical_loss": 3.3637094227598547,
      "tokens_seen": 2835873792
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014210631895687062,
      "loss": 1.3676,
      "theoretical_loss": 3.363703428756098,
      "tokens_seen": 2835939328
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000142086258776329,
      "loss": 1.2888,
      "theoretical_loss": 3.363697434929639,
      "tokens_seen": 2836004864
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014206619859578738,
      "loss": 1.2788,
      "theoretical_loss": 3.3636914412804684,
      "tokens_seen": 2836070400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014204613841524574,
      "loss": 1.2499,
      "theoretical_loss": 3.3636854478085776,
      "tokens_seen": 2836135936
    },
    {
      "epoch": 0.72,
      "objective/train/docs_used": 1030811,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4940600395202637,
      "objective/train/theoretical_loss": 3.3636794545139566,
      "objective/train/tokens_used": 1206660576,
      "theoretical_loss": 3.3636794545139566,
      "tokens_seen": 2836201472
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014202607823470413,
      "loss": 1.4282,
      "theoretical_loss": 3.3636794545139566,
      "tokens_seen": 2836201472
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001420060180541625,
      "loss": 1.2166,
      "theoretical_loss": 3.3636734613965955,
      "tokens_seen": 2836267008
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014198595787362088,
      "loss": 1.2767,
      "theoretical_loss": 3.363667468456486,
      "tokens_seen": 2836332544
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014196589769307924,
      "loss": 1.3173,
      "theoretical_loss": 3.3636614756936183,
      "tokens_seen": 2836398080
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014194583751253763,
      "loss": 1.2247,
      "theoretical_loss": 3.363655483107983,
      "tokens_seen": 2836463616
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000141925777331996,
      "loss": 1.2578,
      "theoretical_loss": 3.3636494906995713,
      "tokens_seen": 2836529152
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014190571715145436,
      "loss": 1.2593,
      "theoretical_loss": 3.363643498468373,
      "tokens_seen": 2836594688
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014188565697091274,
      "loss": 1.3135,
      "theoretical_loss": 3.363637506414379,
      "tokens_seen": 2836660224
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001418655967903711,
      "loss": 1.3631,
      "theoretical_loss": 3.363631514537581,
      "tokens_seen": 2836725760
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001418455366098295,
      "loss": 1.4396,
      "theoretical_loss": 3.363625522837968,
      "tokens_seen": 2836791296
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014182547642928786,
      "loss": 1.3117,
      "theoretical_loss": 3.363619531315532,
      "tokens_seen": 2836856832
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014180541624874625,
      "loss": 1.3365,
      "theoretical_loss": 3.363613539970263,
      "tokens_seen": 2836922368
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001417853560682046,
      "loss": 1.2605,
      "theoretical_loss": 3.3636075488021517,
      "tokens_seen": 2836987904
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000141765295887663,
      "loss": 1.3549,
      "theoretical_loss": 3.3636015578111893,
      "tokens_seen": 2837053440
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014174523570712136,
      "loss": 1.3072,
      "theoretical_loss": 3.3635955669973656,
      "tokens_seen": 2837118976
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014172517552657972,
      "loss": 1.1947,
      "theoretical_loss": 3.3635895763606722,
      "tokens_seen": 2837184512
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001417051153460381,
      "loss": 1.3025,
      "theoretical_loss": 3.363583585901099,
      "tokens_seen": 2837250048
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014168505516549648,
      "loss": 1.3814,
      "theoretical_loss": 3.363577595618637,
      "tokens_seen": 2837315584
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014166499498495486,
      "loss": 1.2895,
      "theoretical_loss": 3.363571605513277,
      "tokens_seen": 2837381120
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014164493480441323,
      "loss": 1.2411,
      "theoretical_loss": 3.3635656155850095,
      "tokens_seen": 2837446656
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014162487462387162,
      "loss": 1.2486,
      "theoretical_loss": 3.363559625833825,
      "tokens_seen": 2837512192
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014160481444332998,
      "loss": 1.2915,
      "theoretical_loss": 3.3635536362597147,
      "tokens_seen": 2837577728
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014158475426278837,
      "loss": 1.3183,
      "theoretical_loss": 3.363547646862669,
      "tokens_seen": 2837643264
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014156469408224673,
      "loss": 1.2194,
      "theoretical_loss": 3.3635416576426787,
      "tokens_seen": 2837708800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001415446339017051,
      "loss": 1.2664,
      "theoretical_loss": 3.3635356685997335,
      "tokens_seen": 2837774336
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014152457372116348,
      "loss": 1.3235,
      "theoretical_loss": 3.3635296797338254,
      "tokens_seen": 2837839872
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014150451354062187,
      "loss": 1.4048,
      "theoretical_loss": 3.3635236910449446,
      "tokens_seen": 2837905408
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014148445336008026,
      "loss": 1.2698,
      "theoretical_loss": 3.3635177025330822,
      "tokens_seen": 2837970944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014146439317953862,
      "loss": 1.3681,
      "theoretical_loss": 3.3635117141982276,
      "tokens_seen": 2838036480
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000141444332998997,
      "loss": 1.292,
      "theoretical_loss": 3.3635057260403727,
      "tokens_seen": 2838102016
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014142427281845537,
      "loss": 1.3044,
      "theoretical_loss": 3.363499738059508,
      "tokens_seen": 2838167552
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014140421263791376,
      "loss": 1.1697,
      "theoretical_loss": 3.363493750255624,
      "tokens_seen": 2838233088
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014138415245737213,
      "loss": 1.2295,
      "theoretical_loss": 3.363487762628711,
      "tokens_seen": 2838298624
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014136409227683052,
      "loss": 1.2287,
      "theoretical_loss": 3.36348177517876,
      "tokens_seen": 2838364160
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014134403209628888,
      "loss": 1.292,
      "theoretical_loss": 3.363475787905762,
      "tokens_seen": 2838429696
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014132397191574724,
      "loss": 1.1975,
      "theoretical_loss": 3.3634698008097077,
      "tokens_seen": 2838495232
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014130391173520563,
      "loss": 1.3534,
      "theoretical_loss": 3.363463813890587,
      "tokens_seen": 2838560768
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000141283851554664,
      "loss": 1.3025,
      "theoretical_loss": 3.363457827148391,
      "tokens_seen": 2838626304
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014126379137412238,
      "loss": 1.1783,
      "theoretical_loss": 3.3634518405831106,
      "tokens_seen": 2838691840
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014124373119358074,
      "loss": 1.1886,
      "theoretical_loss": 3.3634458541947367,
      "tokens_seen": 2838757376
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014122367101303913,
      "loss": 1.2712,
      "theoretical_loss": 3.363439867983259,
      "tokens_seen": 2838822912
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001412036108324975,
      "loss": 1.2046,
      "theoretical_loss": 3.3634338819486693,
      "tokens_seen": 2838888448
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014118355065195588,
      "loss": 1.2057,
      "theoretical_loss": 3.363427896090958,
      "tokens_seen": 2838953984
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014116349047141425,
      "loss": 1.3167,
      "theoretical_loss": 3.3634219104101155,
      "tokens_seen": 2839019520
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001411434302908726,
      "loss": 1.2536,
      "theoretical_loss": 3.3634159249061324,
      "tokens_seen": 2839085056
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000141123370110331,
      "loss": 1.3797,
      "theoretical_loss": 3.363409939579,
      "tokens_seen": 2839150592
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014110330992978936,
      "loss": 1.2651,
      "theoretical_loss": 3.363403954428708,
      "tokens_seen": 2839216128
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014108324974924775,
      "loss": 1.2573,
      "theoretical_loss": 3.363397969455248,
      "tokens_seen": 2839281664
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001410631895687061,
      "loss": 1.2693,
      "theoretical_loss": 3.3633919846586102,
      "tokens_seen": 2839347200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001410431293881645,
      "loss": 1.3165,
      "theoretical_loss": 3.3633860000387856,
      "tokens_seen": 2839412736
    },
    {
      "epoch": 0.72,
      "objective/train/docs_used": 1031989,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3026928901672363,
      "objective/train/theoretical_loss": 3.363380015595765,
      "objective/train/tokens_used": 1209937376,
      "theoretical_loss": 3.363380015595765,
      "tokens_seen": 2839478272
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014102306920762286,
      "loss": 1.3013,
      "theoretical_loss": 3.363380015595765,
      "tokens_seen": 2839478272
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014100300902708125,
      "loss": 1.2961,
      "theoretical_loss": 3.3633740313295384,
      "tokens_seen": 2839543808
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014098294884653961,
      "loss": 1.3192,
      "theoretical_loss": 3.3633680472400975,
      "tokens_seen": 2839609344
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000140962888665998,
      "loss": 1.2157,
      "theoretical_loss": 3.363362063327432,
      "tokens_seen": 2839674880
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014094282848545637,
      "loss": 1.2511,
      "theoretical_loss": 3.3633560795915334,
      "tokens_seen": 2839740416
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014092276830491473,
      "loss": 1.3161,
      "theoretical_loss": 3.363350096032392,
      "tokens_seen": 2839805952
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014090270812437312,
      "loss": 1.2354,
      "theoretical_loss": 3.3633441126499983,
      "tokens_seen": 2839871488
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014088264794383148,
      "loss": 1.1827,
      "theoretical_loss": 3.3633381294443434,
      "tokens_seen": 2839937024
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014086258776328987,
      "loss": 1.2672,
      "theoretical_loss": 3.3633321464154182,
      "tokens_seen": 2840002560
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014084252758274823,
      "loss": 1.2759,
      "theoretical_loss": 3.3633261635632126,
      "tokens_seen": 2840068096
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014082246740220662,
      "loss": 1.4819,
      "theoretical_loss": 3.363320180887718,
      "tokens_seen": 2840133632
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014080240722166498,
      "loss": 1.2415,
      "theoretical_loss": 3.3633141983889248,
      "tokens_seen": 2840199168
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014078234704112337,
      "loss": 1.3209,
      "theoretical_loss": 3.363308216066824,
      "tokens_seen": 2840264704
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014076228686058176,
      "loss": 1.3605,
      "theoretical_loss": 3.363302233921406,
      "tokens_seen": 2840330240
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014074222668004012,
      "loss": 1.3092,
      "theoretical_loss": 3.363296251952661,
      "tokens_seen": 2840395776
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001407221664994985,
      "loss": 1.2424,
      "theoretical_loss": 3.3632902701605807,
      "tokens_seen": 2840461312
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014070210631895688,
      "loss": 1.2125,
      "theoretical_loss": 3.3632842885451555,
      "tokens_seen": 2840526848
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014068204613841526,
      "loss": 1.3118,
      "theoretical_loss": 3.363278307106376,
      "tokens_seen": 2840592384
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014066198595787363,
      "loss": 1.2292,
      "theoretical_loss": 3.3632723258442327,
      "tokens_seen": 2840657920
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014064192577733202,
      "loss": 1.3743,
      "theoretical_loss": 3.3632663447587166,
      "tokens_seen": 2840723456
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014062186559679038,
      "loss": 1.2437,
      "theoretical_loss": 3.363260363849818,
      "tokens_seen": 2840788992
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014060180541624877,
      "loss": 1.2272,
      "theoretical_loss": 3.363254383117529,
      "tokens_seen": 2840854528
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014058174523570713,
      "loss": 1.2499,
      "theoretical_loss": 3.3632484025618385,
      "tokens_seen": 2840920064
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001405616850551655,
      "loss": 1.2195,
      "theoretical_loss": 3.3632424221827377,
      "tokens_seen": 2840985600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014054162487462388,
      "loss": 1.2881,
      "theoretical_loss": 3.363236441980218,
      "tokens_seen": 2841051136
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014052156469408224,
      "loss": 1.2667,
      "theoretical_loss": 3.3632304619542697,
      "tokens_seen": 2841116672
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014050150451354063,
      "loss": 1.3998,
      "theoretical_loss": 3.3632244821048833,
      "tokens_seen": 2841182208
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000140481444332999,
      "loss": 1.3785,
      "theoretical_loss": 3.36321850243205,
      "tokens_seen": 2841247744
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014046138415245739,
      "loss": 1.2861,
      "theoretical_loss": 3.36321252293576,
      "tokens_seen": 2841313280
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014044132397191575,
      "loss": 1.3191,
      "theoretical_loss": 3.363206543616004,
      "tokens_seen": 2841378816
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014042126379137414,
      "loss": 1.3239,
      "theoretical_loss": 3.3632005644727734,
      "tokens_seen": 2841444352
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001404012036108325,
      "loss": 1.3063,
      "theoretical_loss": 3.3631945855060583,
      "tokens_seen": 2841509888
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001403811434302909,
      "loss": 1.3731,
      "theoretical_loss": 3.3631886067158496,
      "tokens_seen": 2841575424
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014036108324974925,
      "loss": 1.2601,
      "theoretical_loss": 3.363182628102138,
      "tokens_seen": 2841640960
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001403410230692076,
      "loss": 1.1639,
      "theoretical_loss": 3.3631766496649145,
      "tokens_seen": 2841706496
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000140320962888666,
      "loss": 1.3155,
      "theoretical_loss": 3.3631706714041694,
      "tokens_seen": 2841772032
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014030090270812436,
      "loss": 1.298,
      "theoretical_loss": 3.3631646933198933,
      "tokens_seen": 2841837568
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014028084252758275,
      "loss": 1.2927,
      "theoretical_loss": 3.3631587154120774,
      "tokens_seen": 2841903104
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014026078234704112,
      "loss": 1.241,
      "theoretical_loss": 3.3631527376807124,
      "tokens_seen": 2841968640
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001402407221664995,
      "loss": 1.2656,
      "theoretical_loss": 3.3631467601257885,
      "tokens_seen": 2842034176
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014022066198595787,
      "loss": 1.2658,
      "theoretical_loss": 3.3631407827472968,
      "tokens_seen": 2842099712
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014020060180541626,
      "loss": 1.3889,
      "theoretical_loss": 3.3631348055452284,
      "tokens_seen": 2842165248
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014018054162487462,
      "loss": 1.2931,
      "theoretical_loss": 3.3631288285195735,
      "tokens_seen": 2842230784
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014016048144433298,
      "loss": 1.3352,
      "theoretical_loss": 3.3631228516703224,
      "tokens_seen": 2842296320
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014014042126379137,
      "loss": 1.2629,
      "theoretical_loss": 3.363116874997467,
      "tokens_seen": 2842361856
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014012036108324973,
      "loss": 1.3733,
      "theoretical_loss": 3.363110898500997,
      "tokens_seen": 2842427392
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014010030090270812,
      "loss": 1.3488,
      "theoretical_loss": 3.363104922180904,
      "tokens_seen": 2842492928
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014008024072216648,
      "loss": 1.2476,
      "theoretical_loss": 3.363098946037178,
      "tokens_seen": 2842558464
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014006018054162487,
      "loss": 1.3065,
      "theoretical_loss": 3.36309297006981,
      "tokens_seen": 2842624000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014004012036108324,
      "loss": 1.254,
      "theoretical_loss": 3.3630869942787904,
      "tokens_seen": 2842689536
    },
    {
      "epoch": 0.72,
      "objective/train/docs_used": 1033097,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.215433120727539,
      "objective/train/theoretical_loss": 3.3630810186641105,
      "objective/train/tokens_used": 1213214176,
      "theoretical_loss": 3.3630810186641105,
      "tokens_seen": 2842755072
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014002006018054163,
      "loss": 1.27,
      "theoretical_loss": 3.3630810186641105,
      "tokens_seen": 2842755072
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014000000000000001,
      "loss": 1.2988,
      "theoretical_loss": 3.363075043225761,
      "tokens_seen": 2842820608
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013997993981945838,
      "loss": 1.2765,
      "theoretical_loss": 3.363069067963732,
      "tokens_seen": 2842886144
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013995987963891677,
      "loss": 1.2565,
      "theoretical_loss": 3.363063092878015,
      "tokens_seen": 2842951680
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013993981945837513,
      "loss": 1.3236,
      "theoretical_loss": 3.3630571179685997,
      "tokens_seen": 2843017216
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013991975927783352,
      "loss": 1.3084,
      "theoretical_loss": 3.363051143235478,
      "tokens_seen": 2843082752
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013989969909729188,
      "loss": 1.1973,
      "theoretical_loss": 3.3630451686786405,
      "tokens_seen": 2843148288
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013987963891675027,
      "loss": 1.2498,
      "theoretical_loss": 3.363039194298077,
      "tokens_seen": 2843213824
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013985957873620863,
      "loss": 1.3285,
      "theoretical_loss": 3.363033220093779,
      "tokens_seen": 2843279360
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013983951855566702,
      "loss": 1.2088,
      "theoretical_loss": 3.363027246065737,
      "tokens_seen": 2843344896
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013981945837512538,
      "loss": 1.3011,
      "theoretical_loss": 3.363021272213942,
      "tokens_seen": 2843410432
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013979939819458377,
      "loss": 1.2459,
      "theoretical_loss": 3.363015298538384,
      "tokens_seen": 2843475968
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013977933801404214,
      "loss": 1.3271,
      "theoretical_loss": 3.3630093250390547,
      "tokens_seen": 2843541504
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001397592778335005,
      "loss": 1.2324,
      "theoretical_loss": 3.3630033517159443,
      "tokens_seen": 2843607040
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001397392176529589,
      "loss": 1.368,
      "theoretical_loss": 3.3629973785690437,
      "tokens_seen": 2843672576
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013971915747241725,
      "loss": 1.244,
      "theoretical_loss": 3.362991405598344,
      "tokens_seen": 2843738112
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013969909729187564,
      "loss": 1.2566,
      "theoretical_loss": 3.3629854328038347,
      "tokens_seen": 2843803648
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000139679037111334,
      "loss": 1.1893,
      "theoretical_loss": 3.3629794601855076,
      "tokens_seen": 2843869184
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001396589769307924,
      "loss": 1.2513,
      "theoretical_loss": 3.3629734877433535,
      "tokens_seen": 2843934720
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013963891675025075,
      "loss": 1.2606,
      "theoretical_loss": 3.362967515477363,
      "tokens_seen": 2844000256
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013961885656970914,
      "loss": 1.3295,
      "theoretical_loss": 3.3629615433875264,
      "tokens_seen": 2844065792
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001395987963891675,
      "loss": 1.2431,
      "theoretical_loss": 3.362955571473835,
      "tokens_seen": 2844131328
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013957873620862587,
      "loss": 1.324,
      "theoretical_loss": 3.362949599736279,
      "tokens_seen": 2844196864
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013955867602808426,
      "loss": 1.3429,
      "theoretical_loss": 3.3629436281748495,
      "tokens_seen": 2844262400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013953861584754262,
      "loss": 1.2445,
      "theoretical_loss": 3.362937656789537,
      "tokens_seen": 2844327936
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000139518555667001,
      "loss": 1.3079,
      "theoretical_loss": 3.3629316855803326,
      "tokens_seen": 2844393472
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013949849548645937,
      "loss": 1.3129,
      "theoretical_loss": 3.362925714547227,
      "tokens_seen": 2844459008
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013947843530591776,
      "loss": 1.2524,
      "theoretical_loss": 3.3629197436902105,
      "tokens_seen": 2844524544
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013945837512537612,
      "loss": 1.2287,
      "theoretical_loss": 3.3629137730092746,
      "tokens_seen": 2844590080
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001394383149448345,
      "loss": 1.3457,
      "theoretical_loss": 3.3629078025044095,
      "tokens_seen": 2844655616
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013941825476429287,
      "loss": 1.2748,
      "theoretical_loss": 3.362901832175606,
      "tokens_seen": 2844721152
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013939819458375123,
      "loss": 1.2505,
      "theoretical_loss": 3.362895862022855,
      "tokens_seen": 2844786688
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013937813440320962,
      "loss": 1.2279,
      "theoretical_loss": 3.362889892046147,
      "tokens_seen": 2844852224
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013935807422266799,
      "loss": 1.2297,
      "theoretical_loss": 3.3628839222454734,
      "tokens_seen": 2844917760
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013933801404212638,
      "loss": 1.3033,
      "theoretical_loss": 3.362877952620824,
      "tokens_seen": 2844983296
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013931795386158474,
      "loss": 1.2289,
      "theoretical_loss": 3.3628719831721905,
      "tokens_seen": 2845048832
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013929789368104313,
      "loss": 1.3884,
      "theoretical_loss": 3.3628660138995627,
      "tokens_seen": 2845114368
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001392778335005015,
      "loss": 1.301,
      "theoretical_loss": 3.3628600448029324,
      "tokens_seen": 2845179904
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001392577733199599,
      "loss": 1.293,
      "theoretical_loss": 3.3628540758822894,
      "tokens_seen": 2845245440
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013923771313941827,
      "loss": 1.2973,
      "theoretical_loss": 3.3628481071376255,
      "tokens_seen": 2845310976
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013921765295887666,
      "loss": 1.1985,
      "theoretical_loss": 3.36284213856893,
      "tokens_seen": 2845376512
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013919759277833502,
      "loss": 1.261,
      "theoretical_loss": 3.362836170176195,
      "tokens_seen": 2845442048
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013917753259779338,
      "loss": 1.2936,
      "theoretical_loss": 3.3628302019594107,
      "tokens_seen": 2845507584
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013915747241725177,
      "loss": 1.2859,
      "theoretical_loss": 3.362824233918568,
      "tokens_seen": 2845573120
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013913741223671013,
      "loss": 1.2863,
      "theoretical_loss": 3.362818266053657,
      "tokens_seen": 2845638656
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013911735205616852,
      "loss": 1.2593,
      "theoretical_loss": 3.3628122983646698,
      "tokens_seen": 2845704192
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013909729187562688,
      "loss": 1.2767,
      "theoretical_loss": 3.3628063308515963,
      "tokens_seen": 2845769728
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013907723169508527,
      "loss": 1.2823,
      "theoretical_loss": 3.362800363514427,
      "tokens_seen": 2845835264
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013905717151454364,
      "loss": 1.2502,
      "theoretical_loss": 3.3627943963531535,
      "tokens_seen": 2845900800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013903711133400203,
      "loss": 1.2711,
      "theoretical_loss": 3.3627884293677655,
      "tokens_seen": 2845966336
    },
    {
      "epoch": 0.72,
      "objective/train/docs_used": 1034283,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.356002926826477,
      "objective/train/theoretical_loss": 3.362782462558255,
      "objective/train/tokens_used": 1216490976,
      "theoretical_loss": 3.362782462558255,
      "tokens_seen": 2846031872
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001390170511534604,
      "loss": 1.382,
      "theoretical_loss": 3.362782462558255,
      "tokens_seen": 2846031872
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013899699097291875,
      "loss": 1.2543,
      "theoretical_loss": 3.362776495924612,
      "tokens_seen": 2846097408
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013897693079237714,
      "loss": 1.3108,
      "theoretical_loss": 3.362770529466827,
      "tokens_seen": 2846162944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001389568706118355,
      "loss": 1.1802,
      "theoretical_loss": 3.362764563184891,
      "tokens_seen": 2846228480
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001389368104312939,
      "loss": 1.2515,
      "theoretical_loss": 3.3627585970787957,
      "tokens_seen": 2846294016
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013891675025075225,
      "loss": 1.277,
      "theoretical_loss": 3.3627526311485307,
      "tokens_seen": 2846359552
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013889669007021064,
      "loss": 1.312,
      "theoretical_loss": 3.362746665394087,
      "tokens_seen": 2846425088
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000138876629889669,
      "loss": 1.2044,
      "theoretical_loss": 3.362740699815456,
      "tokens_seen": 2846490624
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001388565697091274,
      "loss": 1.2223,
      "theoretical_loss": 3.3627347344126277,
      "tokens_seen": 2846556160
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013883650952858576,
      "loss": 1.3408,
      "theoretical_loss": 3.362728769185593,
      "tokens_seen": 2846621696
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013881644934804415,
      "loss": 1.2945,
      "theoretical_loss": 3.3627228041343433,
      "tokens_seen": 2846687232
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001387963891675025,
      "loss": 1.302,
      "theoretical_loss": 3.3627168392588684,
      "tokens_seen": 2846752768
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013877632898696087,
      "loss": 1.2981,
      "theoretical_loss": 3.36271087455916,
      "tokens_seen": 2846818304
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013875626880641926,
      "loss": 1.2763,
      "theoretical_loss": 3.362704910035209,
      "tokens_seen": 2846883840
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013873620862587762,
      "loss": 1.2566,
      "theoretical_loss": 3.3626989456870047,
      "tokens_seen": 2846949376
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000138716148445336,
      "loss": 1.2129,
      "theoretical_loss": 3.362692981514539,
      "tokens_seen": 2847014912
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013869608826479437,
      "loss": 1.3388,
      "theoretical_loss": 3.362687017517803,
      "tokens_seen": 2847080448
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013867602808425276,
      "loss": 1.2517,
      "theoretical_loss": 3.362681053696787,
      "tokens_seen": 2847145984
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013865596790371113,
      "loss": 1.1954,
      "theoretical_loss": 3.3626750900514812,
      "tokens_seen": 2847211520
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013863590772316951,
      "loss": 1.2782,
      "theoretical_loss": 3.362669126581877,
      "tokens_seen": 2847277056
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013861584754262788,
      "loss": 1.3283,
      "theoretical_loss": 3.3626631632879653,
      "tokens_seen": 2847342592
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013859578736208624,
      "loss": 1.3451,
      "theoretical_loss": 3.362657200169737,
      "tokens_seen": 2847408128
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013857572718154463,
      "loss": 1.3019,
      "theoretical_loss": 3.3626512372271824,
      "tokens_seen": 2847473664
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000138555667001003,
      "loss": 1.3031,
      "theoretical_loss": 3.3626452744602924,
      "tokens_seen": 2847539200
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013853560682046138,
      "loss": 1.2594,
      "theoretical_loss": 3.3626393118690574,
      "tokens_seen": 2847604736
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013851554663991974,
      "loss": 1.2217,
      "theoretical_loss": 3.3626333494534695,
      "tokens_seen": 2847670272
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013849548645937816,
      "loss": 1.29,
      "theoretical_loss": 3.362627387213518,
      "tokens_seen": 2847735808
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013847542627883652,
      "loss": 1.2655,
      "theoretical_loss": 3.3626214251491944,
      "tokens_seen": 2847801344
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001384553660982949,
      "loss": 1.2121,
      "theoretical_loss": 3.3626154632604894,
      "tokens_seen": 2847866880
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013843530591775327,
      "loss": 1.265,
      "theoretical_loss": 3.362609501547394,
      "tokens_seen": 2847932416
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013841524573721163,
      "loss": 1.257,
      "theoretical_loss": 3.3626035400098986,
      "tokens_seen": 2847997952
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013839518555667002,
      "loss": 1.2857,
      "theoretical_loss": 3.362597578647994,
      "tokens_seen": 2848063488
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013837512537612839,
      "loss": 1.2495,
      "theoretical_loss": 3.3625916174616712,
      "tokens_seen": 2848129024
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013835506519558678,
      "loss": 1.2006,
      "theoretical_loss": 3.362585656450921,
      "tokens_seen": 2848194560
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013833500501504514,
      "loss": 1.1878,
      "theoretical_loss": 3.362579695615734,
      "tokens_seen": 2848260096
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013831494483450353,
      "loss": 1.2493,
      "theoretical_loss": 3.3625737349561016,
      "tokens_seen": 2848325632
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001382948846539619,
      "loss": 1.218,
      "theoretical_loss": 3.3625677744720135,
      "tokens_seen": 2848391168
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013827482447342028,
      "loss": 1.3153,
      "theoretical_loss": 3.362561814163461,
      "tokens_seen": 2848456704
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013825476429287864,
      "loss": 1.2687,
      "theoretical_loss": 3.362555854030435,
      "tokens_seen": 2848522240
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013823470411233703,
      "loss": 1.2474,
      "theoretical_loss": 3.362549894072927,
      "tokens_seen": 2848587776
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001382146439317954,
      "loss": 1.4171,
      "theoretical_loss": 3.3625439342909265,
      "tokens_seen": 2848653312
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013819458375125375,
      "loss": 1.324,
      "theoretical_loss": 3.362537974684425,
      "tokens_seen": 2848718848
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013817452357071214,
      "loss": 1.3141,
      "theoretical_loss": 3.3625320152534126,
      "tokens_seen": 2848784384
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001381544633901705,
      "loss": 1.279,
      "theoretical_loss": 3.3625260559978813,
      "tokens_seen": 2848849920
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001381344032096289,
      "loss": 1.3671,
      "theoretical_loss": 3.3625200969178213,
      "tokens_seen": 2848915456
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013811434302908726,
      "loss": 1.2466,
      "theoretical_loss": 3.362514138013223,
      "tokens_seen": 2848980992
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013809428284854565,
      "loss": 1.267,
      "theoretical_loss": 3.3625081792840774,
      "tokens_seen": 2849046528
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000138074222668004,
      "loss": 1.2559,
      "theoretical_loss": 3.362502220730376,
      "tokens_seen": 2849112064
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001380541624874624,
      "loss": 1.2957,
      "theoretical_loss": 3.362496262352108,
      "tokens_seen": 2849177600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013803410230692076,
      "loss": 1.3104,
      "theoretical_loss": 3.3624903041492664,
      "tokens_seen": 2849243136
    },
    {
      "epoch": 0.73,
      "objective/train/docs_used": 1035566,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.132224678993225,
      "objective/train/theoretical_loss": 3.3624843461218403,
      "objective/train/tokens_used": 1219767776,
      "theoretical_loss": 3.3624843461218403,
      "tokens_seen": 2849308672
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013801404212637912,
      "loss": 1.2302,
      "theoretical_loss": 3.3624843461218403,
      "tokens_seen": 2849308672
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001379939819458375,
      "loss": 1.2984,
      "theoretical_loss": 3.362478388269821,
      "tokens_seen": 2849374208
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013797392176529588,
      "loss": 1.2868,
      "theoretical_loss": 3.3624724305931992,
      "tokens_seen": 2849439744
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013795386158475426,
      "loss": 1.2623,
      "theoretical_loss": 3.362466473091966,
      "tokens_seen": 2849505280
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013793380140421263,
      "loss": 1.3412,
      "theoretical_loss": 3.3624605157661125,
      "tokens_seen": 2849570816
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013791374122367102,
      "loss": 1.2942,
      "theoretical_loss": 3.3624545586156285,
      "tokens_seen": 2849636352
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013789368104312938,
      "loss": 1.2638,
      "theoretical_loss": 3.3624486016405055,
      "tokens_seen": 2849701888
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013787362086258777,
      "loss": 1.2946,
      "theoretical_loss": 3.362442644840734,
      "tokens_seen": 2849767424
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013785356068204613,
      "loss": 1.1851,
      "theoretical_loss": 3.362436688216305,
      "tokens_seen": 2849832960
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001378335005015045,
      "loss": 1.4534,
      "theoretical_loss": 3.3624307317672093,
      "tokens_seen": 2849898496
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013781344032096288,
      "loss": 1.2632,
      "theoretical_loss": 3.3624247754934378,
      "tokens_seen": 2849964032
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013779338014042124,
      "loss": 1.2205,
      "theoretical_loss": 3.362418819394981,
      "tokens_seen": 2850029568
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013777331995987963,
      "loss": 1.2862,
      "theoretical_loss": 3.36241286347183,
      "tokens_seen": 2850095104
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013775325977933802,
      "loss": 1.2347,
      "theoretical_loss": 3.3624069077239755,
      "tokens_seen": 2850160640
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001377331995987964,
      "loss": 1.3118,
      "theoretical_loss": 3.3624009521514084,
      "tokens_seen": 2850226176
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013771313941825477,
      "loss": 1.2523,
      "theoretical_loss": 3.3623949967541193,
      "tokens_seen": 2850291712
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013769307923771316,
      "loss": 1.2328,
      "theoretical_loss": 3.362389041532099,
      "tokens_seen": 2850357248
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013767301905717153,
      "loss": 1.3733,
      "theoretical_loss": 3.362383086485339,
      "tokens_seen": 2850422784
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013765295887662991,
      "loss": 1.2517,
      "theoretical_loss": 3.3623771316138287,
      "tokens_seen": 2850488320
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013763289869608828,
      "loss": 1.333,
      "theoretical_loss": 3.36237117691756,
      "tokens_seen": 2850553856
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013761283851554664,
      "loss": 1.294,
      "theoretical_loss": 3.362365222396524,
      "tokens_seen": 2850619392
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013759277833500503,
      "loss": 1.3312,
      "theoretical_loss": 3.3623592680507106,
      "tokens_seen": 2850684928
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001375727181544634,
      "loss": 1.3046,
      "theoretical_loss": 3.3623533138801114,
      "tokens_seen": 2850750464
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013755265797392178,
      "loss": 1.3821,
      "theoretical_loss": 3.3623473598847164,
      "tokens_seen": 2850816000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013753259779338014,
      "loss": 1.2815,
      "theoretical_loss": 3.362341406064517,
      "tokens_seen": 2850881536
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013751253761283853,
      "loss": 1.3055,
      "theoretical_loss": 3.362335452419504,
      "tokens_seen": 2850947072
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001374924774322969,
      "loss": 1.2355,
      "theoretical_loss": 3.3623294989496677,
      "tokens_seen": 2851012608
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013747241725175528,
      "loss": 1.2259,
      "theoretical_loss": 3.362323545655,
      "tokens_seen": 2851078144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013745235707121365,
      "loss": 1.2028,
      "theoretical_loss": 3.36231759253549,
      "tokens_seen": 2851143680
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000137432296890672,
      "loss": 1.3734,
      "theoretical_loss": 3.3623116395911303,
      "tokens_seen": 2851209216
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001374122367101304,
      "loss": 1.2446,
      "theoretical_loss": 3.362305686821911,
      "tokens_seen": 2851274752
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013739217652958876,
      "loss": 1.3479,
      "theoretical_loss": 3.3622997342278227,
      "tokens_seen": 2851340288
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013737211634904715,
      "loss": 1.3305,
      "theoretical_loss": 3.362293781808856,
      "tokens_seen": 2851405824
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001373520561685055,
      "loss": 1.2962,
      "theoretical_loss": 3.3622878295650027,
      "tokens_seen": 2851471360
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001373319959879639,
      "loss": 1.4244,
      "theoretical_loss": 3.362281877496253,
      "tokens_seen": 2851536896
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013731193580742226,
      "loss": 1.3285,
      "theoretical_loss": 3.3622759256025976,
      "tokens_seen": 2851602432
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013729187562688065,
      "loss": 1.3128,
      "theoretical_loss": 3.362269973884027,
      "tokens_seen": 2851667968
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013727181544633901,
      "loss": 1.251,
      "theoretical_loss": 3.3622640223405336,
      "tokens_seen": 2851733504
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013725175526579738,
      "loss": 1.1932,
      "theoretical_loss": 3.3622580709721066,
      "tokens_seen": 2851799040
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013723169508525577,
      "loss": 1.1893,
      "theoretical_loss": 3.3622521197787374,
      "tokens_seen": 2851864576
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013721163490471413,
      "loss": 1.3322,
      "theoretical_loss": 3.3622461687604166,
      "tokens_seen": 2851930112
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013719157472417252,
      "loss": 1.2787,
      "theoretical_loss": 3.362240217917136,
      "tokens_seen": 2851995648
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013717151454363088,
      "loss": 1.3186,
      "theoretical_loss": 3.3622342672488847,
      "tokens_seen": 2852061184
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013715145436308927,
      "loss": 1.3197,
      "theoretical_loss": 3.362228316755655,
      "tokens_seen": 2852126720
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013713139418254763,
      "loss": 1.3807,
      "theoretical_loss": 3.3622223664374373,
      "tokens_seen": 2852192256
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013711133400200602,
      "loss": 1.3125,
      "theoretical_loss": 3.3622164162942223,
      "tokens_seen": 2852257792
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013709127382146438,
      "loss": 1.1432,
      "theoretical_loss": 3.3622104663260006,
      "tokens_seen": 2852323328
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013707121364092277,
      "loss": 1.2479,
      "theoretical_loss": 3.3622045165327634,
      "tokens_seen": 2852388864
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013705115346038113,
      "loss": 1.2258,
      "theoretical_loss": 3.362198566914502,
      "tokens_seen": 2852454400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001370310932798395,
      "loss": 1.3217,
      "theoretical_loss": 3.362192617471206,
      "tokens_seen": 2852519936
    },
    {
      "epoch": 0.73,
      "objective/train/docs_used": 1036772,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2533670663833618,
      "objective/train/theoretical_loss": 3.3621866682028676,
      "objective/train/tokens_used": 1223044576,
      "theoretical_loss": 3.3621866682028676,
      "tokens_seen": 2852585472
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001370110330992979,
      "loss": 1.2242,
      "theoretical_loss": 3.3621866682028676,
      "tokens_seen": 2852585472
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013699097291875628,
      "loss": 1.2711,
      "theoretical_loss": 3.3621807191094764,
      "tokens_seen": 2852651008
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013697091273821466,
      "loss": 1.2958,
      "theoretical_loss": 3.3621747701910243,
      "tokens_seen": 2852716544
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013695085255767303,
      "loss": 1.3542,
      "theoretical_loss": 3.362168821447501,
      "tokens_seen": 2852782080
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013693079237713142,
      "loss": 1.3016,
      "theoretical_loss": 3.3621628728788986,
      "tokens_seen": 2852847616
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013691073219658978,
      "loss": 1.3225,
      "theoretical_loss": 3.362156924485207,
      "tokens_seen": 2852913152
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013689067201604817,
      "loss": 1.3364,
      "theoretical_loss": 3.3621509762664177,
      "tokens_seen": 2852978688
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013687061183550653,
      "loss": 1.3183,
      "theoretical_loss": 3.3621450282225207,
      "tokens_seen": 2853044224
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001368505516549649,
      "loss": 1.3261,
      "theoretical_loss": 3.362139080353508,
      "tokens_seen": 2853109760
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013683049147442328,
      "loss": 1.2838,
      "theoretical_loss": 3.362133132659369,
      "tokens_seen": 2853175296
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013681043129388164,
      "loss": 1.2428,
      "theoretical_loss": 3.362127185140096,
      "tokens_seen": 2853240832
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013679037111334003,
      "loss": 1.1676,
      "theoretical_loss": 3.3621212377956784,
      "tokens_seen": 2853306368
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001367703109327984,
      "loss": 1.3305,
      "theoretical_loss": 3.3621152906261083,
      "tokens_seen": 2853371904
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013675025075225678,
      "loss": 1.3409,
      "theoretical_loss": 3.362109343631376,
      "tokens_seen": 2853437440
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013673019057171515,
      "loss": 1.1933,
      "theoretical_loss": 3.3621033968114724,
      "tokens_seen": 2853502976
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013671013039117354,
      "loss": 1.2475,
      "theoretical_loss": 3.362097450166388,
      "tokens_seen": 2853568512
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001366900702106319,
      "loss": 1.3327,
      "theoretical_loss": 3.3620915036961145,
      "tokens_seen": 2853634048
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001366700100300903,
      "loss": 1.2484,
      "theoretical_loss": 3.362085557400642,
      "tokens_seen": 2853699584
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013664994984954865,
      "loss": 1.2361,
      "theoretical_loss": 3.362079611279962,
      "tokens_seen": 2853765120
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000136629889669007,
      "loss": 1.233,
      "theoretical_loss": 3.3620736653340644,
      "tokens_seen": 2853830656
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001366098294884654,
      "loss": 1.2689,
      "theoretical_loss": 3.3620677195629405,
      "tokens_seen": 2853896192
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013658976930792376,
      "loss": 1.2188,
      "theoretical_loss": 3.3620617739665817,
      "tokens_seen": 2853961728
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013656970912738215,
      "loss": 1.1131,
      "theoretical_loss": 3.362055828544978,
      "tokens_seen": 2854027264
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013654964894684052,
      "loss": 1.3007,
      "theoretical_loss": 3.3620498832981207,
      "tokens_seen": 2854092800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001365295887662989,
      "loss": 1.3021,
      "theoretical_loss": 3.362043938226001,
      "tokens_seen": 2854158336
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013650952858575727,
      "loss": 1.3598,
      "theoretical_loss": 3.3620379933286086,
      "tokens_seen": 2854223872
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013648946840521566,
      "loss": 1.2669,
      "theoretical_loss": 3.3620320486059354,
      "tokens_seen": 2854289408
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013646940822467402,
      "loss": 1.3017,
      "theoretical_loss": 3.362026104057972,
      "tokens_seen": 2854354944
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013644934804413238,
      "loss": 1.2837,
      "theoretical_loss": 3.3620201596847092,
      "tokens_seen": 2854420480
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013642928786359077,
      "loss": 1.3427,
      "theoretical_loss": 3.362014215486138,
      "tokens_seen": 2854486016
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013640922768304913,
      "loss": 1.3239,
      "theoretical_loss": 3.362008271462249,
      "tokens_seen": 2854551552
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013638916750250752,
      "loss": 1.3163,
      "theoretical_loss": 3.362002327613033,
      "tokens_seen": 2854617088
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013636910732196588,
      "loss": 1.3888,
      "theoretical_loss": 3.361996383938481,
      "tokens_seen": 2854682624
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013634904714142427,
      "loss": 1.1838,
      "theoretical_loss": 3.361990440438584,
      "tokens_seen": 2854748160
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013632898696088264,
      "loss": 1.2234,
      "theoretical_loss": 3.361984497113333,
      "tokens_seen": 2854813696
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013630892678034103,
      "loss": 1.2601,
      "theoretical_loss": 3.361978553962718,
      "tokens_seen": 2854879232
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001362888665997994,
      "loss": 1.2467,
      "theoretical_loss": 3.3619726109867307,
      "tokens_seen": 2854944768
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013626880641925775,
      "loss": 1.2645,
      "theoretical_loss": 3.361966668185362,
      "tokens_seen": 2855010304
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013624874623871617,
      "loss": 1.3255,
      "theoretical_loss": 3.3619607255586024,
      "tokens_seen": 2855075840
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013622868605817453,
      "loss": 1.2271,
      "theoretical_loss": 3.3619547831064427,
      "tokens_seen": 2855141376
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013620862587763292,
      "loss": 1.3389,
      "theoretical_loss": 3.3619488408288736,
      "tokens_seen": 2855206912
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013618856569709128,
      "loss": 1.3132,
      "theoretical_loss": 3.3619428987258866,
      "tokens_seen": 2855272448
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013616850551654967,
      "loss": 1.2938,
      "theoretical_loss": 3.361936956797472,
      "tokens_seen": 2855337984
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013614844533600803,
      "loss": 1.3458,
      "theoretical_loss": 3.3619310150436212,
      "tokens_seen": 2855403520
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013612838515546642,
      "loss": 1.2606,
      "theoretical_loss": 3.3619250734643247,
      "tokens_seen": 2855469056
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013610832497492478,
      "loss": 1.3885,
      "theoretical_loss": 3.3619191320595734,
      "tokens_seen": 2855534592
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013608826479438317,
      "loss": 1.335,
      "theoretical_loss": 3.361913190829358,
      "tokens_seen": 2855600128
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013606820461384153,
      "loss": 1.3151,
      "theoretical_loss": 3.3619072497736697,
      "tokens_seen": 2855665664
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001360481444332999,
      "loss": 1.1934,
      "theoretical_loss": 3.361901308892499,
      "tokens_seen": 2855731200
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013602808425275829,
      "loss": 1.2572,
      "theoretical_loss": 3.361895368185837,
      "tokens_seen": 2855796736
    },
    {
      "epoch": 0.73,
      "objective/train/docs_used": 1037860,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2703384160995483,
      "objective/train/theoretical_loss": 3.361889427653675,
      "objective/train/tokens_used": 1226321376,
      "theoretical_loss": 3.361889427653675,
      "tokens_seen": 2855862272
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013600802407221665,
      "loss": 1.3591,
      "theoretical_loss": 3.361889427653675,
      "tokens_seen": 2855862272
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013598796389167504,
      "loss": 1.3093,
      "theoretical_loss": 3.3618834872960033,
      "tokens_seen": 2855927808
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001359679037111334,
      "loss": 1.3253,
      "theoretical_loss": 3.3618775471128126,
      "tokens_seen": 2855993344
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001359478435305918,
      "loss": 1.3939,
      "theoretical_loss": 3.3618716071040944,
      "tokens_seen": 2856058880
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013592778335005015,
      "loss": 1.3135,
      "theoretical_loss": 3.361865667269839,
      "tokens_seen": 2856124416
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013590772316950854,
      "loss": 1.2988,
      "theoretical_loss": 3.3618597276100375,
      "tokens_seen": 2856189952
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001358876629889669,
      "loss": 1.3361,
      "theoretical_loss": 3.361853788124681,
      "tokens_seen": 2856255488
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013586760280842527,
      "loss": 1.2995,
      "theoretical_loss": 3.3618478488137598,
      "tokens_seen": 2856321024
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013584754262788365,
      "loss": 1.3154,
      "theoretical_loss": 3.3618419096772656,
      "tokens_seen": 2856386560
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013582748244734202,
      "loss": 1.359,
      "theoretical_loss": 3.3618359707151884,
      "tokens_seen": 2856452096
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001358074222668004,
      "loss": 1.2401,
      "theoretical_loss": 3.3618300319275196,
      "tokens_seen": 2856517632
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013578736208625877,
      "loss": 1.3276,
      "theoretical_loss": 3.36182409331425,
      "tokens_seen": 2856583168
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013576730190571716,
      "loss": 1.2698,
      "theoretical_loss": 3.3618181548753707,
      "tokens_seen": 2856648704
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013574724172517552,
      "loss": 1.2522,
      "theoretical_loss": 3.3618122166108724,
      "tokens_seen": 2856714240
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001357271815446339,
      "loss": 1.3587,
      "theoretical_loss": 3.3618062785207456,
      "tokens_seen": 2856779776
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013570712136409227,
      "loss": 1.2658,
      "theoretical_loss": 3.3618003406049817,
      "tokens_seen": 2856845312
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013568706118355063,
      "loss": 1.2813,
      "theoretical_loss": 3.361794402863571,
      "tokens_seen": 2856910848
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013566700100300902,
      "loss": 1.3108,
      "theoretical_loss": 3.361788465296505,
      "tokens_seen": 2856976384
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013564694082246739,
      "loss": 1.3942,
      "theoretical_loss": 3.3617825279037743,
      "tokens_seen": 2857041920
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013562688064192578,
      "loss": 1.3227,
      "theoretical_loss": 3.36177659068537,
      "tokens_seen": 2857107456
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013560682046138414,
      "loss": 1.3799,
      "theoretical_loss": 3.3617706536412824,
      "tokens_seen": 2857172992
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013558676028084253,
      "loss": 1.3857,
      "theoretical_loss": 3.361764716771503,
      "tokens_seen": 2857238528
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001355667001003009,
      "loss": 1.2706,
      "theoretical_loss": 3.3617587800760225,
      "tokens_seen": 2857304064
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013554663991975928,
      "loss": 1.2951,
      "theoretical_loss": 3.3617528435548314,
      "tokens_seen": 2857369600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013552657973921764,
      "loss": 1.2593,
      "theoretical_loss": 3.3617469072079214,
      "tokens_seen": 2857435136
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013550651955867606,
      "loss": 1.2879,
      "theoretical_loss": 3.3617409710352826,
      "tokens_seen": 2857500672
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013548645937813442,
      "loss": 1.1621,
      "theoretical_loss": 3.361735035036906,
      "tokens_seen": 2857566208
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013546639919759278,
      "loss": 1.1571,
      "theoretical_loss": 3.3617290992127833,
      "tokens_seen": 2857631744
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013544633901705117,
      "loss": 1.2811,
      "theoretical_loss": 3.3617231635629046,
      "tokens_seen": 2857697280
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013542627883650953,
      "loss": 1.2069,
      "theoretical_loss": 3.3617172280872607,
      "tokens_seen": 2857762816
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013540621865596792,
      "loss": 1.3711,
      "theoretical_loss": 3.3617112927858432,
      "tokens_seen": 2857828352
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013538615847542628,
      "loss": 1.249,
      "theoretical_loss": 3.361705357658642,
      "tokens_seen": 2857893888
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013536609829488467,
      "loss": 1.2551,
      "theoretical_loss": 3.361699422705649,
      "tokens_seen": 2857959424
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013534603811434304,
      "loss": 1.3344,
      "theoretical_loss": 3.3616934879268543,
      "tokens_seen": 2858024960
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013532597793380143,
      "loss": 1.2654,
      "theoretical_loss": 3.3616875533222492,
      "tokens_seen": 2858090496
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001353059177532598,
      "loss": 1.2144,
      "theoretical_loss": 3.361681618891825,
      "tokens_seen": 2858156032
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013528585757271815,
      "loss": 1.2414,
      "theoretical_loss": 3.3616756846355718,
      "tokens_seen": 2858221568
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013526579739217654,
      "loss": 1.2314,
      "theoretical_loss": 3.3616697505534807,
      "tokens_seen": 2858287104
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001352457372116349,
      "loss": 1.2107,
      "theoretical_loss": 3.3616638166455424,
      "tokens_seen": 2858352640
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001352256770310933,
      "loss": 1.2958,
      "theoretical_loss": 3.3616578829117487,
      "tokens_seen": 2858418176
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013520561685055165,
      "loss": 1.2691,
      "theoretical_loss": 3.3616519493520896,
      "tokens_seen": 2858483712
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013518555667001004,
      "loss": 1.3637,
      "theoretical_loss": 3.3616460159665564,
      "tokens_seen": 2858549248
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001351654964894684,
      "loss": 1.2776,
      "theoretical_loss": 3.36164008275514,
      "tokens_seen": 2858614784
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001351454363089268,
      "loss": 1.2661,
      "theoretical_loss": 3.361634149717831,
      "tokens_seen": 2858680320
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013512537612838516,
      "loss": 1.2631,
      "theoretical_loss": 3.3616282168546205,
      "tokens_seen": 2858745856
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013510531594784355,
      "loss": 1.2287,
      "theoretical_loss": 3.3616222841654992,
      "tokens_seen": 2858811392
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001350852557673019,
      "loss": 1.2792,
      "theoretical_loss": 3.3616163516504587,
      "tokens_seen": 2858876928
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013506519558676027,
      "loss": 1.2777,
      "theoretical_loss": 3.361610419309489,
      "tokens_seen": 2858942464
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013504513540621866,
      "loss": 1.252,
      "theoretical_loss": 3.3616044871425816,
      "tokens_seen": 2859008000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013502507522567702,
      "loss": 1.287,
      "theoretical_loss": 3.361598555149727,
      "tokens_seen": 2859073536
    },
    {
      "epoch": 0.73,
      "objective/train/docs_used": 1039109,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.074881911277771,
      "objective/train/theoretical_loss": 3.3615926233309166,
      "objective/train/tokens_used": 1229598176,
      "theoretical_loss": 3.3615926233309166,
      "tokens_seen": 2859139072
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001350050150451354,
      "loss": 1.2146,
      "theoretical_loss": 3.3615926233309166,
      "tokens_seen": 2859139072
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013498495486459377,
      "loss": 1.2912,
      "theoretical_loss": 3.3615866916861408,
      "tokens_seen": 2859204608
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013496489468405216,
      "loss": 1.2436,
      "theoretical_loss": 3.361580760215391,
      "tokens_seen": 2859270144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013494483450351052,
      "loss": 1.3489,
      "theoretical_loss": 3.3615748289186573,
      "tokens_seen": 2859335680
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013492477432296891,
      "loss": 1.2693,
      "theoretical_loss": 3.3615688977959315,
      "tokens_seen": 2859401216
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013490471414242728,
      "loss": 1.2564,
      "theoretical_loss": 3.3615629668472042,
      "tokens_seen": 2859466752
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013488465396188564,
      "loss": 1.177,
      "theoretical_loss": 3.361557036072466,
      "tokens_seen": 2859532288
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013486459378134403,
      "loss": 1.2082,
      "theoretical_loss": 3.361551105471708,
      "tokens_seen": 2859597824
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001348445336008024,
      "loss": 1.322,
      "theoretical_loss": 3.3615451750449212,
      "tokens_seen": 2859663360
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013482447342026078,
      "loss": 1.3585,
      "theoretical_loss": 3.361539244792097,
      "tokens_seen": 2859728896
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013480441323971914,
      "loss": 1.358,
      "theoretical_loss": 3.361533314713225,
      "tokens_seen": 2859794432
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013478435305917753,
      "loss": 1.314,
      "theoretical_loss": 3.3615273848082974,
      "tokens_seen": 2859859968
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001347642928786359,
      "loss": 1.2707,
      "theoretical_loss": 3.3615214550773045,
      "tokens_seen": 2859925504
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001347442326980943,
      "loss": 1.3291,
      "theoretical_loss": 3.3615155255202374,
      "tokens_seen": 2859991040
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013472417251755267,
      "loss": 1.3199,
      "theoretical_loss": 3.3615095961370867,
      "tokens_seen": 2860056576
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013470411233701103,
      "loss": 1.2956,
      "theoretical_loss": 3.3615036669278435,
      "tokens_seen": 2860122112
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013468405215646942,
      "loss": 1.3672,
      "theoretical_loss": 3.361497737892499,
      "tokens_seen": 2860187648
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013466399197592779,
      "loss": 1.3036,
      "theoretical_loss": 3.361491809031044,
      "tokens_seen": 2860253184
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013464393179538618,
      "loss": 1.3685,
      "theoretical_loss": 3.361485880343469,
      "tokens_seen": 2860318720
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013462387161484454,
      "loss": 1.2023,
      "theoretical_loss": 3.3614799518297653,
      "tokens_seen": 2860384256
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013460381143430293,
      "loss": 1.31,
      "theoretical_loss": 3.3614740234899236,
      "tokens_seen": 2860449792
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001345837512537613,
      "loss": 1.2839,
      "theoretical_loss": 3.361468095323935,
      "tokens_seen": 2860515328
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013456369107321968,
      "loss": 1.201,
      "theoretical_loss": 3.3614621673317906,
      "tokens_seen": 2860580864
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013454363089267804,
      "loss": 1.3044,
      "theoretical_loss": 3.3614562395134806,
      "tokens_seen": 2860646400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013452357071213643,
      "loss": 1.1609,
      "theoretical_loss": 3.361450311868997,
      "tokens_seen": 2860711936
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001345035105315948,
      "loss": 1.3288,
      "theoretical_loss": 3.3614443843983297,
      "tokens_seen": 2860777472
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013448345035105315,
      "loss": 1.3421,
      "theoretical_loss": 3.36143845710147,
      "tokens_seen": 2860843008
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013446339017051154,
      "loss": 1.3293,
      "theoretical_loss": 3.3614325299784094,
      "tokens_seen": 2860908544
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001344433299899699,
      "loss": 1.2366,
      "theoretical_loss": 3.3614266030291375,
      "tokens_seen": 2860974080
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001344232698094283,
      "loss": 1.2706,
      "theoretical_loss": 3.3614206762536467,
      "tokens_seen": 2861039616
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013440320962888666,
      "loss": 1.2745,
      "theoretical_loss": 3.361414749651927,
      "tokens_seen": 2861105152
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013438314944834505,
      "loss": 1.2862,
      "theoretical_loss": 3.3614088232239694,
      "tokens_seen": 2861170688
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001343630892678034,
      "loss": 1.323,
      "theoretical_loss": 3.361402896969765,
      "tokens_seen": 2861236224
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001343430290872618,
      "loss": 1.2055,
      "theoretical_loss": 3.361396970889305,
      "tokens_seen": 2861301760
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013432296890672016,
      "loss": 1.3283,
      "theoretical_loss": 3.3613910449825797,
      "tokens_seen": 2861367296
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013430290872617852,
      "loss": 1.299,
      "theoretical_loss": 3.3613851192495803,
      "tokens_seen": 2861432832
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001342828485456369,
      "loss": 1.4004,
      "theoretical_loss": 3.361379193690298,
      "tokens_seen": 2861498368
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013426278836509527,
      "loss": 1.2459,
      "theoretical_loss": 3.3613732683047237,
      "tokens_seen": 2861563904
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013424272818455366,
      "loss": 1.3234,
      "theoretical_loss": 3.361367343092848,
      "tokens_seen": 2861629440
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013422266800401203,
      "loss": 1.2787,
      "theoretical_loss": 3.361361418054662,
      "tokens_seen": 2861694976
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013420260782347042,
      "loss": 1.3121,
      "theoretical_loss": 3.361355493190157,
      "tokens_seen": 2861760512
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013418254764292878,
      "loss": 1.2439,
      "theoretical_loss": 3.361349568499323,
      "tokens_seen": 2861826048
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013416248746238717,
      "loss": 1.2753,
      "theoretical_loss": 3.3613436439821514,
      "tokens_seen": 2861891584
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013414242728184553,
      "loss": 1.2078,
      "theoretical_loss": 3.3613377196386334,
      "tokens_seen": 2861957120
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001341223671013039,
      "loss": 1.2074,
      "theoretical_loss": 3.3613317954687596,
      "tokens_seen": 2862022656
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013410230692076228,
      "loss": 1.2907,
      "theoretical_loss": 3.361325871472521,
      "tokens_seen": 2862088192
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013408224674022064,
      "loss": 1.3058,
      "theoretical_loss": 3.361319947649909,
      "tokens_seen": 2862153728
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013406218655967903,
      "loss": 1.3209,
      "theoretical_loss": 3.3613140240009143,
      "tokens_seen": 2862219264
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001340421263791374,
      "loss": 1.1996,
      "theoretical_loss": 3.361308100525527,
      "tokens_seen": 2862284800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013402206619859578,
      "loss": 1.265,
      "theoretical_loss": 3.3613021772237395,
      "tokens_seen": 2862350336
    },
    {
      "epoch": 0.73,
      "objective/train/docs_used": 1040389,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3280324935913086,
      "objective/train/theoretical_loss": 3.3612962540955413,
      "objective/train/tokens_used": 1232874976,
      "theoretical_loss": 3.3612962540955413,
      "tokens_seen": 2862415872
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013400200601805417,
      "loss": 1.3278,
      "theoretical_loss": 3.3612962540955413,
      "tokens_seen": 2862415872
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013398194583751256,
      "loss": 1.2615,
      "theoretical_loss": 3.3612903311409243,
      "tokens_seen": 2862481408
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013396188565697093,
      "loss": 1.1986,
      "theoretical_loss": 3.3612844083598796,
      "tokens_seen": 2862546944
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013394182547642931,
      "loss": 1.2522,
      "theoretical_loss": 3.361278485752397,
      "tokens_seen": 2862612480
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013392176529588768,
      "loss": 1.264,
      "theoretical_loss": 3.3612725633184684,
      "tokens_seen": 2862678016
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013390170511534604,
      "loss": 1.3042,
      "theoretical_loss": 3.3612666410580845,
      "tokens_seen": 2862743552
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013388164493480443,
      "loss": 1.2564,
      "theoretical_loss": 3.361260718971236,
      "tokens_seen": 2862809088
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001338615847542628,
      "loss": 1.2819,
      "theoretical_loss": 3.361254797057914,
      "tokens_seen": 2862874624
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013384152457372118,
      "loss": 1.259,
      "theoretical_loss": 3.36124887531811,
      "tokens_seen": 2862940160
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013382146439317954,
      "loss": 1.4066,
      "theoretical_loss": 3.361242953751814,
      "tokens_seen": 2863005696
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013380140421263793,
      "loss": 1.2806,
      "theoretical_loss": 3.3612370323590173,
      "tokens_seen": 2863071232
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001337813440320963,
      "loss": 1.3296,
      "theoretical_loss": 3.361231111139711,
      "tokens_seen": 2863136768
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013376128385155468,
      "loss": 1.3041,
      "theoretical_loss": 3.3612251900938857,
      "tokens_seen": 2863202304
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013374122367101305,
      "loss": 1.2624,
      "theoretical_loss": 3.3612192692215332,
      "tokens_seen": 2863267840
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001337211634904714,
      "loss": 1.1185,
      "theoretical_loss": 3.3612133485226434,
      "tokens_seen": 2863333376
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001337011033099298,
      "loss": 1.3558,
      "theoretical_loss": 3.361207427997208,
      "tokens_seen": 2863398912
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013368104312938816,
      "loss": 1.2712,
      "theoretical_loss": 3.3612015076452177,
      "tokens_seen": 2863464448
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013366098294884655,
      "loss": 1.1985,
      "theoretical_loss": 3.3611955874666632,
      "tokens_seen": 2863529984
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001336409227683049,
      "loss": 1.2751,
      "theoretical_loss": 3.361189667461536,
      "tokens_seen": 2863595520
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001336208625877633,
      "loss": 1.3259,
      "theoretical_loss": 3.361183747629826,
      "tokens_seen": 2863661056
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013360080240722166,
      "loss": 1.3511,
      "theoretical_loss": 3.3611778279715256,
      "tokens_seen": 2863726592
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013358074222668005,
      "loss": 1.3354,
      "theoretical_loss": 3.3611719084866247,
      "tokens_seen": 2863792128
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013356068204613841,
      "loss": 1.2503,
      "theoretical_loss": 3.361165989175115,
      "tokens_seen": 2863857664
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013354062186559678,
      "loss": 1.2569,
      "theoretical_loss": 3.3611600700369864,
      "tokens_seen": 2863923200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013352056168505517,
      "loss": 1.2663,
      "theoretical_loss": 3.3611541510722307,
      "tokens_seen": 2863988736
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013350050150451353,
      "loss": 1.2834,
      "theoretical_loss": 3.3611482322808386,
      "tokens_seen": 2864054272
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013348044132397192,
      "loss": 1.337,
      "theoretical_loss": 3.361142313662801,
      "tokens_seen": 2864119808
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013346038114343028,
      "loss": 1.2842,
      "theoretical_loss": 3.361136395218109,
      "tokens_seen": 2864185344
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013344032096288867,
      "loss": 1.2324,
      "theoretical_loss": 3.3611304769467534,
      "tokens_seen": 2864250880
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013342026078234703,
      "loss": 1.3692,
      "theoretical_loss": 3.3611245588487257,
      "tokens_seen": 2864316416
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013340020060180542,
      "loss": 1.2172,
      "theoretical_loss": 3.361118640924016,
      "tokens_seen": 2864381952
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013338014042126378,
      "loss": 1.2688,
      "theoretical_loss": 3.361112723172616,
      "tokens_seen": 2864447488
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013336008024072217,
      "loss": 1.3351,
      "theoretical_loss": 3.361106805594516,
      "tokens_seen": 2864513024
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013334002006018053,
      "loss": 1.2518,
      "theoretical_loss": 3.361100888189707,
      "tokens_seen": 2864578560
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001333199598796389,
      "loss": 1.1999,
      "theoretical_loss": 3.361094970958181,
      "tokens_seen": 2864644096
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013329989969909729,
      "loss": 1.3603,
      "theoretical_loss": 3.361089053899928,
      "tokens_seen": 2864709632
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013327983951855565,
      "loss": 1.3838,
      "theoretical_loss": 3.3610831370149388,
      "tokens_seen": 2864775168
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013325977933801404,
      "loss": 1.3036,
      "theoretical_loss": 3.361077220303205,
      "tokens_seen": 2864840704
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013323971915747243,
      "loss": 1.2994,
      "theoretical_loss": 3.3610713037647173,
      "tokens_seen": 2864906240
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013321965897693082,
      "loss": 1.3162,
      "theoretical_loss": 3.3610653873994667,
      "tokens_seen": 2864971776
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013319959879638918,
      "loss": 1.398,
      "theoretical_loss": 3.361059471207444,
      "tokens_seen": 2865037312
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013317953861584757,
      "loss": 1.2654,
      "theoretical_loss": 3.3610535551886405,
      "tokens_seen": 2865102848
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013315947843530593,
      "loss": 1.2778,
      "theoretical_loss": 3.361047639343047,
      "tokens_seen": 2865168384
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001331394182547643,
      "loss": 1.2952,
      "theoretical_loss": 3.3610417236706542,
      "tokens_seen": 2865233920
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013311935807422268,
      "loss": 1.2807,
      "theoretical_loss": 3.3610358081714535,
      "tokens_seen": 2865299456
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013309929789368104,
      "loss": 1.3352,
      "theoretical_loss": 3.3610298928454356,
      "tokens_seen": 2865364992
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013307923771313943,
      "loss": 1.2755,
      "theoretical_loss": 3.361023977692591,
      "tokens_seen": 2865430528
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001330591775325978,
      "loss": 1.164,
      "theoretical_loss": 3.361018062712912,
      "tokens_seen": 2865496064
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013303911735205618,
      "loss": 1.2582,
      "theoretical_loss": 3.3610121479063886,
      "tokens_seen": 2865561600
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013301905717151455,
      "loss": 1.2961,
      "theoretical_loss": 3.3610062332730117,
      "tokens_seen": 2865627136
    },
    {
      "epoch": 0.74,
      "objective/train/docs_used": 1041644,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.493830680847168,
      "objective/train/theoretical_loss": 3.3610003188127724,
      "objective/train/tokens_used": 1236151776,
      "theoretical_loss": 3.3610003188127724,
      "tokens_seen": 2865692672
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013299899699097294,
      "loss": 1.4425,
      "theoretical_loss": 3.3610003188127724,
      "tokens_seen": 2865692672
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001329789368104313,
      "loss": 1.3133,
      "theoretical_loss": 3.360994404525662,
      "tokens_seen": 2865758208
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001329588766298897,
      "loss": 1.3048,
      "theoretical_loss": 3.3609884904116716,
      "tokens_seen": 2865823744
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013293881644934805,
      "loss": 1.2552,
      "theoretical_loss": 3.3609825764707915,
      "tokens_seen": 2865889280
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001329187562688064,
      "loss": 1.2318,
      "theoretical_loss": 3.3609766627030133,
      "tokens_seen": 2865954816
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001328986960882648,
      "loss": 1.3245,
      "theoretical_loss": 3.3609707491083274,
      "tokens_seen": 2866020352
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013287863590772316,
      "loss": 1.2276,
      "theoretical_loss": 3.3609648356867248,
      "tokens_seen": 2866085888
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013285857572718155,
      "loss": 1.1997,
      "theoretical_loss": 3.360958922438197,
      "tokens_seen": 2866151424
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013283851554663992,
      "loss": 1.3198,
      "theoretical_loss": 3.360953009362735,
      "tokens_seen": 2866216960
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001328184553660983,
      "loss": 1.261,
      "theoretical_loss": 3.360947096460329,
      "tokens_seen": 2866282496
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013279839518555667,
      "loss": 1.1908,
      "theoretical_loss": 3.360941183730971,
      "tokens_seen": 2866348032
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013277833500501506,
      "loss": 1.354,
      "theoretical_loss": 3.360935271174651,
      "tokens_seen": 2866413568
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013275827482447342,
      "loss": 1.2838,
      "theoretical_loss": 3.3609293587913607,
      "tokens_seen": 2866479104
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013273821464393178,
      "loss": 1.248,
      "theoretical_loss": 3.360923446581091,
      "tokens_seen": 2866544640
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013271815446339017,
      "loss": 1.3487,
      "theoretical_loss": 3.360917534543832,
      "tokens_seen": 2866610176
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013269809428284853,
      "loss": 1.4034,
      "theoretical_loss": 3.360911622679576,
      "tokens_seen": 2866675712
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013267803410230692,
      "loss": 1.3126,
      "theoretical_loss": 3.3609057109883134,
      "tokens_seen": 2866741248
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013265797392176528,
      "loss": 1.3419,
      "theoretical_loss": 3.3608997994700345,
      "tokens_seen": 2866806784
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013263791374122367,
      "loss": 1.2162,
      "theoretical_loss": 3.3608938881247314,
      "tokens_seen": 2866872320
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013261785356068204,
      "loss": 1.3526,
      "theoretical_loss": 3.3608879769523945,
      "tokens_seen": 2866937856
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013259779338014043,
      "loss": 1.267,
      "theoretical_loss": 3.3608820659530148,
      "tokens_seen": 2867003392
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001325777331995988,
      "loss": 1.348,
      "theoretical_loss": 3.3608761551265838,
      "tokens_seen": 2867068928
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013255767301905715,
      "loss": 1.1352,
      "theoretical_loss": 3.360870244473092,
      "tokens_seen": 2867134464
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013253761283851554,
      "loss": 1.2705,
      "theoretical_loss": 3.36086433399253,
      "tokens_seen": 2867200000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001325175526579739,
      "loss": 1.3239,
      "theoretical_loss": 3.3608584236848893,
      "tokens_seen": 2867265536
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013249749247743232,
      "loss": 1.2941,
      "theoretical_loss": 3.360852513550161,
      "tokens_seen": 2867331072
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013247743229689068,
      "loss": 1.3941,
      "theoretical_loss": 3.360846603588336,
      "tokens_seen": 2867396608
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013245737211634907,
      "loss": 1.2477,
      "theoretical_loss": 3.360840693799405,
      "tokens_seen": 2867462144
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013243731193580743,
      "loss": 1.208,
      "theoretical_loss": 3.3608347841833597,
      "tokens_seen": 2867527680
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013241725175526582,
      "loss": 1.2796,
      "theoretical_loss": 3.36082887474019,
      "tokens_seen": 2867593216
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013239719157472418,
      "loss": 1.1734,
      "theoretical_loss": 3.360822965469888,
      "tokens_seen": 2867658752
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013237713139418257,
      "loss": 1.3543,
      "theoretical_loss": 3.3608170563724435,
      "tokens_seen": 2867724288
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013235707121364093,
      "loss": 1.1779,
      "theoretical_loss": 3.3608111474478486,
      "tokens_seen": 2867789824
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001323370110330993,
      "loss": 1.2838,
      "theoretical_loss": 3.360805238696094,
      "tokens_seen": 2867855360
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013231695085255769,
      "loss": 1.382,
      "theoretical_loss": 3.3607993301171706,
      "tokens_seen": 2867920896
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013229689067201605,
      "loss": 1.3273,
      "theoretical_loss": 3.360793421711069,
      "tokens_seen": 2867986432
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013227683049147444,
      "loss": 1.2484,
      "theoretical_loss": 3.360787513477781,
      "tokens_seen": 2868051968
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001322567703109328,
      "loss": 1.3204,
      "theoretical_loss": 3.360781605417297,
      "tokens_seen": 2868117504
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001322367101303912,
      "loss": 1.2832,
      "theoretical_loss": 3.3607756975296077,
      "tokens_seen": 2868183040
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013221664994984955,
      "loss": 1.3164,
      "theoretical_loss": 3.360769789814705,
      "tokens_seen": 2868248576
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013219658976930794,
      "loss": 1.3658,
      "theoretical_loss": 3.3607638822725794,
      "tokens_seen": 2868314112
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001321765295887663,
      "loss": 1.2019,
      "theoretical_loss": 3.3607579749032217,
      "tokens_seen": 2868379648
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013215646940822467,
      "loss": 1.3563,
      "theoretical_loss": 3.360752067706623,
      "tokens_seen": 2868445184
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013213640922768305,
      "loss": 1.2468,
      "theoretical_loss": 3.3607461606827753,
      "tokens_seen": 2868510720
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013211634904714142,
      "loss": 1.2134,
      "theoretical_loss": 3.360740253831668,
      "tokens_seen": 2868576256
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001320962888665998,
      "loss": 1.2556,
      "theoretical_loss": 3.360734347153293,
      "tokens_seen": 2868641792
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013207622868605817,
      "loss": 1.3016,
      "theoretical_loss": 3.360728440647641,
      "tokens_seen": 2868707328
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013205616850551656,
      "loss": 1.3514,
      "theoretical_loss": 3.3607225343147036,
      "tokens_seen": 2868772864
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013203610832497492,
      "loss": 1.3265,
      "theoretical_loss": 3.360716628154471,
      "tokens_seen": 2868838400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001320160481444333,
      "loss": 1.3533,
      "theoretical_loss": 3.360710722166935,
      "tokens_seen": 2868903936
    },
    {
      "epoch": 0.74,
      "objective/train/docs_used": 1042847,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3076987266540527,
      "objective/train/theoretical_loss": 3.3607048163520856,
      "objective/train/tokens_used": 1239428576,
      "theoretical_loss": 3.3607048163520856,
      "tokens_seen": 2868969472
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013199598796389167,
      "loss": 1.3181,
      "theoretical_loss": 3.3607048163520856,
      "tokens_seen": 2868969472
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013197592778335003,
      "loss": 1.4183,
      "theoretical_loss": 3.3606989107099148,
      "tokens_seen": 2869035008
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013195586760280842,
      "loss": 1.201,
      "theoretical_loss": 3.360693005240413,
      "tokens_seen": 2869100544
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013193580742226679,
      "loss": 1.3155,
      "theoretical_loss": 3.360687099943571,
      "tokens_seen": 2869166080
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013191574724172517,
      "loss": 1.2559,
      "theoretical_loss": 3.3606811948193807,
      "tokens_seen": 2869231616
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013189568706118354,
      "loss": 1.2662,
      "theoretical_loss": 3.3606752898678325,
      "tokens_seen": 2869297152
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013187562688064193,
      "loss": 1.2727,
      "theoretical_loss": 3.3606693850889178,
      "tokens_seen": 2869362688
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001318555667001003,
      "loss": 1.298,
      "theoretical_loss": 3.3606634804826268,
      "tokens_seen": 2869428224
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013183550651955868,
      "loss": 1.2553,
      "theoretical_loss": 3.3606575760489514,
      "tokens_seen": 2869493760
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013181544633901704,
      "loss": 1.225,
      "theoretical_loss": 3.360651671787882,
      "tokens_seen": 2869559296
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013179538615847543,
      "loss": 1.3704,
      "theoretical_loss": 3.36064576769941,
      "tokens_seen": 2869624832
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001317753259779338,
      "loss": 1.3302,
      "theoretical_loss": 3.3606398637835264,
      "tokens_seen": 2869690368
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013175526579739218,
      "loss": 1.2347,
      "theoretical_loss": 3.360633960040222,
      "tokens_seen": 2869755904
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013173520561685057,
      "loss": 1.3212,
      "theoretical_loss": 3.360628056469488,
      "tokens_seen": 2869821440
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013171514543630893,
      "loss": 1.2598,
      "theoretical_loss": 3.360622153071315,
      "tokens_seen": 2869886976
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013169508525576732,
      "loss": 1.242,
      "theoretical_loss": 3.3606162498456946,
      "tokens_seen": 2869952512
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013167502507522568,
      "loss": 1.242,
      "theoretical_loss": 3.360610346792617,
      "tokens_seen": 2870018048
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013165496489468407,
      "loss": 1.2927,
      "theoretical_loss": 3.3606044439120746,
      "tokens_seen": 2870083584
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013163490471414244,
      "loss": 1.336,
      "theoretical_loss": 3.360598541204057,
      "tokens_seen": 2870149120
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013161484453360083,
      "loss": 1.3318,
      "theoretical_loss": 3.3605926386685563,
      "tokens_seen": 2870214656
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001315947843530592,
      "loss": 1.2307,
      "theoretical_loss": 3.360586736305563,
      "tokens_seen": 2870280192
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013157472417251755,
      "loss": 1.2576,
      "theoretical_loss": 3.3605808341150674,
      "tokens_seen": 2870345728
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013155466399197594,
      "loss": 1.2808,
      "theoretical_loss": 3.360574932097062,
      "tokens_seen": 2870411264
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001315346038114343,
      "loss": 1.2832,
      "theoretical_loss": 3.3605690302515367,
      "tokens_seen": 2870476800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001315145436308927,
      "loss": 1.3916,
      "theoretical_loss": 3.360563128578483,
      "tokens_seen": 2870542336
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013149448345035105,
      "loss": 1.3497,
      "theoretical_loss": 3.3605572270778916,
      "tokens_seen": 2870607872
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013147442326980944,
      "loss": 1.3053,
      "theoretical_loss": 3.3605513257497543,
      "tokens_seen": 2870673408
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001314543630892678,
      "loss": 1.261,
      "theoretical_loss": 3.3605454245940614,
      "tokens_seen": 2870738944
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001314343029087262,
      "loss": 1.2618,
      "theoretical_loss": 3.3605395236108038,
      "tokens_seen": 2870804480
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013141424272818456,
      "loss": 1.3406,
      "theoretical_loss": 3.3605336227999727,
      "tokens_seen": 2870870016
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013139418254764295,
      "loss": 1.1237,
      "theoretical_loss": 3.3605277221615597,
      "tokens_seen": 2870935552
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001313741223671013,
      "loss": 1.3064,
      "theoretical_loss": 3.3605218216955555,
      "tokens_seen": 2871001088
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013135406218655967,
      "loss": 1.1355,
      "theoretical_loss": 3.3605159214019507,
      "tokens_seen": 2871066624
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013133400200601806,
      "loss": 1.263,
      "theoretical_loss": 3.3605100212807364,
      "tokens_seen": 2871132160
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013131394182547642,
      "loss": 1.3038,
      "theoretical_loss": 3.3605041213319042,
      "tokens_seen": 2871197696
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001312938816449348,
      "loss": 1.3152,
      "theoretical_loss": 3.360498221555445,
      "tokens_seen": 2871263232
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013127382146439317,
      "loss": 1.3616,
      "theoretical_loss": 3.360492321951349,
      "tokens_seen": 2871328768
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013125376128385156,
      "loss": 1.2023,
      "theoretical_loss": 3.360486422519608,
      "tokens_seen": 2871394304
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013123370110330992,
      "loss": 1.2863,
      "theoretical_loss": 3.3604805232602133,
      "tokens_seen": 2871459840
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013121364092276831,
      "loss": 1.2776,
      "theoretical_loss": 3.360474624173155,
      "tokens_seen": 2871525376
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013119358074222668,
      "loss": 1.2298,
      "theoretical_loss": 3.360468725258425,
      "tokens_seen": 2871590912
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013117352056168504,
      "loss": 1.2521,
      "theoretical_loss": 3.360462826516014,
      "tokens_seen": 2871656448
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013115346038114343,
      "loss": 1.1979,
      "theoretical_loss": 3.360456927945913,
      "tokens_seen": 2871721984
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001311334002006018,
      "loss": 1.2906,
      "theoretical_loss": 3.3604510295481127,
      "tokens_seen": 2871787520
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013111334002006018,
      "loss": 1.3014,
      "theoretical_loss": 3.3604451313226047,
      "tokens_seen": 2871853056
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013109327983951854,
      "loss": 1.2654,
      "theoretical_loss": 3.3604392332693798,
      "tokens_seen": 2871918592
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013107321965897693,
      "loss": 1.3539,
      "theoretical_loss": 3.360433335388429,
      "tokens_seen": 2871984128
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001310531594784353,
      "loss": 1.307,
      "theoretical_loss": 3.3604274376797436,
      "tokens_seen": 2872049664
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013103309929789368,
      "loss": 1.2816,
      "theoretical_loss": 3.360421540143314,
      "tokens_seen": 2872115200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013101303911735204,
      "loss": 1.2879,
      "theoretical_loss": 3.3604156427791323,
      "tokens_seen": 2872180736
    },
    {
      "epoch": 0.74,
      "objective/train/docs_used": 1044056,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.30495285987854,
      "objective/train/theoretical_loss": 3.360409745587188,
      "objective/train/tokens_used": 1242705376,
      "theoretical_loss": 3.360409745587188,
      "tokens_seen": 2872246272
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013099297893681043,
      "loss": 1.2925,
      "theoretical_loss": 3.360409745587188,
      "tokens_seen": 2872246272
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013097291875626882,
      "loss": 1.2028,
      "theoretical_loss": 3.360403848567474,
      "tokens_seen": 2872311808
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013095285857572719,
      "loss": 1.2174,
      "theoretical_loss": 3.36039795171998,
      "tokens_seen": 2872377344
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013093279839518558,
      "loss": 1.2567,
      "theoretical_loss": 3.3603920550446977,
      "tokens_seen": 2872442880
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013091273821464394,
      "loss": 1.2407,
      "theoretical_loss": 3.3603861585416173,
      "tokens_seen": 2872508416
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013089267803410233,
      "loss": 1.2635,
      "theoretical_loss": 3.3603802622107306,
      "tokens_seen": 2872573952
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001308726178535607,
      "loss": 1.1757,
      "theoretical_loss": 3.360374366052029,
      "tokens_seen": 2872639488
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013085255767301908,
      "loss": 1.2838,
      "theoretical_loss": 3.3603684700655023,
      "tokens_seen": 2872705024
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013083249749247744,
      "loss": 1.3039,
      "theoretical_loss": 3.3603625742511425,
      "tokens_seen": 2872770560
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013081243731193583,
      "loss": 1.306,
      "theoretical_loss": 3.36035667860894,
      "tokens_seen": 2872836096
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001307923771313942,
      "loss": 1.2423,
      "theoretical_loss": 3.3603507831388866,
      "tokens_seen": 2872901632
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013077231695085255,
      "loss": 1.2316,
      "theoretical_loss": 3.360344887840973,
      "tokens_seen": 2872967168
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013075225677031094,
      "loss": 1.2444,
      "theoretical_loss": 3.3603389927151905,
      "tokens_seen": 2873032704
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001307321965897693,
      "loss": 1.2335,
      "theoretical_loss": 3.3603330977615293,
      "tokens_seen": 2873098240
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001307121364092277,
      "loss": 1.3172,
      "theoretical_loss": 3.3603272029799816,
      "tokens_seen": 2873163776
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013069207622868606,
      "loss": 1.2737,
      "theoretical_loss": 3.3603213083705374,
      "tokens_seen": 2873229312
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013067201604814445,
      "loss": 1.2664,
      "theoretical_loss": 3.3603154139331886,
      "tokens_seen": 2873294848
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001306519558676028,
      "loss": 1.1685,
      "theoretical_loss": 3.3603095196679256,
      "tokens_seen": 2873360384
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001306318956870612,
      "loss": 1.2288,
      "theoretical_loss": 3.3603036255747396,
      "tokens_seen": 2873425920
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013061183550651956,
      "loss": 1.3427,
      "theoretical_loss": 3.360297731653622,
      "tokens_seen": 2873491456
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013059177532597792,
      "loss": 1.3497,
      "theoretical_loss": 3.3602918379045636,
      "tokens_seen": 2873556992
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001305717151454363,
      "loss": 1.2567,
      "theoretical_loss": 3.3602859443275555,
      "tokens_seen": 2873622528
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013055165496489467,
      "loss": 1.382,
      "theoretical_loss": 3.3602800509225887,
      "tokens_seen": 2873688064
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013053159478435306,
      "loss": 1.2403,
      "theoretical_loss": 3.360274157689654,
      "tokens_seen": 2873753600
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013051153460381143,
      "loss": 1.2919,
      "theoretical_loss": 3.3602682646287434,
      "tokens_seen": 2873819136
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013049147442326982,
      "loss": 1.21,
      "theoretical_loss": 3.360262371739847,
      "tokens_seen": 2873884672
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013047141424272818,
      "loss": 1.2831,
      "theoretical_loss": 3.360256479022956,
      "tokens_seen": 2873950208
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013045135406218657,
      "loss": 1.3281,
      "theoretical_loss": 3.360250586478062,
      "tokens_seen": 2874015744
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013043129388164493,
      "loss": 1.3184,
      "theoretical_loss": 3.3602446941051554,
      "tokens_seen": 2874081280
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001304112337011033,
      "loss": 1.2069,
      "theoretical_loss": 3.3602388019042273,
      "tokens_seen": 2874146816
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013039117352056168,
      "loss": 1.2442,
      "theoretical_loss": 3.3602329098752692,
      "tokens_seen": 2874212352
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013037111334002004,
      "loss": 1.3352,
      "theoretical_loss": 3.360227018018272,
      "tokens_seen": 2874277888
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013035105315947843,
      "loss": 1.2521,
      "theoretical_loss": 3.3602211263332267,
      "tokens_seen": 2874343424
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001303309929789368,
      "loss": 1.3301,
      "theoretical_loss": 3.3602152348201244,
      "tokens_seen": 2874408960
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013031093279839518,
      "loss": 1.282,
      "theoretical_loss": 3.3602093434789557,
      "tokens_seen": 2874474496
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013029087261785355,
      "loss": 1.1713,
      "theoretical_loss": 3.3602034523097126,
      "tokens_seen": 2874540032
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013027081243731194,
      "loss": 1.21,
      "theoretical_loss": 3.3601975613123853,
      "tokens_seen": 2874605568
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013025075225677033,
      "loss": 1.1898,
      "theoretical_loss": 3.360191670486966,
      "tokens_seen": 2874671104
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013023069207622871,
      "loss": 1.3336,
      "theoretical_loss": 3.3601857798334445,
      "tokens_seen": 2874736640
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013021063189568708,
      "loss": 1.272,
      "theoretical_loss": 3.360179889351812,
      "tokens_seen": 2874802176
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013019057171514544,
      "loss": 1.3136,
      "theoretical_loss": 3.3601739990420603,
      "tokens_seen": 2874867712
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013017051153460383,
      "loss": 1.3086,
      "theoretical_loss": 3.36016810890418,
      "tokens_seen": 2874933248
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001301504513540622,
      "loss": 1.2509,
      "theoretical_loss": 3.360162218938162,
      "tokens_seen": 2874998784
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013013039117352058,
      "loss": 1.4024,
      "theoretical_loss": 3.360156329143998,
      "tokens_seen": 2875064320
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013011033099297894,
      "loss": 1.3123,
      "theoretical_loss": 3.3601504395216786,
      "tokens_seen": 2875129856
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013009027081243733,
      "loss": 1.4079,
      "theoretical_loss": 3.360144550071195,
      "tokens_seen": 2875195392
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001300702106318957,
      "loss": 1.297,
      "theoretical_loss": 3.360138660792538,
      "tokens_seen": 2875260928
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013005015045135408,
      "loss": 1.4151,
      "theoretical_loss": 3.3601327716856995,
      "tokens_seen": 2875326464
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013003009027081245,
      "loss": 1.2381,
      "theoretical_loss": 3.360126882750669,
      "tokens_seen": 2875392000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001300100300902708,
      "loss": 1.2952,
      "theoretical_loss": 3.3601209939874392,
      "tokens_seen": 2875457536
    },
    {
      "epoch": 0.74,
      "objective/train/docs_used": 1045259,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1272075176239014,
      "objective/train/theoretical_loss": 3.360115105396001,
      "objective/train/tokens_used": 1245982176,
      "theoretical_loss": 3.360115105396001,
      "tokens_seen": 2875523072
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001299899699097292,
      "loss": 1.1347,
      "theoretical_loss": 3.360115105396001,
      "tokens_seen": 2875523072
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012996990972918756,
      "loss": 1.325,
      "theoretical_loss": 3.360109216976344,
      "tokens_seen": 2875588608
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012994984954864595,
      "loss": 1.2679,
      "theoretical_loss": 3.3601033287284605,
      "tokens_seen": 2875654144
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001299297893681043,
      "loss": 1.3538,
      "theoretical_loss": 3.360097440652342,
      "tokens_seen": 2875719680
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001299097291875627,
      "loss": 1.2612,
      "theoretical_loss": 3.3600915527479778,
      "tokens_seen": 2875785216
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012988966900702106,
      "loss": 1.2171,
      "theoretical_loss": 3.3600856650153608,
      "tokens_seen": 2875850752
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012986960882647945,
      "loss": 1.3469,
      "theoretical_loss": 3.360079777454481,
      "tokens_seen": 2875916288
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012984954864593781,
      "loss": 1.3091,
      "theoretical_loss": 3.3600738900653306,
      "tokens_seen": 2875981824
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012982948846539618,
      "loss": 1.2796,
      "theoretical_loss": 3.3600680028478993,
      "tokens_seen": 2876047360
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012980942828485457,
      "loss": 1.3176,
      "theoretical_loss": 3.3600621158021786,
      "tokens_seen": 2876112896
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012978936810431293,
      "loss": 1.3179,
      "theoretical_loss": 3.36005622892816,
      "tokens_seen": 2876178432
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012976930792377132,
      "loss": 1.3048,
      "theoretical_loss": 3.3600503422258345,
      "tokens_seen": 2876243968
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012974924774322968,
      "loss": 1.3129,
      "theoretical_loss": 3.360044455695193,
      "tokens_seen": 2876309504
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012972918756268807,
      "loss": 1.2316,
      "theoretical_loss": 3.360038569336227,
      "tokens_seen": 2876375040
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012970912738214643,
      "loss": 1.3501,
      "theoretical_loss": 3.3600326831489267,
      "tokens_seen": 2876440576
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012968906720160482,
      "loss": 1.3354,
      "theoretical_loss": 3.3600267971332833,
      "tokens_seen": 2876506112
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012966900702106318,
      "loss": 1.3307,
      "theoretical_loss": 3.360020911289289,
      "tokens_seen": 2876571648
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012964894684052157,
      "loss": 1.2026,
      "theoretical_loss": 3.360015025616934,
      "tokens_seen": 2876637184
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012962888665997993,
      "loss": 1.2982,
      "theoretical_loss": 3.3600091401162095,
      "tokens_seen": 2876702720
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001296088264794383,
      "loss": 1.3225,
      "theoretical_loss": 3.360003254787107,
      "tokens_seen": 2876768256
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012958876629889669,
      "loss": 1.2878,
      "theoretical_loss": 3.3599973696296166,
      "tokens_seen": 2876833792
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012956870611835505,
      "loss": 1.2253,
      "theoretical_loss": 3.35999148464373,
      "tokens_seen": 2876899328
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012954864593781344,
      "loss": 1.3596,
      "theoretical_loss": 3.3599855998294386,
      "tokens_seen": 2876964864
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001295285857572718,
      "loss": 1.3773,
      "theoretical_loss": 3.359979715186733,
      "tokens_seen": 2877030400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001295085255767302,
      "loss": 1.3419,
      "theoretical_loss": 3.3599738307156044,
      "tokens_seen": 2877095936
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012948846539618858,
      "loss": 1.2522,
      "theoretical_loss": 3.359967946416044,
      "tokens_seen": 2877161472
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012946840521564697,
      "loss": 1.2733,
      "theoretical_loss": 3.3599620622880426,
      "tokens_seen": 2877227008
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012944834503510533,
      "loss": 1.3679,
      "theoretical_loss": 3.3599561783315917,
      "tokens_seen": 2877292544
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001294282848545637,
      "loss": 1.2515,
      "theoretical_loss": 3.3599502945466826,
      "tokens_seen": 2877358080
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012940822467402208,
      "loss": 1.2765,
      "theoretical_loss": 3.3599444109333056,
      "tokens_seen": 2877423616
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012938816449348044,
      "loss": 1.2555,
      "theoretical_loss": 3.3599385274914524,
      "tokens_seen": 2877489152
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012936810431293883,
      "loss": 1.2391,
      "theoretical_loss": 3.359932644221114,
      "tokens_seen": 2877554688
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001293480441323972,
      "loss": 1.3326,
      "theoretical_loss": 3.359926761122281,
      "tokens_seen": 2877620224
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012932798395185558,
      "loss": 1.2456,
      "theoretical_loss": 3.359920878194945,
      "tokens_seen": 2877685760
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012930792377131395,
      "loss": 1.3433,
      "theoretical_loss": 3.359914995439097,
      "tokens_seen": 2877751296
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012928786359077234,
      "loss": 1.2622,
      "theoretical_loss": 3.3599091128547283,
      "tokens_seen": 2877816832
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001292678034102307,
      "loss": 1.2421,
      "theoretical_loss": 3.3599032304418297,
      "tokens_seen": 2877882368
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001292477432296891,
      "loss": 1.3695,
      "theoretical_loss": 3.359897348200392,
      "tokens_seen": 2877947904
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012922768304914745,
      "loss": 1.2896,
      "theoretical_loss": 3.3598914661304073,
      "tokens_seen": 2878013440
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001292076228686058,
      "loss": 1.3571,
      "theoretical_loss": 3.359885584231866,
      "tokens_seen": 2878078976
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001291875626880642,
      "loss": 1.2737,
      "theoretical_loss": 3.3598797025047586,
      "tokens_seen": 2878144512
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012916750250752256,
      "loss": 1.2156,
      "theoretical_loss": 3.3598738209490775,
      "tokens_seen": 2878210048
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012914744232698095,
      "loss": 1.4239,
      "theoretical_loss": 3.359867939564813,
      "tokens_seen": 2878275584
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012912738214643932,
      "loss": 1.3553,
      "theoretical_loss": 3.359862058351956,
      "tokens_seen": 2878341120
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001291073219658977,
      "loss": 1.2511,
      "theoretical_loss": 3.3598561773104985,
      "tokens_seen": 2878406656
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012908726178535607,
      "loss": 1.239,
      "theoretical_loss": 3.3598502964404307,
      "tokens_seen": 2878472192
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012906720160481446,
      "loss": 1.2741,
      "theoretical_loss": 3.3598444157417444,
      "tokens_seen": 2878537728
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012904714142427282,
      "loss": 1.2905,
      "theoretical_loss": 3.35983853521443,
      "tokens_seen": 2878603264
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012902708124373118,
      "loss": 1.3805,
      "theoretical_loss": 3.3598326548584794,
      "tokens_seen": 2878668800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012900702106318957,
      "loss": 1.3362,
      "theoretical_loss": 3.359826774673883,
      "tokens_seen": 2878734336
    },
    {
      "epoch": 0.74,
      "objective/train/docs_used": 1046332,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2998217344284058,
      "objective/train/theoretical_loss": 3.3598208946606323,
      "objective/train/tokens_used": 1249258976,
      "theoretical_loss": 3.3598208946606323,
      "tokens_seen": 2878799872
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012898696088264793,
      "loss": 1.2947,
      "theoretical_loss": 3.3598208946606323,
      "tokens_seen": 2878799872
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012896690070210632,
      "loss": 1.1581,
      "theoretical_loss": 3.359815014818718,
      "tokens_seen": 2878865408
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012894684052156468,
      "loss": 1.2567,
      "theoretical_loss": 3.359809135148132,
      "tokens_seen": 2878930944
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012892678034102307,
      "loss": 1.1676,
      "theoretical_loss": 3.3598032556488646,
      "tokens_seen": 2878996480
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012890672016048144,
      "loss": 1.3338,
      "theoretical_loss": 3.3597973763209072,
      "tokens_seen": 2879062016
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012888665997993982,
      "loss": 1.3064,
      "theoretical_loss": 3.3597914971642515,
      "tokens_seen": 2879127552
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001288665997993982,
      "loss": 1.3314,
      "theoretical_loss": 3.359785618178887,
      "tokens_seen": 2879193088
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012884653961885655,
      "loss": 1.3627,
      "theoretical_loss": 3.359779739364807,
      "tokens_seen": 2879258624
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012882647943831494,
      "loss": 1.1772,
      "theoretical_loss": 3.359773860722001,
      "tokens_seen": 2879324160
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001288064192577733,
      "loss": 1.2653,
      "theoretical_loss": 3.3597679822504602,
      "tokens_seen": 2879389696
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001287863590772317,
      "loss": 1.3859,
      "theoretical_loss": 3.3597621039501764,
      "tokens_seen": 2879455232
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012876629889669005,
      "loss": 1.1782,
      "theoretical_loss": 3.3597562258211404,
      "tokens_seen": 2879520768
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012874623871614847,
      "loss": 1.3524,
      "theoretical_loss": 3.3597503478633435,
      "tokens_seen": 2879586304
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012872617853560683,
      "loss": 1.2798,
      "theoretical_loss": 3.3597444700767762,
      "tokens_seen": 2879651840
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012870611835506522,
      "loss": 1.1491,
      "theoretical_loss": 3.3597385924614303,
      "tokens_seen": 2879717376
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012868605817452358,
      "loss": 1.2291,
      "theoretical_loss": 3.3597327150172966,
      "tokens_seen": 2879782912
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012866599799398197,
      "loss": 1.3303,
      "theoretical_loss": 3.3597268377443665,
      "tokens_seen": 2879848448
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012864593781344033,
      "loss": 1.2701,
      "theoretical_loss": 3.3597209606426306,
      "tokens_seen": 2879913984
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001286258776328987,
      "loss": 1.3429,
      "theoretical_loss": 3.3597150837120804,
      "tokens_seen": 2879979520
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012860581745235709,
      "loss": 1.3296,
      "theoretical_loss": 3.359709206952707,
      "tokens_seen": 2880045056
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012858575727181545,
      "loss": 1.2772,
      "theoretical_loss": 3.3597033303645016,
      "tokens_seen": 2880110592
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012856569709127384,
      "loss": 1.2677,
      "theoretical_loss": 3.3596974539474553,
      "tokens_seen": 2880176128
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001285456369107322,
      "loss": 1.2368,
      "theoretical_loss": 3.3596915777015584,
      "tokens_seen": 2880241664
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001285255767301906,
      "loss": 1.2463,
      "theoretical_loss": 3.3596857016268036,
      "tokens_seen": 2880307200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012850551654964895,
      "loss": 1.3352,
      "theoretical_loss": 3.3596798257231804,
      "tokens_seen": 2880372736
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012848545636910734,
      "loss": 1.3348,
      "theoretical_loss": 3.3596739499906807,
      "tokens_seen": 2880438272
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001284653961885657,
      "loss": 1.2882,
      "theoretical_loss": 3.3596680744292957,
      "tokens_seen": 2880503808
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012844533600802407,
      "loss": 1.2548,
      "theoretical_loss": 3.3596621990390165,
      "tokens_seen": 2880569344
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012842527582748245,
      "loss": 1.3124,
      "theoretical_loss": 3.3596563238198343,
      "tokens_seen": 2880634880
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012840521564694082,
      "loss": 1.1857,
      "theoretical_loss": 3.35965044877174,
      "tokens_seen": 2880700416
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001283851554663992,
      "loss": 1.2723,
      "theoretical_loss": 3.359644573894725,
      "tokens_seen": 2880765952
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012836509528585757,
      "loss": 1.176,
      "theoretical_loss": 3.3596386991887797,
      "tokens_seen": 2880831488
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012834503510531596,
      "loss": 1.3352,
      "theoretical_loss": 3.359632824653896,
      "tokens_seen": 2880897024
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012832497492477432,
      "loss": 1.1698,
      "theoretical_loss": 3.3596269502900644,
      "tokens_seen": 2880962560
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001283049147442327,
      "loss": 1.2782,
      "theoretical_loss": 3.3596210760972767,
      "tokens_seen": 2881028096
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012828485456369107,
      "loss": 1.2754,
      "theoretical_loss": 3.3596152020755237,
      "tokens_seen": 2881093632
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012826479438314943,
      "loss": 1.2992,
      "theoretical_loss": 3.3596093282247965,
      "tokens_seen": 2881159168
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012824473420260782,
      "loss": 1.2769,
      "theoretical_loss": 3.3596034545450864,
      "tokens_seen": 2881224704
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012822467402206619,
      "loss": 1.376,
      "theoretical_loss": 3.3595975810363843,
      "tokens_seen": 2881290240
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012820461384152457,
      "loss": 1.3276,
      "theoretical_loss": 3.3595917076986814,
      "tokens_seen": 2881355776
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012818455366098294,
      "loss": 1.2317,
      "theoretical_loss": 3.359585834531969,
      "tokens_seen": 2881421312
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012816449348044133,
      "loss": 1.2515,
      "theoretical_loss": 3.359579961536238,
      "tokens_seen": 2881486848
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001281444332998997,
      "loss": 1.311,
      "theoretical_loss": 3.3595740887114798,
      "tokens_seen": 2881552384
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012812437311935808,
      "loss": 1.3302,
      "theoretical_loss": 3.359568216057685,
      "tokens_seen": 2881617920
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012810431293881644,
      "loss": 1.2785,
      "theoretical_loss": 3.3595623435748454,
      "tokens_seen": 2881683456
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012808425275827483,
      "loss": 1.1957,
      "theoretical_loss": 3.3595564712629518,
      "tokens_seen": 2881748992
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001280641925777332,
      "loss": 1.3127,
      "theoretical_loss": 3.3595505991219956,
      "tokens_seen": 2881814528
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012804413239719155,
      "loss": 1.3311,
      "theoretical_loss": 3.3595447271519676,
      "tokens_seen": 2881880064
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012802407221664994,
      "loss": 1.237,
      "theoretical_loss": 3.359538855352859,
      "tokens_seen": 2881945600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001280040120361083,
      "loss": 1.1972,
      "theoretical_loss": 3.3595329837246606,
      "tokens_seen": 2882011136
    },
    {
      "epoch": 0.75,
      "objective/train/docs_used": 1047629,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.259054183959961,
      "objective/train/theoretical_loss": 3.3595271122673642,
      "objective/train/tokens_used": 1252535776,
      "theoretical_loss": 3.3595271122673642,
      "tokens_seen": 2882076672
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012798395185556672,
      "loss": 1.2636,
      "theoretical_loss": 3.3595271122673642,
      "tokens_seen": 2882076672
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012796389167502508,
      "loss": 1.2657,
      "theoretical_loss": 3.359521240980961,
      "tokens_seen": 2882142208
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012794383149448347,
      "loss": 1.2269,
      "theoretical_loss": 3.359515369865442,
      "tokens_seen": 2882207744
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012792377131394184,
      "loss": 1.3754,
      "theoretical_loss": 3.359509498920797,
      "tokens_seen": 2882273280
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012790371113340023,
      "loss": 1.3001,
      "theoretical_loss": 3.3595036281470194,
      "tokens_seen": 2882338816
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001278836509528586,
      "loss": 1.1938,
      "theoretical_loss": 3.3594977575440987,
      "tokens_seen": 2882404352
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012786359077231695,
      "loss": 1.1626,
      "theoretical_loss": 3.359491887112027,
      "tokens_seen": 2882469888
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012784353059177534,
      "loss": 1.3039,
      "theoretical_loss": 3.359486016850795,
      "tokens_seen": 2882535424
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001278234704112337,
      "loss": 1.2614,
      "theoretical_loss": 3.3594801467603936,
      "tokens_seen": 2882600960
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001278034102306921,
      "loss": 1.3059,
      "theoretical_loss": 3.3594742768408143,
      "tokens_seen": 2882666496
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012778335005015045,
      "loss": 1.2943,
      "theoretical_loss": 3.359468407092048,
      "tokens_seen": 2882732032
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012776328986960884,
      "loss": 1.3141,
      "theoretical_loss": 3.3594625375140863,
      "tokens_seen": 2882797568
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001277432296890672,
      "loss": 1.354,
      "theoretical_loss": 3.35945666810692,
      "tokens_seen": 2882863104
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001277231695085256,
      "loss": 1.3261,
      "theoretical_loss": 3.35945079887054,
      "tokens_seen": 2882928640
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012770310932798396,
      "loss": 1.2708,
      "theoretical_loss": 3.359444929804938,
      "tokens_seen": 2882994176
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012768304914744232,
      "loss": 1.2833,
      "theoretical_loss": 3.359439060910105,
      "tokens_seen": 2883059712
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001276629889669007,
      "loss": 1.273,
      "theoretical_loss": 3.359433192186032,
      "tokens_seen": 2883125248
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012764292878635907,
      "loss": 1.3292,
      "theoretical_loss": 3.35942732363271,
      "tokens_seen": 2883190784
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012762286860581746,
      "loss": 1.2661,
      "theoretical_loss": 3.3594214552501307,
      "tokens_seen": 2883256320
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012760280842527582,
      "loss": 1.2378,
      "theoretical_loss": 3.3594155870382845,
      "tokens_seen": 2883321856
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001275827482447342,
      "loss": 1.1999,
      "theoretical_loss": 3.359409718997163,
      "tokens_seen": 2883387392
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012756268806419257,
      "loss": 1.3329,
      "theoretical_loss": 3.359403851126758,
      "tokens_seen": 2883452928
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012754262788365096,
      "loss": 1.3805,
      "theoretical_loss": 3.359397983427059,
      "tokens_seen": 2883518464
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012752256770310932,
      "loss": 1.2621,
      "theoretical_loss": 3.3593921158980584,
      "tokens_seen": 2883584000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012750250752256771,
      "loss": 1.2705,
      "theoretical_loss": 3.359386248539747,
      "tokens_seen": 2883649536
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012748244734202608,
      "loss": 1.2912,
      "theoretical_loss": 3.3593803813521164,
      "tokens_seen": 2883715072
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012746238716148444,
      "loss": 1.3411,
      "theoretical_loss": 3.3593745143351574,
      "tokens_seen": 2883780608
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012744232698094283,
      "loss": 1.1657,
      "theoretical_loss": 3.3593686474888607,
      "tokens_seen": 2883846144
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001274222668004012,
      "loss": 1.3279,
      "theoretical_loss": 3.359362780813218,
      "tokens_seen": 2883911680
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012740220661985958,
      "loss": 1.2878,
      "theoretical_loss": 3.3593569143082207,
      "tokens_seen": 2883977216
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012738214643931794,
      "loss": 1.2543,
      "theoretical_loss": 3.3593510479738597,
      "tokens_seen": 2884042752
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012736208625877633,
      "loss": 1.3126,
      "theoretical_loss": 3.3593451818101254,
      "tokens_seen": 2884108288
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001273420260782347,
      "loss": 1.2738,
      "theoretical_loss": 3.35933931581701,
      "tokens_seen": 2884173824
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012732196589769308,
      "loss": 1.3123,
      "theoretical_loss": 3.3593334499945042,
      "tokens_seen": 2884239360
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012730190571715144,
      "loss": 1.3151,
      "theoretical_loss": 3.3593275843425996,
      "tokens_seen": 2884304896
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001272818455366098,
      "loss": 1.2079,
      "theoretical_loss": 3.3593217188612865,
      "tokens_seen": 2884370432
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001272617853560682,
      "loss": 1.289,
      "theoretical_loss": 3.359315853550557,
      "tokens_seen": 2884435968
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012724172517552659,
      "loss": 1.25,
      "theoretical_loss": 3.3593099884104016,
      "tokens_seen": 2884501504
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012722166499498497,
      "loss": 1.2274,
      "theoretical_loss": 3.359304123440812,
      "tokens_seen": 2884567040
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012720160481444334,
      "loss": 1.2939,
      "theoretical_loss": 3.359298258641779,
      "tokens_seen": 2884632576
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012718154463390173,
      "loss": 1.2804,
      "theoretical_loss": 3.3592923940132935,
      "tokens_seen": 2884698112
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001271614844533601,
      "loss": 1.3013,
      "theoretical_loss": 3.359286529555347,
      "tokens_seen": 2884763648
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012714142427281848,
      "loss": 1.2153,
      "theoretical_loss": 3.359280665267931,
      "tokens_seen": 2884829184
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012712136409227684,
      "loss": 1.3348,
      "theoretical_loss": 3.3592748011510363,
      "tokens_seen": 2884894720
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012710130391173523,
      "loss": 1.177,
      "theoretical_loss": 3.359268937204654,
      "tokens_seen": 2884960256
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001270812437311936,
      "loss": 1.1492,
      "theoretical_loss": 3.3592630734287754,
      "tokens_seen": 2885025792
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012706118355065195,
      "loss": 1.1847,
      "theoretical_loss": 3.359257209823392,
      "tokens_seen": 2885091328
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012704112337011034,
      "loss": 1.2752,
      "theoretical_loss": 3.359251346388494,
      "tokens_seen": 2885156864
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001270210631895687,
      "loss": 1.2747,
      "theoretical_loss": 3.3592454831240732,
      "tokens_seen": 2885222400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001270010030090271,
      "loss": 1.3102,
      "theoretical_loss": 3.359239620030121,
      "tokens_seen": 2885287936
    },
    {
      "epoch": 0.75,
      "objective/train/docs_used": 1048861,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4479341506958008,
      "objective/train/theoretical_loss": 3.3592337571066286,
      "objective/train/tokens_used": 1255812576,
      "theoretical_loss": 3.3592337571066286,
      "tokens_seen": 2885353472
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012698094282848546,
      "loss": 1.3971,
      "theoretical_loss": 3.3592337571066286,
      "tokens_seen": 2885353472
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012696088264794385,
      "loss": 1.3006,
      "theoretical_loss": 3.3592278943535865,
      "tokens_seen": 2885419008
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001269408224674022,
      "loss": 1.2556,
      "theoretical_loss": 3.3592220317709867,
      "tokens_seen": 2885484544
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001269207622868606,
      "loss": 1.3132,
      "theoretical_loss": 3.3592161693588194,
      "tokens_seen": 2885550080
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012690070210631896,
      "loss": 1.258,
      "theoretical_loss": 3.3592103071170767,
      "tokens_seen": 2885615616
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012688064192577732,
      "loss": 1.1862,
      "theoretical_loss": 3.359204445045749,
      "tokens_seen": 2885681152
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001268605817452357,
      "loss": 1.359,
      "theoretical_loss": 3.359198583144828,
      "tokens_seen": 2885746688
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012684052156469407,
      "loss": 1.3561,
      "theoretical_loss": 3.3591927214143054,
      "tokens_seen": 2885812224
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012682046138415246,
      "loss": 1.2621,
      "theoretical_loss": 3.359186859854171,
      "tokens_seen": 2885877760
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012680040120361083,
      "loss": 1.1891,
      "theoretical_loss": 3.3591809984644168,
      "tokens_seen": 2885943296
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012678034102306922,
      "loss": 1.2165,
      "theoretical_loss": 3.359175137245034,
      "tokens_seen": 2886008832
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012676028084252758,
      "loss": 1.2729,
      "theoretical_loss": 3.3591692761960137,
      "tokens_seen": 2886074368
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012674022066198597,
      "loss": 1.3222,
      "theoretical_loss": 3.359163415317347,
      "tokens_seen": 2886139904
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012672016048144433,
      "loss": 1.174,
      "theoretical_loss": 3.359157554609025,
      "tokens_seen": 2886205440
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001267001003009027,
      "loss": 1.3416,
      "theoretical_loss": 3.3591516940710395,
      "tokens_seen": 2886270976
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012668004012036108,
      "loss": 1.2618,
      "theoretical_loss": 3.359145833703381,
      "tokens_seen": 2886336512
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012665997993981944,
      "loss": 1.3255,
      "theoretical_loss": 3.35913997350604,
      "tokens_seen": 2886402048
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012663991975927783,
      "loss": 1.2621,
      "theoretical_loss": 3.3591341134790094,
      "tokens_seen": 2886467584
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001266198595787362,
      "loss": 1.2575,
      "theoretical_loss": 3.3591282536222797,
      "tokens_seen": 2886533120
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012659979939819458,
      "loss": 1.2627,
      "theoretical_loss": 3.3591223939358414,
      "tokens_seen": 2886598656
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012657973921765295,
      "loss": 1.246,
      "theoretical_loss": 3.3591165344196865,
      "tokens_seen": 2886664192
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012655967903711134,
      "loss": 1.2892,
      "theoretical_loss": 3.3591106750738056,
      "tokens_seen": 2886729728
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001265396188565697,
      "loss": 1.2543,
      "theoretical_loss": 3.3591048158981907,
      "tokens_seen": 2886795264
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001265195586760281,
      "loss": 1.3226,
      "theoretical_loss": 3.359098956892832,
      "tokens_seen": 2886860800
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012649949849548648,
      "loss": 1.3302,
      "theoretical_loss": 3.359093098057721,
      "tokens_seen": 2886926336
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012647943831494484,
      "loss": 1.3453,
      "theoretical_loss": 3.3590872393928493,
      "tokens_seen": 2886991872
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012645937813440323,
      "loss": 1.2373,
      "theoretical_loss": 3.359081380898208,
      "tokens_seen": 2887057408
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001264393179538616,
      "loss": 1.2703,
      "theoretical_loss": 3.359075522573788,
      "tokens_seen": 2887122944
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012641925777331998,
      "loss": 1.2318,
      "theoretical_loss": 3.3590696644195805,
      "tokens_seen": 2887188480
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012639919759277834,
      "loss": 1.2219,
      "theoretical_loss": 3.359063806435577,
      "tokens_seen": 2887254016
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012637913741223673,
      "loss": 1.27,
      "theoretical_loss": 3.3590579486217687,
      "tokens_seen": 2887319552
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001263590772316951,
      "loss": 1.2922,
      "theoretical_loss": 3.359052090978146,
      "tokens_seen": 2887385088
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012633901705115348,
      "loss": 1.2746,
      "theoretical_loss": 3.359046233504701,
      "tokens_seen": 2887450624
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012631895687061185,
      "loss": 1.3157,
      "theoretical_loss": 3.3590403762014245,
      "tokens_seen": 2887516160
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001262988966900702,
      "loss": 1.2406,
      "theoretical_loss": 3.3590345190683077,
      "tokens_seen": 2887581696
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001262788365095286,
      "loss": 1.3354,
      "theoretical_loss": 3.359028662105342,
      "tokens_seen": 2887647232
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012625877632898696,
      "loss": 1.224,
      "theoretical_loss": 3.3590228053125184,
      "tokens_seen": 2887712768
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012623871614844535,
      "loss": 1.2338,
      "theoretical_loss": 3.3590169486898285,
      "tokens_seen": 2887778304
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001262186559679037,
      "loss": 1.2563,
      "theoretical_loss": 3.3590110922372625,
      "tokens_seen": 2887843840
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001261985957873621,
      "loss": 1.3234,
      "theoretical_loss": 3.359005235954813,
      "tokens_seen": 2887909376
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012617853560682046,
      "loss": 1.2603,
      "theoretical_loss": 3.3589993798424693,
      "tokens_seen": 2887974912
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012615847542627885,
      "loss": 1.2571,
      "theoretical_loss": 3.3589935239002244,
      "tokens_seen": 2888040448
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012613841524573721,
      "loss": 1.2066,
      "theoretical_loss": 3.3589876681280693,
      "tokens_seen": 2888105984
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012611835506519558,
      "loss": 1.3657,
      "theoretical_loss": 3.358981812525994,
      "tokens_seen": 2888171520
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012609829488465397,
      "loss": 1.3804,
      "theoretical_loss": 3.358975957093991,
      "tokens_seen": 2888237056
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012607823470411233,
      "loss": 1.2582,
      "theoretical_loss": 3.3589701018320506,
      "tokens_seen": 2888302592
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012605817452357072,
      "loss": 1.2406,
      "theoretical_loss": 3.3589642467401646,
      "tokens_seen": 2888368128
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012603811434302908,
      "loss": 1.2726,
      "theoretical_loss": 3.358958391818324,
      "tokens_seen": 2888433664
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012601805416248747,
      "loss": 1.3003,
      "theoretical_loss": 3.3589525370665196,
      "tokens_seen": 2888499200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012599799398194583,
      "loss": 1.2988,
      "theoretical_loss": 3.358946682484743,
      "tokens_seen": 2888564736
    },
    {
      "epoch": 0.75,
      "objective/train/docs_used": 1050131,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3088003396987915,
      "objective/train/theoretical_loss": 3.3589408280729858,
      "objective/train/tokens_used": 1259089376,
      "theoretical_loss": 3.3589408280729858,
      "tokens_seen": 2888630272
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012597793380140422,
      "loss": 1.351,
      "theoretical_loss": 3.3589408280729858,
      "tokens_seen": 2888630272
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012595787362086258,
      "loss": 1.3932,
      "theoretical_loss": 3.3589349738312384,
      "tokens_seen": 2888695808
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012593781344032097,
      "loss": 1.292,
      "theoretical_loss": 3.3589291197594924,
      "tokens_seen": 2888761344
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012591775325977933,
      "loss": 1.3416,
      "theoretical_loss": 3.3589232658577393,
      "tokens_seen": 2888826880
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001258976930792377,
      "loss": 1.3498,
      "theoretical_loss": 3.3589174121259693,
      "tokens_seen": 2888892416
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012587763289869609,
      "loss": 1.2653,
      "theoretical_loss": 3.358911558564175,
      "tokens_seen": 2888957952
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012585757271815445,
      "loss": 1.2262,
      "theoretical_loss": 3.3589057051723463,
      "tokens_seen": 2889023488
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012583751253761284,
      "loss": 1.2866,
      "theoretical_loss": 3.3588998519504756,
      "tokens_seen": 2889089024
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001258174523570712,
      "loss": 1.3262,
      "theoretical_loss": 3.3588939988985533,
      "tokens_seen": 2889154560
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001257973921765296,
      "loss": 1.358,
      "theoretical_loss": 3.3588881460165707,
      "tokens_seen": 2889220096
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012577733199598795,
      "loss": 1.2262,
      "theoretical_loss": 3.3588822933045193,
      "tokens_seen": 2889285632
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012575727181544634,
      "loss": 1.2981,
      "theoretical_loss": 3.35887644076239,
      "tokens_seen": 2889351168
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012573721163490473,
      "loss": 1.2124,
      "theoretical_loss": 3.3588705883901744,
      "tokens_seen": 2889416704
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001257171514543631,
      "loss": 1.2657,
      "theoretical_loss": 3.358864736187863,
      "tokens_seen": 2889482240
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012569709127382148,
      "loss": 1.2874,
      "theoretical_loss": 3.358858884155448,
      "tokens_seen": 2889547776
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012567703109327984,
      "loss": 1.1942,
      "theoretical_loss": 3.35885303229292,
      "tokens_seen": 2889613312
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012565697091273823,
      "loss": 1.3924,
      "theoretical_loss": 3.3588471806002707,
      "tokens_seen": 2889678848
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001256369107321966,
      "loss": 1.342,
      "theoretical_loss": 3.3588413290774906,
      "tokens_seen": 2889744384
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012561685055165498,
      "loss": 1.1956,
      "theoretical_loss": 3.358835477724571,
      "tokens_seen": 2889809920
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012559679037111335,
      "loss": 1.3283,
      "theoretical_loss": 3.3588296265415036,
      "tokens_seen": 2889875456
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012557673019057174,
      "loss": 1.1962,
      "theoretical_loss": 3.3588237755282795,
      "tokens_seen": 2889940992
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001255566700100301,
      "loss": 1.2736,
      "theoretical_loss": 3.3588179246848897,
      "tokens_seen": 2890006528
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001255366098294885,
      "loss": 1.2906,
      "theoretical_loss": 3.3588120740113254,
      "tokens_seen": 2890072064
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012551654964894685,
      "loss": 1.2743,
      "theoretical_loss": 3.358806223507578,
      "tokens_seen": 2890137600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001254964894684052,
      "loss": 1.2443,
      "theoretical_loss": 3.358800373173639,
      "tokens_seen": 2890203136
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001254764292878636,
      "loss": 1.2607,
      "theoretical_loss": 3.358794523009499,
      "tokens_seen": 2890268672
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012545636910732196,
      "loss": 1.24,
      "theoretical_loss": 3.35878867301515,
      "tokens_seen": 2890334208
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012543630892678035,
      "loss": 1.3611,
      "theoretical_loss": 3.358782823190582,
      "tokens_seen": 2890399744
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012541624874623872,
      "loss": 1.237,
      "theoretical_loss": 3.3587769735357873,
      "tokens_seen": 2890465280
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001253961885656971,
      "loss": 1.2598,
      "theoretical_loss": 3.3587711240507567,
      "tokens_seen": 2890530816
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012537612838515547,
      "loss": 1.2704,
      "theoretical_loss": 3.3587652747354815,
      "tokens_seen": 2890596352
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012535606820461386,
      "loss": 1.248,
      "theoretical_loss": 3.358759425589953,
      "tokens_seen": 2890661888
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012533600802407222,
      "loss": 1.3579,
      "theoretical_loss": 3.3587535766141627,
      "tokens_seen": 2890727424
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012531594784353058,
      "loss": 1.2317,
      "theoretical_loss": 3.358747727808101,
      "tokens_seen": 2890792960
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012529588766298897,
      "loss": 1.1502,
      "theoretical_loss": 3.35874187917176,
      "tokens_seen": 2890858496
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012527582748244733,
      "loss": 1.3642,
      "theoretical_loss": 3.3587360307051304,
      "tokens_seen": 2890924032
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012525576730190572,
      "loss": 1.3189,
      "theoretical_loss": 3.3587301824082036,
      "tokens_seen": 2890989568
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012523570712136408,
      "loss": 1.2503,
      "theoretical_loss": 3.3587243342809705,
      "tokens_seen": 2891055104
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012521564694082247,
      "loss": 1.2277,
      "theoretical_loss": 3.358718486323423,
      "tokens_seen": 2891120640
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012519558676028084,
      "loss": 1.238,
      "theoretical_loss": 3.3587126385355517,
      "tokens_seen": 2891186176
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012517552657973922,
      "loss": 1.2468,
      "theoretical_loss": 3.3587067909173483,
      "tokens_seen": 2891251712
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001251554663991976,
      "loss": 1.3207,
      "theoretical_loss": 3.3587009434688033,
      "tokens_seen": 2891317248
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012513540621865595,
      "loss": 1.242,
      "theoretical_loss": 3.3586950961899094,
      "tokens_seen": 2891382784
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012511534603811434,
      "loss": 1.2189,
      "theoretical_loss": 3.358689249080656,
      "tokens_seen": 2891448320
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001250952858575727,
      "loss": 1.3497,
      "theoretical_loss": 3.3586834021410357,
      "tokens_seen": 2891513856
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001250752256770311,
      "loss": 1.3459,
      "theoretical_loss": 3.358677555371039,
      "tokens_seen": 2891579392
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012505516549648945,
      "loss": 1.2288,
      "theoretical_loss": 3.3586717087706575,
      "tokens_seen": 2891644928
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012503510531594784,
      "loss": 1.2225,
      "theoretical_loss": 3.3586658623398824,
      "tokens_seen": 2891710464
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001250150451354062,
      "loss": 1.2826,
      "theoretical_loss": 3.358660016078705,
      "tokens_seen": 2891776000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001249949849548646,
      "loss": 1.3092,
      "theoretical_loss": 3.3586541699871164,
      "tokens_seen": 2891841536
    },
    {
      "epoch": 0.75,
      "objective/train/docs_used": 1051420,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3007147312164307,
      "objective/train/theoretical_loss": 3.3586483240651077,
      "objective/train/tokens_used": 1262366176,
      "theoretical_loss": 3.3586483240651077,
      "tokens_seen": 2891907072
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012497492477432298,
      "loss": 1.2745,
      "theoretical_loss": 3.3586483240651077,
      "tokens_seen": 2891907072
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012495486459378134,
      "loss": 1.2569,
      "theoretical_loss": 3.35864247831267,
      "tokens_seen": 2891972608
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001249348044132397,
      "loss": 1.2302,
      "theoretical_loss": 3.3586366327297954,
      "tokens_seen": 2892038144
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001249147442326981,
      "loss": 1.2995,
      "theoretical_loss": 3.358630787316474,
      "tokens_seen": 2892103680
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012489468405215646,
      "loss": 1.2435,
      "theoretical_loss": 3.358624942072698,
      "tokens_seen": 2892169216
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012487462387161485,
      "loss": 1.3936,
      "theoretical_loss": 3.3586190969984577,
      "tokens_seen": 2892234752
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001248545636910732,
      "loss": 1.1891,
      "theoretical_loss": 3.358613252093745,
      "tokens_seen": 2892300288
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001248345035105316,
      "loss": 1.2804,
      "theoretical_loss": 3.3586074073585515,
      "tokens_seen": 2892365824
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012481444332998996,
      "loss": 1.2798,
      "theoretical_loss": 3.3586015627928676,
      "tokens_seen": 2892431360
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012479438314944835,
      "loss": 1.2063,
      "theoretical_loss": 3.358595718396685,
      "tokens_seen": 2892496896
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012477432296890674,
      "loss": 1.3339,
      "theoretical_loss": 3.358589874169995,
      "tokens_seen": 2892562432
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001247542627883651,
      "loss": 1.3072,
      "theoretical_loss": 3.3585840301127883,
      "tokens_seen": 2892627968
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012473420260782346,
      "loss": 1.2976,
      "theoretical_loss": 3.3585781862250568,
      "tokens_seen": 2892693504
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012471414242728185,
      "loss": 1.297,
      "theoretical_loss": 3.3585723425067915,
      "tokens_seen": 2892759040
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012469408224674022,
      "loss": 1.319,
      "theoretical_loss": 3.3585664989579835,
      "tokens_seen": 2892824576
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001246740220661986,
      "loss": 1.3235,
      "theoretical_loss": 3.358560655578624,
      "tokens_seen": 2892890112
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012465396188565697,
      "loss": 1.3441,
      "theoretical_loss": 3.358554812368705,
      "tokens_seen": 2892955648
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012463390170511536,
      "loss": 1.182,
      "theoretical_loss": 3.3585489693282167,
      "tokens_seen": 2893021184
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012461384152457372,
      "loss": 1.2126,
      "theoretical_loss": 3.358543126457151,
      "tokens_seen": 2893086720
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001245937813440321,
      "loss": 1.3363,
      "theoretical_loss": 3.358537283755499,
      "tokens_seen": 2893152256
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012457372116349047,
      "loss": 1.2319,
      "theoretical_loss": 3.358531441223252,
      "tokens_seen": 2893217792
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012455366098294883,
      "loss": 1.4171,
      "theoretical_loss": 3.358525598860401,
      "tokens_seen": 2893283328
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012453360080240722,
      "loss": 1.2531,
      "theoretical_loss": 3.3585197566669374,
      "tokens_seen": 2893348864
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012451354062186559,
      "loss": 1.3336,
      "theoretical_loss": 3.3585139146428524,
      "tokens_seen": 2893414400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012449348044132397,
      "loss": 1.3165,
      "theoretical_loss": 3.358508072788138,
      "tokens_seen": 2893479936
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012447342026078234,
      "loss": 1.2905,
      "theoretical_loss": 3.3585022311027837,
      "tokens_seen": 2893545472
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012445336008024073,
      "loss": 1.2053,
      "theoretical_loss": 3.3584963895867825,
      "tokens_seen": 2893611008
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012443329989969912,
      "loss": 1.2622,
      "theoretical_loss": 3.358490548240125,
      "tokens_seen": 2893676544
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012441323971915748,
      "loss": 1.2656,
      "theoretical_loss": 3.3584847070628023,
      "tokens_seen": 2893742080
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012439317953861587,
      "loss": 1.3273,
      "theoretical_loss": 3.358478866054806,
      "tokens_seen": 2893807616
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012437311935807423,
      "loss": 1.2308,
      "theoretical_loss": 3.358473025216127,
      "tokens_seen": 2893873152
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001243530591775326,
      "loss": 1.2575,
      "theoretical_loss": 3.3584671845467566,
      "tokens_seen": 2893938688
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012433299899699098,
      "loss": 1.1865,
      "theoretical_loss": 3.3584613440466864,
      "tokens_seen": 2894004224
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012431293881644934,
      "loss": 1.2262,
      "theoretical_loss": 3.3584555037159074,
      "tokens_seen": 2894069760
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012429287863590773,
      "loss": 1.2379,
      "theoretical_loss": 3.3584496635544108,
      "tokens_seen": 2894135296
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001242728184553661,
      "loss": 1.2788,
      "theoretical_loss": 3.3584438235621876,
      "tokens_seen": 2894200832
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012425275827482448,
      "loss": 1.1976,
      "theoretical_loss": 3.35843798373923,
      "tokens_seen": 2894266368
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012423269809428285,
      "loss": 1.357,
      "theoretical_loss": 3.3584321440855285,
      "tokens_seen": 2894331904
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012421263791374124,
      "loss": 1.2663,
      "theoretical_loss": 3.3584263046010747,
      "tokens_seen": 2894397440
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001241925777331996,
      "loss": 1.316,
      "theoretical_loss": 3.3584204652858594,
      "tokens_seen": 2894462976
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000124172517552658,
      "loss": 1.3264,
      "theoretical_loss": 3.3584146261398744,
      "tokens_seen": 2894528512
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012415245737211635,
      "loss": 1.2514,
      "theoretical_loss": 3.3584087871631105,
      "tokens_seen": 2894594048
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001241323971915747,
      "loss": 1.3369,
      "theoretical_loss": 3.3584029483555593,
      "tokens_seen": 2894659584
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001241123370110331,
      "loss": 1.2765,
      "theoretical_loss": 3.3583971097172123,
      "tokens_seen": 2894725120
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012409227683049146,
      "loss": 1.3588,
      "theoretical_loss": 3.3583912712480597,
      "tokens_seen": 2894790656
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012407221664994985,
      "loss": 1.2898,
      "theoretical_loss": 3.358385432948094,
      "tokens_seen": 2894856192
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012405215646940824,
      "loss": 1.3736,
      "theoretical_loss": 3.358379594817306,
      "tokens_seen": 2894921728
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001240320962888666,
      "loss": 1.2924,
      "theoretical_loss": 3.358373756855687,
      "tokens_seen": 2894987264
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000124012036108325,
      "loss": 1.2636,
      "theoretical_loss": 3.3583679190632276,
      "tokens_seen": 2895052800
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012399197592778336,
      "loss": 1.2431,
      "theoretical_loss": 3.35836208143992,
      "tokens_seen": 2895118336
    },
    {
      "epoch": 0.75,
      "objective/train/docs_used": 1052611,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3063397407531738,
      "objective/train/theoretical_loss": 3.3583562439857557,
      "objective/train/tokens_used": 1265642976,
      "theoretical_loss": 3.3583562439857557,
      "tokens_seen": 2895183872
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012397191574724172,
      "loss": 1.2549,
      "theoretical_loss": 3.3583562439857557,
      "tokens_seen": 2895183872
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001239518555667001,
      "loss": 1.2484,
      "theoretical_loss": 3.358350406700725,
      "tokens_seen": 2895249408
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012393179538615847,
      "loss": 1.2705,
      "theoretical_loss": 3.358344569584819,
      "tokens_seen": 2895314944
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012391173520561686,
      "loss": 1.2912,
      "theoretical_loss": 3.35833873263803,
      "tokens_seen": 2895380480
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012389167502507522,
      "loss": 1.2337,
      "theoretical_loss": 3.3583328958603493,
      "tokens_seen": 2895446016
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001238716148445336,
      "loss": 1.2566,
      "theoretical_loss": 3.358327059251767,
      "tokens_seen": 2895511552
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012385155466399197,
      "loss": 1.3419,
      "theoretical_loss": 3.3583212228122754,
      "tokens_seen": 2895577088
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012383149448345036,
      "loss": 1.2042,
      "theoretical_loss": 3.3583153865418653,
      "tokens_seen": 2895642624
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012381143430290872,
      "loss": 1.3001,
      "theoretical_loss": 3.3583095504405285,
      "tokens_seen": 2895708160
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012379137412236711,
      "loss": 1.2776,
      "theoretical_loss": 3.3583037145082555,
      "tokens_seen": 2895773696
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012377131394182548,
      "loss": 1.1903,
      "theoretical_loss": 3.358297878745038,
      "tokens_seen": 2895839232
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012375125376128384,
      "loss": 1.3128,
      "theoretical_loss": 3.3582920431508674,
      "tokens_seen": 2895904768
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012373119358074223,
      "loss": 1.2044,
      "theoretical_loss": 3.358286207725735,
      "tokens_seen": 2895970304
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001237111334002006,
      "loss": 1.2816,
      "theoretical_loss": 3.3582803724696313,
      "tokens_seen": 2896035840
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012369107321965898,
      "loss": 1.2737,
      "theoretical_loss": 3.358274537382549,
      "tokens_seen": 2896101376
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012367101303911737,
      "loss": 1.2983,
      "theoretical_loss": 3.3582687024644775,
      "tokens_seen": 2896166912
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012365095285857573,
      "loss": 1.2941,
      "theoretical_loss": 3.35826286771541,
      "tokens_seen": 2896232448
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012363089267803412,
      "loss": 1.3383,
      "theoretical_loss": 3.3582570331353367,
      "tokens_seen": 2896297984
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012361083249749248,
      "loss": 1.218,
      "theoretical_loss": 3.358251198724249,
      "tokens_seen": 2896363520
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012359077231695087,
      "loss": 1.3084,
      "theoretical_loss": 3.3582453644821384,
      "tokens_seen": 2896429056
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012357071213640923,
      "loss": 1.199,
      "theoretical_loss": 3.358239530408996,
      "tokens_seen": 2896494592
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001235506519558676,
      "loss": 1.2717,
      "theoretical_loss": 3.358233696504813,
      "tokens_seen": 2896560128
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012353059177532599,
      "loss": 1.3929,
      "theoretical_loss": 3.3582278627695814,
      "tokens_seen": 2896625664
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012351053159478435,
      "loss": 1.2284,
      "theoretical_loss": 3.358222029203292,
      "tokens_seen": 2896691200
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012349047141424274,
      "loss": 1.3018,
      "theoretical_loss": 3.358216195805935,
      "tokens_seen": 2896756736
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001234704112337011,
      "loss": 1.2322,
      "theoretical_loss": 3.3582103625775037,
      "tokens_seen": 2896822272
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001234503510531595,
      "loss": 1.3656,
      "theoretical_loss": 3.358204529517988,
      "tokens_seen": 2896887808
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012343029087261785,
      "loss": 1.1616,
      "theoretical_loss": 3.358198696627379,
      "tokens_seen": 2896953344
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012341023069207624,
      "loss": 1.2544,
      "theoretical_loss": 3.3581928639056695,
      "tokens_seen": 2897018880
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001233901705115346,
      "loss": 1.3367,
      "theoretical_loss": 3.3581870313528492,
      "tokens_seen": 2897084416
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012337011033099296,
      "loss": 1.3122,
      "theoretical_loss": 3.3581811989689103,
      "tokens_seen": 2897149952
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012335005015045135,
      "loss": 1.2067,
      "theoretical_loss": 3.358175366753844,
      "tokens_seen": 2897215488
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012332998996990972,
      "loss": 1.2107,
      "theoretical_loss": 3.358169534707641,
      "tokens_seen": 2897281024
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001233099297893681,
      "loss": 1.2981,
      "theoretical_loss": 3.3581637028302933,
      "tokens_seen": 2897346560
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001232898696088265,
      "loss": 1.2868,
      "theoretical_loss": 3.358157871121792,
      "tokens_seen": 2897412096
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012326980942828486,
      "loss": 1.2071,
      "theoretical_loss": 3.358152039582128,
      "tokens_seen": 2897477632
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012324974924774325,
      "loss": 1.2432,
      "theoretical_loss": 3.3581462082112927,
      "tokens_seen": 2897543168
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001232296890672016,
      "loss": 1.2332,
      "theoretical_loss": 3.358140377009278,
      "tokens_seen": 2897608704
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012320962888666,
      "loss": 1.2381,
      "theoretical_loss": 3.3581345459760747,
      "tokens_seen": 2897674240
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012318956870611836,
      "loss": 1.2103,
      "theoretical_loss": 3.358128715111674,
      "tokens_seen": 2897739776
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012316950852557672,
      "loss": 1.2493,
      "theoretical_loss": 3.3581228844160678,
      "tokens_seen": 2897805312
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001231494483450351,
      "loss": 1.2953,
      "theoretical_loss": 3.3581170538892464,
      "tokens_seen": 2897870848
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012312938816449347,
      "loss": 1.2338,
      "theoretical_loss": 3.358111223531202,
      "tokens_seen": 2897936384
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012310932798395186,
      "loss": 1.2654,
      "theoretical_loss": 3.3581053933419254,
      "tokens_seen": 2898001920
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012308926780341023,
      "loss": 1.3701,
      "theoretical_loss": 3.3580995633214084,
      "tokens_seen": 2898067456
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012306920762286862,
      "loss": 1.2994,
      "theoretical_loss": 3.3580937334696417,
      "tokens_seen": 2898132992
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012304914744232698,
      "loss": 1.2217,
      "theoretical_loss": 3.3580879037866165,
      "tokens_seen": 2898198528
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012302908726178537,
      "loss": 1.2893,
      "theoretical_loss": 3.3580820742723247,
      "tokens_seen": 2898264064
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012300902708124373,
      "loss": 1.3858,
      "theoretical_loss": 3.3580762449267576,
      "tokens_seen": 2898329600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001229889669007021,
      "loss": 1.3045,
      "theoretical_loss": 3.358070415749906,
      "tokens_seen": 2898395136
    },
    {
      "epoch": 0.76,
      "objective/train/docs_used": 1053920,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2260544300079346,
      "objective/train/theoretical_loss": 3.3580645867417616,
      "objective/train/tokens_used": 1268919776,
      "theoretical_loss": 3.3580645867417616,
      "tokens_seen": 2898460672
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012296890672016048,
      "loss": 1.2255,
      "theoretical_loss": 3.3580645867417616,
      "tokens_seen": 2898460672
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012294884653961884,
      "loss": 1.2847,
      "theoretical_loss": 3.358058757902316,
      "tokens_seen": 2898526208
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012292878635907723,
      "loss": 1.1922,
      "theoretical_loss": 3.358052929231559,
      "tokens_seen": 2898591744
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012290872617853562,
      "loss": 1.2579,
      "theoretical_loss": 3.358047100729484,
      "tokens_seen": 2898657280
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012288866599799398,
      "loss": 1.2243,
      "theoretical_loss": 3.358041272396081,
      "tokens_seen": 2898722816
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012286860581745237,
      "loss": 1.2643,
      "theoretical_loss": 3.358035444231341,
      "tokens_seen": 2898788352
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012284854563691074,
      "loss": 1.2704,
      "theoretical_loss": 3.3580296162352563,
      "tokens_seen": 2898853888
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012282848545636912,
      "loss": 1.3148,
      "theoretical_loss": 3.358023788407818,
      "tokens_seen": 2898919424
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001228084252758275,
      "loss": 1.1846,
      "theoretical_loss": 3.3580179607490166,
      "tokens_seen": 2898984960
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012278836509528585,
      "loss": 1.3181,
      "theoretical_loss": 3.358012133258845,
      "tokens_seen": 2899050496
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012276830491474424,
      "loss": 1.3236,
      "theoretical_loss": 3.358006305937293,
      "tokens_seen": 2899116032
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001227482447342026,
      "loss": 1.2018,
      "theoretical_loss": 3.358000478784352,
      "tokens_seen": 2899181568
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000122728184553661,
      "loss": 1.3174,
      "theoretical_loss": 3.357994651800014,
      "tokens_seen": 2899247104
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012270812437311935,
      "loss": 1.2147,
      "theoretical_loss": 3.35798882498427,
      "tokens_seen": 2899312640
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012268806419257774,
      "loss": 1.1632,
      "theoretical_loss": 3.3579829983371114,
      "tokens_seen": 2899378176
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001226680040120361,
      "loss": 1.3318,
      "theoretical_loss": 3.35797717185853,
      "tokens_seen": 2899443712
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001226479438314945,
      "loss": 1.3189,
      "theoretical_loss": 3.357971345548516,
      "tokens_seen": 2899509248
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012262788365095286,
      "loss": 1.2445,
      "theoretical_loss": 3.3579655194070614,
      "tokens_seen": 2899574784
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012260782347041122,
      "loss": 1.2511,
      "theoretical_loss": 3.3579596934341573,
      "tokens_seen": 2899640320
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001225877632898696,
      "loss": 1.2532,
      "theoretical_loss": 3.3579538676297953,
      "tokens_seen": 2899705856
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012256770310932797,
      "loss": 1.2349,
      "theoretical_loss": 3.357948041993967,
      "tokens_seen": 2899771392
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012254764292878636,
      "loss": 1.2493,
      "theoretical_loss": 3.3579422165266624,
      "tokens_seen": 2899836928
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012252758274824475,
      "loss": 1.3252,
      "theoretical_loss": 3.357936391227874,
      "tokens_seen": 2899902464
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001225075225677031,
      "loss": 1.3464,
      "theoretical_loss": 3.357930566097593,
      "tokens_seen": 2899968000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001224874623871615,
      "loss": 1.3642,
      "theoretical_loss": 3.35792474113581,
      "tokens_seen": 2900033536
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012246740220661986,
      "loss": 1.2585,
      "theoretical_loss": 3.357918916342517,
      "tokens_seen": 2900099072
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012244734202607825,
      "loss": 1.2647,
      "theoretical_loss": 3.357913091717706,
      "tokens_seen": 2900164608
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001224272818455366,
      "loss": 1.3259,
      "theoretical_loss": 3.3579072672613663,
      "tokens_seen": 2900230144
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012240722166499498,
      "loss": 1.2925,
      "theoretical_loss": 3.357901442973491,
      "tokens_seen": 2900295680
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012238716148445336,
      "loss": 1.2456,
      "theoretical_loss": 3.3578956188540707,
      "tokens_seen": 2900361216
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012236710130391173,
      "loss": 1.3255,
      "theoretical_loss": 3.357889794903097,
      "tokens_seen": 2900426752
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012234704112337012,
      "loss": 1.2737,
      "theoretical_loss": 3.3578839711205606,
      "tokens_seen": 2900492288
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012232698094282848,
      "loss": 1.2088,
      "theoretical_loss": 3.3578781475064536,
      "tokens_seen": 2900557824
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012230692076228687,
      "loss": 1.339,
      "theoretical_loss": 3.3578723240607666,
      "tokens_seen": 2900623360
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012228686058174523,
      "loss": 1.2755,
      "theoretical_loss": 3.3578665007834916,
      "tokens_seen": 2900688896
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012226680040120362,
      "loss": 1.3068,
      "theoretical_loss": 3.35786067767462,
      "tokens_seen": 2900754432
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012224674022066198,
      "loss": 1.301,
      "theoretical_loss": 3.3578548547341422,
      "tokens_seen": 2900819968
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012222668004012037,
      "loss": 1.3002,
      "theoretical_loss": 3.3578490319620506,
      "tokens_seen": 2900885504
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012220661985957873,
      "loss": 1.2695,
      "theoretical_loss": 3.3578432093583355,
      "tokens_seen": 2900951040
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001221865596790371,
      "loss": 1.2566,
      "theoretical_loss": 3.357837386922989,
      "tokens_seen": 2901016576
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012216649949849549,
      "loss": 1.2543,
      "theoretical_loss": 3.3578315646560024,
      "tokens_seen": 2901082112
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012214643931795387,
      "loss": 1.145,
      "theoretical_loss": 3.3578257425573663,
      "tokens_seen": 2901147648
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012212637913741224,
      "loss": 1.2466,
      "theoretical_loss": 3.357819920627073,
      "tokens_seen": 2901213184
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012210631895687063,
      "loss": 1.3696,
      "theoretical_loss": 3.3578140988651133,
      "tokens_seen": 2901278720
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000122086258776329,
      "loss": 1.3178,
      "theoretical_loss": 3.357808277271478,
      "tokens_seen": 2901344256
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012206619859578736,
      "loss": 1.2602,
      "theoretical_loss": 3.35780245584616,
      "tokens_seen": 2901409792
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012204613841524574,
      "loss": 1.3121,
      "theoretical_loss": 3.357796634589149,
      "tokens_seen": 2901475328
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012202607823470412,
      "loss": 1.2096,
      "theoretical_loss": 3.3577908135004373,
      "tokens_seen": 2901540864
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012200601805416249,
      "loss": 1.2788,
      "theoretical_loss": 3.3577849925800156,
      "tokens_seen": 2901606400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012198595787362087,
      "loss": 1.2508,
      "theoretical_loss": 3.3577791718278758,
      "tokens_seen": 2901671936
    },
    {
      "epoch": 0.76,
      "objective/train/docs_used": 1055165,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.303365707397461,
      "objective/train/theoretical_loss": 3.357773351244009,
      "objective/train/tokens_used": 1272196576,
      "theoretical_loss": 3.357773351244009,
      "tokens_seen": 2901737472
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012196589769307924,
      "loss": 1.3423,
      "theoretical_loss": 3.357773351244009,
      "tokens_seen": 2901737472
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012194583751253762,
      "loss": 1.3355,
      "theoretical_loss": 3.3577675308284065,
      "tokens_seen": 2901803008
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012192577733199598,
      "loss": 1.1717,
      "theoretical_loss": 3.3577617105810598,
      "tokens_seen": 2901868544
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012190571715145436,
      "loss": 1.4403,
      "theoretical_loss": 3.3577558905019598,
      "tokens_seen": 2901934080
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012188565697091273,
      "loss": 1.3367,
      "theoretical_loss": 3.357750070591098,
      "tokens_seen": 2901999616
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012186559679037111,
      "loss": 1.3513,
      "theoretical_loss": 3.3577442508484663,
      "tokens_seen": 2902065152
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012184553660982948,
      "loss": 1.2988,
      "theoretical_loss": 3.3577384312740555,
      "tokens_seen": 2902130688
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012182547642928786,
      "loss": 1.2557,
      "theoretical_loss": 3.357732611867857,
      "tokens_seen": 2902196224
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012180541624874625,
      "loss": 1.2659,
      "theoretical_loss": 3.357726792629862,
      "tokens_seen": 2902261760
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012178535606820463,
      "loss": 1.37,
      "theoretical_loss": 3.3577209735600624,
      "tokens_seen": 2902327296
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000121765295887663,
      "loss": 1.3019,
      "theoretical_loss": 3.357715154658449,
      "tokens_seen": 2902392832
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012174523570712138,
      "loss": 1.401,
      "theoretical_loss": 3.3577093359250134,
      "tokens_seen": 2902458368
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012172517552657974,
      "loss": 1.3092,
      "theoretical_loss": 3.357703517359747,
      "tokens_seen": 2902523904
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012170511534603811,
      "loss": 1.2745,
      "theoretical_loss": 3.3576976989626406,
      "tokens_seen": 2902589440
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012168505516549649,
      "loss": 1.2987,
      "theoretical_loss": 3.357691880733686,
      "tokens_seen": 2902654976
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012166499498495487,
      "loss": 1.2132,
      "theoretical_loss": 3.3576860626728746,
      "tokens_seen": 2902720512
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012164493480441324,
      "loss": 1.2113,
      "theoretical_loss": 3.3576802447801977,
      "tokens_seen": 2902786048
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012162487462387162,
      "loss": 1.2508,
      "theoretical_loss": 3.3576744270556462,
      "tokens_seen": 2902851584
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012160481444333,
      "loss": 1.2411,
      "theoretical_loss": 3.3576686094992123,
      "tokens_seen": 2902917120
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012158475426278837,
      "loss": 1.258,
      "theoretical_loss": 3.3576627921108866,
      "tokens_seen": 2902982656
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012156469408224675,
      "loss": 1.3663,
      "theoretical_loss": 3.357656974890661,
      "tokens_seen": 2903048192
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012154463390170512,
      "loss": 1.2618,
      "theoretical_loss": 3.3576511578385264,
      "tokens_seen": 2903113728
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012152457372116348,
      "loss": 1.3469,
      "theoretical_loss": 3.357645340954474,
      "tokens_seen": 2903179264
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012150451354062186,
      "loss": 1.2503,
      "theoretical_loss": 3.357639524238496,
      "tokens_seen": 2903244800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012148445336008024,
      "loss": 1.2973,
      "theoretical_loss": 3.357633707690583,
      "tokens_seen": 2903310336
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012146439317953861,
      "loss": 1.1646,
      "theoretical_loss": 3.3576278913107265,
      "tokens_seen": 2903375872
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012144433299899699,
      "loss": 1.234,
      "theoretical_loss": 3.357622075098918,
      "tokens_seen": 2903441408
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012142427281845538,
      "loss": 1.3465,
      "theoretical_loss": 3.3576162590551486,
      "tokens_seen": 2903506944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012140421263791375,
      "loss": 1.2033,
      "theoretical_loss": 3.35761044317941,
      "tokens_seen": 2903572480
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012138415245737213,
      "loss": 1.2084,
      "theoretical_loss": 3.3576046274716935,
      "tokens_seen": 2903638016
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001213640922768305,
      "loss": 1.3632,
      "theoretical_loss": 3.35759881193199,
      "tokens_seen": 2903703552
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012134403209628888,
      "loss": 1.3129,
      "theoretical_loss": 3.357592996560291,
      "tokens_seen": 2903769088
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012132397191574724,
      "loss": 1.286,
      "theoretical_loss": 3.3575871813565885,
      "tokens_seen": 2903834624
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012130391173520562,
      "loss": 1.1585,
      "theoretical_loss": 3.3575813663208733,
      "tokens_seen": 2903900160
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012128385155466399,
      "loss": 1.3248,
      "theoretical_loss": 3.3575755514531367,
      "tokens_seen": 2903965696
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012126379137412237,
      "loss": 1.3052,
      "theoretical_loss": 3.35756973675337,
      "tokens_seen": 2904031232
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012124373119358074,
      "loss": 1.3088,
      "theoretical_loss": 3.3575639222215656,
      "tokens_seen": 2904096768
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012122367101303912,
      "loss": 1.2514,
      "theoretical_loss": 3.357558107857713,
      "tokens_seen": 2904162304
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001212036108324975,
      "loss": 1.3064,
      "theoretical_loss": 3.357552293661805,
      "tokens_seen": 2904227840
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012118355065195587,
      "loss": 1.1858,
      "theoretical_loss": 3.357546479633833,
      "tokens_seen": 2904293376
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012116349047141425,
      "loss": 1.3208,
      "theoretical_loss": 3.3575406657737874,
      "tokens_seen": 2904358912
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012114343029087261,
      "loss": 1.3063,
      "theoretical_loss": 3.3575348520816597,
      "tokens_seen": 2904424448
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012112337011033099,
      "loss": 1.2641,
      "theoretical_loss": 3.357529038557442,
      "tokens_seen": 2904489984
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012110330992978936,
      "loss": 1.2615,
      "theoretical_loss": 3.357523225201126,
      "tokens_seen": 2904555520
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012108324974924774,
      "loss": 1.254,
      "theoretical_loss": 3.3575174120127014,
      "tokens_seen": 2904621056
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012106318956870611,
      "loss": 1.2646,
      "theoretical_loss": 3.3575115989921605,
      "tokens_seen": 2904686592
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001210431293881645,
      "loss": 1.2277,
      "theoretical_loss": 3.357505786139495,
      "tokens_seen": 2904752128
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012102306920762288,
      "loss": 1.2648,
      "theoretical_loss": 3.357499973454696,
      "tokens_seen": 2904817664
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012100300902708125,
      "loss": 1.2993,
      "theoretical_loss": 3.3574941609377547,
      "tokens_seen": 2904883200
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012098294884653963,
      "loss": 1.2185,
      "theoretical_loss": 3.3574883485886624,
      "tokens_seen": 2904948736
    },
    {
      "epoch": 0.76,
      "objective/train/docs_used": 1056274,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4536802768707275,
      "objective/train/theoretical_loss": 3.357482536407411,
      "objective/train/tokens_used": 1275473376,
      "theoretical_loss": 3.357482536407411,
      "tokens_seen": 2905014272
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000120962888665998,
      "loss": 1.3944,
      "theoretical_loss": 3.357482536407411,
      "tokens_seen": 2905014272
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012094282848545637,
      "loss": 1.2635,
      "theoretical_loss": 3.357476724393991,
      "tokens_seen": 2905079808
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012092276830491474,
      "loss": 1.2549,
      "theoretical_loss": 3.357470912548395,
      "tokens_seen": 2905145344
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012090270812437312,
      "loss": 1.3087,
      "theoretical_loss": 3.3574651008706127,
      "tokens_seen": 2905210880
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001208826479438315,
      "loss": 1.2233,
      "theoretical_loss": 3.357459289360637,
      "tokens_seen": 2905276416
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012086258776328987,
      "loss": 1.2891,
      "theoretical_loss": 3.3574534780184586,
      "tokens_seen": 2905341952
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012084252758274825,
      "loss": 1.2403,
      "theoretical_loss": 3.357447666844069,
      "tokens_seen": 2905407488
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012082246740220662,
      "loss": 1.2183,
      "theoretical_loss": 3.3574418558374592,
      "tokens_seen": 2905473024
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000120802407221665,
      "loss": 1.3268,
      "theoretical_loss": 3.357436044998621,
      "tokens_seen": 2905538560
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012078234704112337,
      "loss": 1.2874,
      "theoretical_loss": 3.3574302343275457,
      "tokens_seen": 2905604096
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012076228686058175,
      "loss": 1.2693,
      "theoretical_loss": 3.357424423824225,
      "tokens_seen": 2905669632
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012074222668004011,
      "loss": 1.3742,
      "theoretical_loss": 3.3574186134886492,
      "tokens_seen": 2905735168
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012072216649949849,
      "loss": 1.2258,
      "theoretical_loss": 3.357412803320811,
      "tokens_seen": 2905800704
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012070210631895686,
      "loss": 1.2762,
      "theoretical_loss": 3.357406993320701,
      "tokens_seen": 2905866240
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012068204613841525,
      "loss": 1.2559,
      "theoretical_loss": 3.3574011834883106,
      "tokens_seen": 2905931776
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012066198595787363,
      "loss": 1.3421,
      "theoretical_loss": 3.3573953738236315,
      "tokens_seen": 2905997312
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000120641925777332,
      "loss": 1.2924,
      "theoretical_loss": 3.3573895643266547,
      "tokens_seen": 2906062848
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012062186559679038,
      "loss": 1.3278,
      "theoretical_loss": 3.3573837549973717,
      "tokens_seen": 2906128384
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012060180541624876,
      "loss": 1.2662,
      "theoretical_loss": 3.357377945835774,
      "tokens_seen": 2906193920
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012058174523570713,
      "loss": 1.2327,
      "theoretical_loss": 3.357372136841853,
      "tokens_seen": 2906259456
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012056168505516551,
      "loss": 1.2646,
      "theoretical_loss": 3.3573663280156,
      "tokens_seen": 2906324992
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012054162487462387,
      "loss": 1.1859,
      "theoretical_loss": 3.357360519357006,
      "tokens_seen": 2906390528
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012052156469408225,
      "loss": 1.3437,
      "theoretical_loss": 3.3573547108660633,
      "tokens_seen": 2906456064
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012050150451354062,
      "loss": 1.2724,
      "theoretical_loss": 3.3573489025427623,
      "tokens_seen": 2906521600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000120481444332999,
      "loss": 1.2344,
      "theoretical_loss": 3.357343094387095,
      "tokens_seen": 2906587136
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012046138415245737,
      "loss": 1.3557,
      "theoretical_loss": 3.357337286399053,
      "tokens_seen": 2906652672
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012044132397191575,
      "loss": 1.2315,
      "theoretical_loss": 3.357331478578627,
      "tokens_seen": 2906718208
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012042126379137412,
      "loss": 1.3091,
      "theoretical_loss": 3.3573256709258086,
      "tokens_seen": 2906783744
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001204012036108325,
      "loss": 1.2848,
      "theoretical_loss": 3.3573198634405887,
      "tokens_seen": 2906849280
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012038114343029088,
      "loss": 1.358,
      "theoretical_loss": 3.35731405612296,
      "tokens_seen": 2906914816
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012036108324974924,
      "loss": 1.2356,
      "theoretical_loss": 3.357308248972913,
      "tokens_seen": 2906980352
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012034102306920761,
      "loss": 1.1913,
      "theoretical_loss": 3.357302441990439,
      "tokens_seen": 2907045888
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012032096288866599,
      "loss": 1.2853,
      "theoretical_loss": 3.35729663517553,
      "tokens_seen": 2907111424
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012030090270812438,
      "loss": 1.2886,
      "theoretical_loss": 3.3572908285281766,
      "tokens_seen": 2907176960
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012028084252758276,
      "loss": 1.2432,
      "theoretical_loss": 3.3572850220483703,
      "tokens_seen": 2907242496
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012026078234704113,
      "loss": 1.1851,
      "theoretical_loss": 3.3572792157361033,
      "tokens_seen": 2907308032
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012024072216649951,
      "loss": 1.2722,
      "theoretical_loss": 3.3572734095913663,
      "tokens_seen": 2907373568
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012022066198595788,
      "loss": 1.3336,
      "theoretical_loss": 3.357267603614151,
      "tokens_seen": 2907439104
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012020060180541626,
      "loss": 1.2288,
      "theoretical_loss": 3.357261797804448,
      "tokens_seen": 2907504640
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012018054162487463,
      "loss": 1.3346,
      "theoretical_loss": 3.35725599216225,
      "tokens_seen": 2907570176
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000120160481444333,
      "loss": 1.1662,
      "theoretical_loss": 3.3572501866875477,
      "tokens_seen": 2907635712
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012014042126379137,
      "loss": 1.297,
      "theoretical_loss": 3.357244381380332,
      "tokens_seen": 2907701248
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012012036108324975,
      "loss": 1.287,
      "theoretical_loss": 3.357238576240595,
      "tokens_seen": 2907766784
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012010030090270812,
      "loss": 1.1871,
      "theoretical_loss": 3.357232771268328,
      "tokens_seen": 2907832320
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001200802407221665,
      "loss": 1.3915,
      "theoretical_loss": 3.3572269664635224,
      "tokens_seen": 2907897856
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012006018054162488,
      "loss": 1.3082,
      "theoretical_loss": 3.3572211618261694,
      "tokens_seen": 2907963392
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012004012036108325,
      "loss": 1.1708,
      "theoretical_loss": 3.35721535735626,
      "tokens_seen": 2908028928
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012002006018054163,
      "loss": 1.1656,
      "theoretical_loss": 3.357209553053787,
      "tokens_seen": 2908094464
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012,
      "loss": 1.293,
      "theoretical_loss": 3.35720374891874,
      "tokens_seen": 2908160000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011997993981945838,
      "loss": 1.2081,
      "theoretical_loss": 3.357197944951112,
      "tokens_seen": 2908225536
    },
    {
      "epoch": 0.76,
      "objective/train/docs_used": 1057433,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.301594614982605,
      "objective/train/theoretical_loss": 3.357192141150893,
      "objective/train/tokens_used": 1278750176,
      "theoretical_loss": 3.357192141150893,
      "tokens_seen": 2908291072
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011995987963891674,
      "loss": 1.2957,
      "theoretical_loss": 3.357192141150893,
      "tokens_seen": 2908291072
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011993981945837512,
      "loss": 1.263,
      "theoretical_loss": 3.3571863375180753,
      "tokens_seen": 2908356608
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001199197592778335,
      "loss": 1.3379,
      "theoretical_loss": 3.35718053405265,
      "tokens_seen": 2908422144
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011989969909729188,
      "loss": 1.2291,
      "theoretical_loss": 3.357174730754609,
      "tokens_seen": 2908487680
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011987963891675026,
      "loss": 1.2519,
      "theoretical_loss": 3.3571689276239427,
      "tokens_seen": 2908553216
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011985957873620863,
      "loss": 1.2557,
      "theoretical_loss": 3.3571631246606435,
      "tokens_seen": 2908618752
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011983951855566701,
      "loss": 1.2438,
      "theoretical_loss": 3.357157321864702,
      "tokens_seen": 2908684288
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011981945837512539,
      "loss": 1.2389,
      "theoretical_loss": 3.35715151923611,
      "tokens_seen": 2908749824
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011979939819458376,
      "loss": 1.2059,
      "theoretical_loss": 3.3571457167748595,
      "tokens_seen": 2908815360
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011977933801404214,
      "loss": 1.2509,
      "theoretical_loss": 3.357139914480941,
      "tokens_seen": 2908880896
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001197592778335005,
      "loss": 1.2672,
      "theoretical_loss": 3.3571341123543457,
      "tokens_seen": 2908946432
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011973921765295887,
      "loss": 1.0964,
      "theoretical_loss": 3.357128310395066,
      "tokens_seen": 2909011968
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011971915747241725,
      "loss": 1.2918,
      "theoretical_loss": 3.3571225086030925,
      "tokens_seen": 2909077504
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011969909729187563,
      "loss": 1.3304,
      "theoretical_loss": 3.357116706978417,
      "tokens_seen": 2909143040
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000119679037111334,
      "loss": 1.3241,
      "theoretical_loss": 3.357110905521031,
      "tokens_seen": 2909208576
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011965897693079238,
      "loss": 1.2552,
      "theoretical_loss": 3.3571051042309255,
      "tokens_seen": 2909274112
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011963891675025075,
      "loss": 1.349,
      "theoretical_loss": 3.3570993031080922,
      "tokens_seen": 2909339648
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011961885656970913,
      "loss": 1.1997,
      "theoretical_loss": 3.357093502152522,
      "tokens_seen": 2909405184
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001195987963891675,
      "loss": 1.2821,
      "theoretical_loss": 3.3570877013642075,
      "tokens_seen": 2909470720
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011957873620862587,
      "loss": 1.2409,
      "theoretical_loss": 3.3570819007431387,
      "tokens_seen": 2909536256
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011955867602808424,
      "loss": 1.3237,
      "theoretical_loss": 3.357076100289308,
      "tokens_seen": 2909601792
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011953861584754263,
      "loss": 1.3533,
      "theoretical_loss": 3.3570703000027065,
      "tokens_seen": 2909667328
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011951855566700101,
      "loss": 1.3026,
      "theoretical_loss": 3.3570644998833257,
      "tokens_seen": 2909732864
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011949849548645938,
      "loss": 1.1834,
      "theoretical_loss": 3.357058699931157,
      "tokens_seen": 2909798400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011947843530591776,
      "loss": 1.2628,
      "theoretical_loss": 3.3570529001461913,
      "tokens_seen": 2909863936
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011945837512537614,
      "loss": 1.2491,
      "theoretical_loss": 3.357047100528421,
      "tokens_seen": 2909929472
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011943831494483451,
      "loss": 1.1878,
      "theoretical_loss": 3.357041301077836,
      "tokens_seen": 2909995008
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011941825476429289,
      "loss": 1.3373,
      "theoretical_loss": 3.3570355017944293,
      "tokens_seen": 2910060544
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011939819458375126,
      "loss": 1.3134,
      "theoretical_loss": 3.357029702678192,
      "tokens_seen": 2910126080
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011937813440320963,
      "loss": 1.2403,
      "theoretical_loss": 3.3570239037291145,
      "tokens_seen": 2910191616
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000119358074222668,
      "loss": 1.1522,
      "theoretical_loss": 3.357018104947189,
      "tokens_seen": 2910257152
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011933801404212638,
      "loss": 1.3349,
      "theoretical_loss": 3.357012306332407,
      "tokens_seen": 2910322688
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011931795386158475,
      "loss": 1.4379,
      "theoretical_loss": 3.35700650788476,
      "tokens_seen": 2910388224
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011929789368104313,
      "loss": 1.2491,
      "theoretical_loss": 3.3570007096042387,
      "tokens_seen": 2910453760
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001192778335005015,
      "loss": 1.183,
      "theoretical_loss": 3.3569949114908355,
      "tokens_seen": 2910519296
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011925777331995988,
      "loss": 1.1454,
      "theoretical_loss": 3.3569891135445413,
      "tokens_seen": 2910584832
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011923771313941826,
      "loss": 1.3843,
      "theoretical_loss": 3.356983315765347,
      "tokens_seen": 2910650368
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011921765295887663,
      "loss": 1.2253,
      "theoretical_loss": 3.356977518153245,
      "tokens_seen": 2910715904
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011919759277833501,
      "loss": 1.3764,
      "theoretical_loss": 3.356971720708226,
      "tokens_seen": 2910781440
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011917753259779338,
      "loss": 1.3867,
      "theoretical_loss": 3.3569659234302818,
      "tokens_seen": 2910846976
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011915747241725176,
      "loss": 1.2544,
      "theoretical_loss": 3.3569601263194038,
      "tokens_seen": 2910912512
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011913741223671014,
      "loss": 1.3672,
      "theoretical_loss": 3.3569543293755832,
      "tokens_seen": 2910978048
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011911735205616851,
      "loss": 1.3181,
      "theoretical_loss": 3.3569485325988113,
      "tokens_seen": 2911043584
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011909729187562689,
      "loss": 1.2403,
      "theoretical_loss": 3.3569427359890804,
      "tokens_seen": 2911109120
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011907723169508526,
      "loss": 1.4067,
      "theoretical_loss": 3.3569369395463813,
      "tokens_seen": 2911174656
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011905717151454364,
      "loss": 1.2597,
      "theoretical_loss": 3.356931143270705,
      "tokens_seen": 2911240192
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011903711133400201,
      "loss": 1.2838,
      "theoretical_loss": 3.3569253471620435,
      "tokens_seen": 2911305728
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011901705115346039,
      "loss": 1.2984,
      "theoretical_loss": 3.356919551220388,
      "tokens_seen": 2911371264
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011899699097291877,
      "loss": 1.2265,
      "theoretical_loss": 3.3569137554457305,
      "tokens_seen": 2911436800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011897693079237713,
      "loss": 1.2851,
      "theoretical_loss": 3.3569079598380616,
      "tokens_seen": 2911502336
    },
    {
      "epoch": 0.76,
      "objective/train/docs_used": 1058588,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3200011253356934,
      "objective/train/theoretical_loss": 3.3569021643973733,
      "objective/train/tokens_used": 1282026976,
      "theoretical_loss": 3.3569021643973733,
      "tokens_seen": 2911567872
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001189568706118355,
      "loss": 1.2851,
      "theoretical_loss": 3.3569021643973733,
      "tokens_seen": 2911567872
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011893681043129388,
      "loss": 1.2325,
      "theoretical_loss": 3.356896369123657,
      "tokens_seen": 2911633408
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011891675025075226,
      "loss": 1.1928,
      "theoretical_loss": 3.3568905740169033,
      "tokens_seen": 2911698944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011889669007021063,
      "loss": 1.2467,
      "theoretical_loss": 3.3568847790771046,
      "tokens_seen": 2911764480
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011887662988966901,
      "loss": 1.2583,
      "theoretical_loss": 3.3568789843042524,
      "tokens_seen": 2911830016
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011885656970912738,
      "loss": 1.2635,
      "theoretical_loss": 3.3568731896983373,
      "tokens_seen": 2911895552
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011883650952858576,
      "loss": 1.2502,
      "theoretical_loss": 3.3568673952593513,
      "tokens_seen": 2911961088
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011881644934804413,
      "loss": 1.2589,
      "theoretical_loss": 3.3568616009872856,
      "tokens_seen": 2912026624
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011879638916750251,
      "loss": 1.3119,
      "theoretical_loss": 3.356855806882132,
      "tokens_seen": 2912092160
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011877632898696089,
      "loss": 1.3859,
      "theoretical_loss": 3.3568500129438816,
      "tokens_seen": 2912157696
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011875626880641926,
      "loss": 1.3816,
      "theoretical_loss": 3.3568442191725256,
      "tokens_seen": 2912223232
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011873620862587764,
      "loss": 1.2385,
      "theoretical_loss": 3.356838425568056,
      "tokens_seen": 2912288768
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011871614844533601,
      "loss": 1.3177,
      "theoretical_loss": 3.3568326321304642,
      "tokens_seen": 2912354304
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011869608826479439,
      "loss": 1.3196,
      "theoretical_loss": 3.3568268388597415,
      "tokens_seen": 2912419840
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011867602808425276,
      "loss": 1.2776,
      "theoretical_loss": 3.356821045755879,
      "tokens_seen": 2912485376
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011865596790371114,
      "loss": 1.2592,
      "theoretical_loss": 3.3568152528188686,
      "tokens_seen": 2912550912
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011863590772316952,
      "loss": 1.3153,
      "theoretical_loss": 3.3568094600487015,
      "tokens_seen": 2912616448
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011861584754262789,
      "loss": 1.1542,
      "theoretical_loss": 3.356803667445369,
      "tokens_seen": 2912681984
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011859578736208625,
      "loss": 1.331,
      "theoretical_loss": 3.356797875008863,
      "tokens_seen": 2912747520
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011857572718154463,
      "loss": 1.2723,
      "theoretical_loss": 3.3567920827391746,
      "tokens_seen": 2912813056
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000118555667001003,
      "loss": 1.3059,
      "theoretical_loss": 3.3567862906362955,
      "tokens_seen": 2912878592
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011853560682046138,
      "loss": 1.2813,
      "theoretical_loss": 3.3567804987002168,
      "tokens_seen": 2912944128
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011851554663991976,
      "loss": 1.1765,
      "theoretical_loss": 3.35677470693093,
      "tokens_seen": 2913009664
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011849548645937813,
      "loss": 1.1466,
      "theoretical_loss": 3.3567689153284266,
      "tokens_seen": 2913075200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011847542627883651,
      "loss": 1.2231,
      "theoretical_loss": 3.3567631238926983,
      "tokens_seen": 2913140736
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011845536609829488,
      "loss": 1.3583,
      "theoretical_loss": 3.3567573326237365,
      "tokens_seen": 2913206272
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011843530591775326,
      "loss": 1.2452,
      "theoretical_loss": 3.3567515415215325,
      "tokens_seen": 2913271808
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011841524573721165,
      "loss": 1.3854,
      "theoretical_loss": 3.3567457505860774,
      "tokens_seen": 2913337344
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011839518555667001,
      "loss": 1.2305,
      "theoretical_loss": 3.356739959817363,
      "tokens_seen": 2913402880
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011837512537612839,
      "loss": 1.2568,
      "theoretical_loss": 3.356734169215381,
      "tokens_seen": 2913468416
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011835506519558676,
      "loss": 1.2091,
      "theoretical_loss": 3.3567283787801223,
      "tokens_seen": 2913533952
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011833500501504514,
      "loss": 1.3707,
      "theoretical_loss": 3.356722588511579,
      "tokens_seen": 2913599488
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011831494483450352,
      "loss": 1.3394,
      "theoretical_loss": 3.356716798409742,
      "tokens_seen": 2913665024
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011829488465396189,
      "loss": 1.2965,
      "theoretical_loss": 3.356711008474603,
      "tokens_seen": 2913730560
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011827482447342027,
      "loss": 1.2505,
      "theoretical_loss": 3.3567052187061535,
      "tokens_seen": 2913796096
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011825476429287864,
      "loss": 1.2279,
      "theoretical_loss": 3.3566994291043843,
      "tokens_seen": 2913861632
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011823470411233702,
      "loss": 1.2859,
      "theoretical_loss": 3.356693639669288,
      "tokens_seen": 2913927168
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011821464393179538,
      "loss": 1.1989,
      "theoretical_loss": 3.356687850400855,
      "tokens_seen": 2913992704
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011819458375125376,
      "loss": 1.2564,
      "theoretical_loss": 3.3566820612990775,
      "tokens_seen": 2914058240
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011817452357071213,
      "loss": 1.2986,
      "theoretical_loss": 3.3566762723639467,
      "tokens_seen": 2914123776
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011815446339017051,
      "loss": 1.2156,
      "theoretical_loss": 3.3566704835954537,
      "tokens_seen": 2914189312
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011813440320962888,
      "loss": 1.2844,
      "theoretical_loss": 3.3566646949935905,
      "tokens_seen": 2914254848
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011811434302908726,
      "loss": 1.2067,
      "theoretical_loss": 3.3566589065583483,
      "tokens_seen": 2914320384
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011809428284854564,
      "loss": 1.2674,
      "theoretical_loss": 3.3566531182897186,
      "tokens_seen": 2914385920
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011807422266800401,
      "loss": 1.2731,
      "theoretical_loss": 3.356647330187693,
      "tokens_seen": 2914451456
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001180541624874624,
      "loss": 1.3727,
      "theoretical_loss": 3.3566415422522624,
      "tokens_seen": 2914516992
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011803410230692078,
      "loss": 1.2784,
      "theoretical_loss": 3.356635754483419,
      "tokens_seen": 2914582528
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011801404212637914,
      "loss": 1.2426,
      "theoretical_loss": 3.3566299668811537,
      "tokens_seen": 2914648064
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011799398194583751,
      "loss": 1.1966,
      "theoretical_loss": 3.3566241794454585,
      "tokens_seen": 2914713600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011797392176529589,
      "loss": 1.2821,
      "theoretical_loss": 3.356618392176324,
      "tokens_seen": 2914779136
    },
    {
      "epoch": 0.77,
      "objective/train/docs_used": 1059748,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2119574546813965,
      "objective/train/theoretical_loss": 3.3566126050737424,
      "objective/train/tokens_used": 1285303776,
      "theoretical_loss": 3.3566126050737424,
      "tokens_seen": 2914844672
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011795386158475427,
      "loss": 1.332,
      "theoretical_loss": 3.3566126050737424,
      "tokens_seen": 2914844672
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011793380140421264,
      "loss": 1.2718,
      "theoretical_loss": 3.3566068181377053,
      "tokens_seen": 2914910208
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011791374122367102,
      "loss": 1.3486,
      "theoretical_loss": 3.3566010313682035,
      "tokens_seen": 2914975744
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001178936810431294,
      "loss": 1.2795,
      "theoretical_loss": 3.356595244765229,
      "tokens_seen": 2915041280
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011787362086258777,
      "loss": 1.2299,
      "theoretical_loss": 3.3565894583287728,
      "tokens_seen": 2915106816
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011785356068204615,
      "loss": 1.187,
      "theoretical_loss": 3.356583672058827,
      "tokens_seen": 2915172352
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011783350050150452,
      "loss": 1.2874,
      "theoretical_loss": 3.356577885955382,
      "tokens_seen": 2915237888
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011781344032096288,
      "loss": 1.3674,
      "theoretical_loss": 3.3565721000184308,
      "tokens_seen": 2915303424
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011779338014042126,
      "loss": 1.4094,
      "theoretical_loss": 3.3565663142479636,
      "tokens_seen": 2915368960
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011777331995987963,
      "loss": 1.2934,
      "theoretical_loss": 3.356560528643972,
      "tokens_seen": 2915434496
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011775325977933801,
      "loss": 1.314,
      "theoretical_loss": 3.3565547432064484,
      "tokens_seen": 2915500032
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011773319959879639,
      "loss": 1.3932,
      "theoretical_loss": 3.356548957935383,
      "tokens_seen": 2915565568
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011771313941825476,
      "loss": 1.2776,
      "theoretical_loss": 3.356543172830768,
      "tokens_seen": 2915631104
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011769307923771314,
      "loss": 1.3302,
      "theoretical_loss": 3.356537387892595,
      "tokens_seen": 2915696640
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011767301905717153,
      "loss": 1.2146,
      "theoretical_loss": 3.356531603120855,
      "tokens_seen": 2915762176
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001176529588766299,
      "loss": 1.3351,
      "theoretical_loss": 3.35652581851554,
      "tokens_seen": 2915827712
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011763289869608828,
      "loss": 1.3367,
      "theoretical_loss": 3.356520034076641,
      "tokens_seen": 2915893248
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011761283851554664,
      "loss": 1.2307,
      "theoretical_loss": 3.3565142498041496,
      "tokens_seen": 2915958784
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011759277833500502,
      "loss": 1.1699,
      "theoretical_loss": 3.3565084656980573,
      "tokens_seen": 2916024320
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011757271815446339,
      "loss": 1.3098,
      "theoretical_loss": 3.356502681758356,
      "tokens_seen": 2916089856
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011755265797392177,
      "loss": 1.2805,
      "theoretical_loss": 3.356496897985036,
      "tokens_seen": 2916155392
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011753259779338014,
      "loss": 1.2671,
      "theoretical_loss": 3.3564911143780902,
      "tokens_seen": 2916220928
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011751253761283852,
      "loss": 1.2808,
      "theoretical_loss": 3.356485330937509,
      "tokens_seen": 2916286464
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001174924774322969,
      "loss": 1.2282,
      "theoretical_loss": 3.3564795476632847,
      "tokens_seen": 2916352000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011747241725175527,
      "loss": 1.1891,
      "theoretical_loss": 3.356473764555408,
      "tokens_seen": 2916417536
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011745235707121365,
      "loss": 1.2068,
      "theoretical_loss": 3.356467981613871,
      "tokens_seen": 2916483072
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011743229689067201,
      "loss": 1.2595,
      "theoretical_loss": 3.3564621988386647,
      "tokens_seen": 2916548608
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011741223671013039,
      "loss": 1.2834,
      "theoretical_loss": 3.356456416229781,
      "tokens_seen": 2916614144
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011739217652958876,
      "loss": 1.2401,
      "theoretical_loss": 3.356450633787211,
      "tokens_seen": 2916679680
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011737211634904714,
      "loss": 1.3473,
      "theoretical_loss": 3.3564448515109464,
      "tokens_seen": 2916745216
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011735205616850551,
      "loss": 1.3066,
      "theoretical_loss": 3.3564390694009787,
      "tokens_seen": 2916810752
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011733199598796389,
      "loss": 1.2135,
      "theoretical_loss": 3.3564332874572993,
      "tokens_seen": 2916876288
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011731193580742226,
      "loss": 1.2235,
      "theoretical_loss": 3.3564275056799,
      "tokens_seen": 2916941824
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011729187562688065,
      "loss": 1.3134,
      "theoretical_loss": 3.3564217240687713,
      "tokens_seen": 2917007360
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011727181544633903,
      "loss": 1.2554,
      "theoretical_loss": 3.3564159426239057,
      "tokens_seen": 2917072896
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001172517552657974,
      "loss": 1.2583,
      "theoretical_loss": 3.3564101613452944,
      "tokens_seen": 2917138432
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011723169508525577,
      "loss": 1.3428,
      "theoretical_loss": 3.356404380232929,
      "tokens_seen": 2917203968
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011721163490471414,
      "loss": 1.2906,
      "theoretical_loss": 3.3563985992868006,
      "tokens_seen": 2917269504
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011719157472417252,
      "loss": 1.2445,
      "theoretical_loss": 3.356392818506901,
      "tokens_seen": 2917335040
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001171715145436309,
      "loss": 1.3416,
      "theoretical_loss": 3.3563870378932217,
      "tokens_seen": 2917400576
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011715145436308927,
      "loss": 1.208,
      "theoretical_loss": 3.3563812574457534,
      "tokens_seen": 2917466112
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011713139418254765,
      "loss": 1.3958,
      "theoretical_loss": 3.356375477164489,
      "tokens_seen": 2917531648
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011711133400200602,
      "loss": 1.2348,
      "theoretical_loss": 3.356369697049419,
      "tokens_seen": 2917597184
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001170912738214644,
      "loss": 1.2404,
      "theoretical_loss": 3.3563639171005355,
      "tokens_seen": 2917662720
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011707121364092277,
      "loss": 1.2669,
      "theoretical_loss": 3.3563581373178293,
      "tokens_seen": 2917728256
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011705115346038115,
      "loss": 1.3331,
      "theoretical_loss": 3.3563523577012924,
      "tokens_seen": 2917793792
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011703109327983951,
      "loss": 1.2615,
      "theoretical_loss": 3.356346578250916,
      "tokens_seen": 2917859328
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011701103309929789,
      "loss": 1.3063,
      "theoretical_loss": 3.3563407989666914,
      "tokens_seen": 2917924864
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011699097291875626,
      "loss": 1.2499,
      "theoretical_loss": 3.356335019848611,
      "tokens_seen": 2917990400
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011697091273821464,
      "loss": 1.2496,
      "theoretical_loss": 3.356329240896665,
      "tokens_seen": 2918055936
    },
    {
      "epoch": 0.77,
      "objective/train/docs_used": 1060857,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3790392875671387,
      "objective/train/theoretical_loss": 3.356323462110846,
      "objective/train/tokens_used": 1288580576,
      "theoretical_loss": 3.356323462110846,
      "tokens_seen": 2918121472
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011695085255767302,
      "loss": 1.3866,
      "theoretical_loss": 3.356323462110846,
      "tokens_seen": 2918121472
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011693079237713139,
      "loss": 1.2707,
      "theoretical_loss": 3.3563176834911452,
      "tokens_seen": 2918187008
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011691073219658978,
      "loss": 1.3591,
      "theoretical_loss": 3.3563119050375536,
      "tokens_seen": 2918252544
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011689067201604816,
      "loss": 1.1274,
      "theoretical_loss": 3.3563061267500633,
      "tokens_seen": 2918318080
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011687061183550653,
      "loss": 1.3603,
      "theoretical_loss": 3.3563003486286656,
      "tokens_seen": 2918383616
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011685055165496491,
      "loss": 1.3425,
      "theoretical_loss": 3.356294570673352,
      "tokens_seen": 2918449152
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011683049147442327,
      "loss": 1.2642,
      "theoretical_loss": 3.3562887928841136,
      "tokens_seen": 2918514688
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011681043129388165,
      "loss": 1.2044,
      "theoretical_loss": 3.356283015260942,
      "tokens_seen": 2918580224
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011679037111334002,
      "loss": 1.3268,
      "theoretical_loss": 3.35627723780383,
      "tokens_seen": 2918645760
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001167703109327984,
      "loss": 1.3613,
      "theoretical_loss": 3.3562714605127675,
      "tokens_seen": 2918711296
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011675025075225677,
      "loss": 1.2259,
      "theoretical_loss": 3.3562656833877464,
      "tokens_seen": 2918776832
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011673019057171515,
      "loss": 1.3234,
      "theoretical_loss": 3.3562599064287584,
      "tokens_seen": 2918842368
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011671013039117352,
      "loss": 1.3335,
      "theoretical_loss": 3.3562541296357953,
      "tokens_seen": 2918907904
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001166900702106319,
      "loss": 1.2858,
      "theoretical_loss": 3.356248353008848,
      "tokens_seen": 2918973440
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011667001003009028,
      "loss": 1.196,
      "theoretical_loss": 3.356242576547908,
      "tokens_seen": 2919038976
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011664994984954864,
      "loss": 1.2403,
      "theoretical_loss": 3.356236800252967,
      "tokens_seen": 2919104512
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011662988966900701,
      "loss": 1.3344,
      "theoretical_loss": 3.3562310241240167,
      "tokens_seen": 2919170048
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011660982948846539,
      "loss": 1.3802,
      "theoretical_loss": 3.356225248161049,
      "tokens_seen": 2919235584
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011658976930792377,
      "loss": 1.296,
      "theoretical_loss": 3.3562194723640544,
      "tokens_seen": 2919301120
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011656970912738214,
      "loss": 1.3802,
      "theoretical_loss": 3.356213696733025,
      "tokens_seen": 2919366656
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011654964894684053,
      "loss": 1.1916,
      "theoretical_loss": 3.356207921267952,
      "tokens_seen": 2919432192
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011652958876629891,
      "loss": 1.349,
      "theoretical_loss": 3.356202145968827,
      "tokens_seen": 2919497728
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011650952858575728,
      "loss": 1.2874,
      "theoretical_loss": 3.356196370835642,
      "tokens_seen": 2919563264
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011648946840521566,
      "loss": 1.3145,
      "theoretical_loss": 3.3561905958683877,
      "tokens_seen": 2919628800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011646940822467403,
      "loss": 1.2463,
      "theoretical_loss": 3.356184821067056,
      "tokens_seen": 2919694336
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001164493480441324,
      "loss": 1.3217,
      "theoretical_loss": 3.3561790464316386,
      "tokens_seen": 2919759872
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011642928786359077,
      "loss": 1.27,
      "theoretical_loss": 3.3561732719621267,
      "tokens_seen": 2919825408
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011640922768304915,
      "loss": 1.2888,
      "theoretical_loss": 3.356167497658512,
      "tokens_seen": 2919890944
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011638916750250752,
      "loss": 1.2051,
      "theoretical_loss": 3.356161723520786,
      "tokens_seen": 2919956480
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001163691073219659,
      "loss": 1.2705,
      "theoretical_loss": 3.35615594954894,
      "tokens_seen": 2920022016
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011634904714142428,
      "loss": 1.2671,
      "theoretical_loss": 3.356150175742966,
      "tokens_seen": 2920087552
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011632898696088265,
      "loss": 1.3574,
      "theoretical_loss": 3.3561444021028546,
      "tokens_seen": 2920153088
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011630892678034103,
      "loss": 1.2741,
      "theoretical_loss": 3.3561386286285986,
      "tokens_seen": 2920218624
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001162888665997994,
      "loss": 1.3651,
      "theoretical_loss": 3.356132855320188,
      "tokens_seen": 2920284160
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011626880641925778,
      "loss": 1.3317,
      "theoretical_loss": 3.3561270821776157,
      "tokens_seen": 2920349696
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011624874623871614,
      "loss": 1.3371,
      "theoretical_loss": 3.3561213092008724,
      "tokens_seen": 2920415232
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011622868605817452,
      "loss": 1.3,
      "theoretical_loss": 3.35611553638995,
      "tokens_seen": 2920480768
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011620862587763289,
      "loss": 1.3815,
      "theoretical_loss": 3.35610976374484,
      "tokens_seen": 2920546304
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011618856569709127,
      "loss": 1.2739,
      "theoretical_loss": 3.3561039912655333,
      "tokens_seen": 2920611840
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011616850551654966,
      "loss": 1.3136,
      "theoretical_loss": 3.356098218952022,
      "tokens_seen": 2920677376
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011614844533600803,
      "loss": 1.2431,
      "theoretical_loss": 3.356092446804298,
      "tokens_seen": 2920742912
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011612838515546641,
      "loss": 1.2444,
      "theoretical_loss": 3.3560866748223517,
      "tokens_seen": 2920808448
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011610832497492478,
      "loss": 1.3174,
      "theoretical_loss": 3.3560809030061756,
      "tokens_seen": 2920873984
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011608826479438316,
      "loss": 1.2236,
      "theoretical_loss": 3.3560751313557606,
      "tokens_seen": 2920939520
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011606820461384154,
      "loss": 1.3986,
      "theoretical_loss": 3.356069359871099,
      "tokens_seen": 2921005056
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001160481444332999,
      "loss": 1.2207,
      "theoretical_loss": 3.3560635885521815,
      "tokens_seen": 2921070592
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011602808425275827,
      "loss": 1.3446,
      "theoretical_loss": 3.356057817399,
      "tokens_seen": 2921136128
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011600802407221665,
      "loss": 1.2191,
      "theoretical_loss": 3.3560520464115458,
      "tokens_seen": 2921201664
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011598796389167503,
      "loss": 1.2412,
      "theoretical_loss": 3.3560462755898106,
      "tokens_seen": 2921267200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001159679037111334,
      "loss": 1.23,
      "theoretical_loss": 3.3560405049337865,
      "tokens_seen": 2921332736
    },
    {
      "epoch": 0.77,
      "objective/train/docs_used": 1062024,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.446655511856079,
      "objective/train/theoretical_loss": 3.3560347344434636,
      "objective/train/tokens_used": 1291857376,
      "theoretical_loss": 3.3560347344434636,
      "tokens_seen": 2921398272
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011594784353059178,
      "loss": 1.4145,
      "theoretical_loss": 3.3560347344434636,
      "tokens_seen": 2921398272
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011592778335005015,
      "loss": 1.2557,
      "theoretical_loss": 3.356028964118835,
      "tokens_seen": 2921463808
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011590772316950853,
      "loss": 1.3249,
      "theoretical_loss": 3.356023193959891,
      "tokens_seen": 2921529344
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001158876629889669,
      "loss": 1.237,
      "theoretical_loss": 3.356017423966623,
      "tokens_seen": 2921594880
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011586760280842527,
      "loss": 1.2585,
      "theoretical_loss": 3.3560116541390244,
      "tokens_seen": 2921660416
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011584754262788364,
      "loss": 1.1971,
      "theoretical_loss": 3.3560058844770846,
      "tokens_seen": 2921725952
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011582748244734202,
      "loss": 1.3269,
      "theoretical_loss": 3.3560001149807963,
      "tokens_seen": 2921791488
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001158074222668004,
      "loss": 1.2063,
      "theoretical_loss": 3.3559943456501506,
      "tokens_seen": 2921857024
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011578736208625878,
      "loss": 1.3635,
      "theoretical_loss": 3.3559885764851396,
      "tokens_seen": 2921922560
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011576730190571716,
      "loss": 1.284,
      "theoretical_loss": 3.355982807485754,
      "tokens_seen": 2921988096
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011574724172517554,
      "loss": 1.3149,
      "theoretical_loss": 3.3559770386519854,
      "tokens_seen": 2922053632
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011572718154463391,
      "loss": 1.2923,
      "theoretical_loss": 3.3559712699838258,
      "tokens_seen": 2922119168
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011570712136409229,
      "loss": 1.2851,
      "theoretical_loss": 3.3559655014812666,
      "tokens_seen": 2922184704
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011568706118355066,
      "loss": 1.2962,
      "theoretical_loss": 3.3559597331443,
      "tokens_seen": 2922250240
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011566700100300903,
      "loss": 1.3059,
      "theoretical_loss": 3.355953964972916,
      "tokens_seen": 2922315776
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001156469408224674,
      "loss": 1.2163,
      "theoretical_loss": 3.3559481969671072,
      "tokens_seen": 2922381312
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011562688064192578,
      "loss": 1.1857,
      "theoretical_loss": 3.355942429126865,
      "tokens_seen": 2922446848
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011560682046138415,
      "loss": 1.3108,
      "theoretical_loss": 3.3559366614521804,
      "tokens_seen": 2922512384
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011558676028084253,
      "loss": 1.368,
      "theoretical_loss": 3.355930893943046,
      "tokens_seen": 2922577920
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001155667001003009,
      "loss": 1.3746,
      "theoretical_loss": 3.3559251265994523,
      "tokens_seen": 2922643456
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011554663991975928,
      "loss": 1.2374,
      "theoretical_loss": 3.3559193594213914,
      "tokens_seen": 2922708992
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011552657973921766,
      "loss": 1.3339,
      "theoretical_loss": 3.3559135924088546,
      "tokens_seen": 2922774528
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011550651955867603,
      "loss": 1.3011,
      "theoretical_loss": 3.3559078255618338,
      "tokens_seen": 2922840064
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011548645937813441,
      "loss": 1.2958,
      "theoretical_loss": 3.3559020588803197,
      "tokens_seen": 2922905600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011546639919759277,
      "loss": 1.2747,
      "theoretical_loss": 3.3558962923643048,
      "tokens_seen": 2922971136
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011544633901705115,
      "loss": 1.2239,
      "theoretical_loss": 3.3558905260137806,
      "tokens_seen": 2923036672
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011542627883650953,
      "loss": 1.2791,
      "theoretical_loss": 3.3558847598287374,
      "tokens_seen": 2923102208
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011540621865596791,
      "loss": 1.2826,
      "theoretical_loss": 3.355878993809168,
      "tokens_seen": 2923167744
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011538615847542629,
      "loss": 1.2651,
      "theoretical_loss": 3.355873227955064,
      "tokens_seen": 2923233280
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011536609829488466,
      "loss": 1.2447,
      "theoretical_loss": 3.355867462266416,
      "tokens_seen": 2923298816
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011534603811434304,
      "loss": 1.3747,
      "theoretical_loss": 3.355861696743216,
      "tokens_seen": 2923364352
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011532597793380141,
      "loss": 1.2445,
      "theoretical_loss": 3.355855931385456,
      "tokens_seen": 2923429888
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011530591775325979,
      "loss": 1.2848,
      "theoretical_loss": 3.355850166193127,
      "tokens_seen": 2923495424
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011528585757271815,
      "loss": 1.2731,
      "theoretical_loss": 3.3558444011662205,
      "tokens_seen": 2923560960
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011526579739217653,
      "loss": 1.286,
      "theoretical_loss": 3.3558386363047283,
      "tokens_seen": 2923626496
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001152457372116349,
      "loss": 1.3196,
      "theoretical_loss": 3.3558328716086416,
      "tokens_seen": 2923692032
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011522567703109328,
      "loss": 1.3804,
      "theoretical_loss": 3.355827107077953,
      "tokens_seen": 2923757568
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011520561685055165,
      "loss": 1.3095,
      "theoretical_loss": 3.3558213427126526,
      "tokens_seen": 2923823104
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011518555667001003,
      "loss": 1.2644,
      "theoretical_loss": 3.3558155785127326,
      "tokens_seen": 2923888640
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001151654964894684,
      "loss": 1.2042,
      "theoretical_loss": 3.355809814478185,
      "tokens_seen": 2923954176
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011514543630892678,
      "loss": 1.2809,
      "theoretical_loss": 3.3558040506090006,
      "tokens_seen": 2924019712
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011512537612838516,
      "loss": 1.2034,
      "theoretical_loss": 3.3557982869051712,
      "tokens_seen": 2924085248
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011510531594784353,
      "loss": 1.3033,
      "theoretical_loss": 3.3557925233666888,
      "tokens_seen": 2924150784
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001150852557673019,
      "loss": 1.3918,
      "theoretical_loss": 3.3557867599935443,
      "tokens_seen": 2924216320
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011506519558676027,
      "loss": 1.3332,
      "theoretical_loss": 3.3557809967857293,
      "tokens_seen": 2924281856
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011504513540621866,
      "loss": 1.3195,
      "theoretical_loss": 3.355775233743236,
      "tokens_seen": 2924347392
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011502507522567704,
      "loss": 1.3215,
      "theoretical_loss": 3.355769470866055,
      "tokens_seen": 2924412928
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011500501504513541,
      "loss": 1.4775,
      "theoretical_loss": 3.355763708154179,
      "tokens_seen": 2924478464
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011498495486459379,
      "loss": 1.369,
      "theoretical_loss": 3.355757945607598,
      "tokens_seen": 2924544000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011498495486459379,
      "loss": 1.2242,
      "theoretical_loss": 3.3557521832263055,
      "tokens_seen": 2924609536
    },
    {
      "epoch": 0.77,
      "objective/train/docs_used": 1063229,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2917735576629639,
      "objective/train/theoretical_loss": 3.3557464210102914,
      "objective/train/tokens_used": 1295134176,
      "theoretical_loss": 3.3557464210102914,
      "tokens_seen": 2924675072
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011496489468405216,
      "loss": 1.3079,
      "theoretical_loss": 3.3557464210102914,
      "tokens_seen": 2924675072
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011494483450351054,
      "loss": 1.298,
      "theoretical_loss": 3.355740658959548,
      "tokens_seen": 2924740608
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011492477432296892,
      "loss": 1.3317,
      "theoretical_loss": 3.355734897074067,
      "tokens_seen": 2924806144
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011490471414242729,
      "loss": 1.3744,
      "theoretical_loss": 3.3557291353538394,
      "tokens_seen": 2924871680
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011488465396188565,
      "loss": 1.3396,
      "theoretical_loss": 3.3557233737988574,
      "tokens_seen": 2924937216
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011486459378134403,
      "loss": 1.2951,
      "theoretical_loss": 3.355717612409112,
      "tokens_seen": 2925002752
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011486459378134403,
      "loss": 1.4221,
      "theoretical_loss": 3.355711851184595,
      "tokens_seen": 2925068288
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001148445336008024,
      "loss": 1.4226,
      "theoretical_loss": 3.355706090125298,
      "tokens_seen": 2925133824
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011482447342026078,
      "loss": 1.2511,
      "theoretical_loss": 3.3557003292312126,
      "tokens_seen": 2925199360
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011480441323971916,
      "loss": 1.2314,
      "theoretical_loss": 3.3556945685023303,
      "tokens_seen": 2925264896
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011478435305917753,
      "loss": 1.228,
      "theoretical_loss": 3.355688807938642,
      "tokens_seen": 2925330432
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011476429287863591,
      "loss": 1.2571,
      "theoretical_loss": 3.3556830475401407,
      "tokens_seen": 2925395968
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011474423269809428,
      "loss": 1.2227,
      "theoretical_loss": 3.3556772873068166,
      "tokens_seen": 2925461504
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011472417251755266,
      "loss": 1.3631,
      "theoretical_loss": 3.355671527238662,
      "tokens_seen": 2925527040
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011470411233701104,
      "loss": 1.328,
      "theoretical_loss": 3.3556657673356685,
      "tokens_seen": 2925592576
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001146840521564694,
      "loss": 1.3338,
      "theoretical_loss": 3.355660007597827,
      "tokens_seen": 2925658112
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011466399197592779,
      "loss": 1.3326,
      "theoretical_loss": 3.3556542480251297,
      "tokens_seen": 2925723648
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011464393179538616,
      "loss": 1.3686,
      "theoretical_loss": 3.355648488617568,
      "tokens_seen": 2925789184
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011462387161484454,
      "loss": 1.3599,
      "theoretical_loss": 3.3556427293751336,
      "tokens_seen": 2925854720
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011460381143430292,
      "loss": 1.2965,
      "theoretical_loss": 3.3556369702978177,
      "tokens_seen": 2925920256
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011458375125376129,
      "loss": 1.3105,
      "theoretical_loss": 3.355631211385612,
      "tokens_seen": 2925985792
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011456369107321967,
      "loss": 1.2652,
      "theoretical_loss": 3.3556254526385083,
      "tokens_seen": 2926051328
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011454363089267804,
      "loss": 1.2803,
      "theoretical_loss": 3.355619694056498,
      "tokens_seen": 2926116864
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011452357071213642,
      "loss": 1.3198,
      "theoretical_loss": 3.355613935639573,
      "tokens_seen": 2926182400
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011450351053159478,
      "loss": 1.1843,
      "theoretical_loss": 3.3556081773877238,
      "tokens_seen": 2926247936
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011448345035105316,
      "loss": 1.316,
      "theoretical_loss": 3.355602419300943,
      "tokens_seen": 2926313472
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011446339017051153,
      "loss": 1.329,
      "theoretical_loss": 3.3555966613792223,
      "tokens_seen": 2926379008
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011444332998996991,
      "loss": 1.149,
      "theoretical_loss": 3.3555909036225526,
      "tokens_seen": 2926444544
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011442326980942828,
      "loss": 1.2594,
      "theoretical_loss": 3.3555851460309256,
      "tokens_seen": 2926510080
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011440320962888666,
      "loss": 1.3694,
      "theoretical_loss": 3.3555793886043332,
      "tokens_seen": 2926575616
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011438314944834504,
      "loss": 1.314,
      "theoretical_loss": 3.3555736313427666,
      "tokens_seen": 2926641152
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011436308926780341,
      "loss": 1.2509,
      "theoretical_loss": 3.355567874246218,
      "tokens_seen": 2926706688
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011434302908726179,
      "loss": 1.2499,
      "theoretical_loss": 3.355562117314678,
      "tokens_seen": 2926772224
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011432296890672016,
      "loss": 1.317,
      "theoretical_loss": 3.355556360548139,
      "tokens_seen": 2926837760
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011430290872617853,
      "loss": 1.2198,
      "theoretical_loss": 3.3555506039465923,
      "tokens_seen": 2926903296
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011428284854563691,
      "loss": 1.3026,
      "theoretical_loss": 3.3555448475100293,
      "tokens_seen": 2926968832
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011426278836509529,
      "loss": 1.2603,
      "theoretical_loss": 3.355539091238442,
      "tokens_seen": 2927034368
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011424272818455367,
      "loss": 1.2875,
      "theoretical_loss": 3.3555333351318213,
      "tokens_seen": 2927099904
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011422266800401204,
      "loss": 1.2543,
      "theoretical_loss": 3.3555275791901593,
      "tokens_seen": 2927165440
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011420260782347042,
      "loss": 1.2982,
      "theoretical_loss": 3.3555218234134476,
      "tokens_seen": 2927230976
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001141825476429288,
      "loss": 1.3203,
      "theoretical_loss": 3.3555160678016778,
      "tokens_seen": 2927296512
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011416248746238717,
      "loss": 1.3163,
      "theoretical_loss": 3.3555103123548413,
      "tokens_seen": 2927362048
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011414242728184554,
      "loss": 1.2206,
      "theoretical_loss": 3.35550455707293,
      "tokens_seen": 2927427584
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011412236710130392,
      "loss": 1.1722,
      "theoretical_loss": 3.3554988019559344,
      "tokens_seen": 2927493120
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011410230692076228,
      "loss": 1.3449,
      "theoretical_loss": 3.3554930470038475,
      "tokens_seen": 2927558656
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011408224674022066,
      "loss": 1.2625,
      "theoretical_loss": 3.35548729221666,
      "tokens_seen": 2927624192
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011406218655967903,
      "loss": 1.3185,
      "theoretical_loss": 3.355481537594364,
      "tokens_seen": 2927689728
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011404212637913741,
      "loss": 1.2693,
      "theoretical_loss": 3.3554757831369506,
      "tokens_seen": 2927755264
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011402206619859579,
      "loss": 1.1886,
      "theoretical_loss": 3.3554700288444117,
      "tokens_seen": 2927820800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011400200601805416,
      "loss": 1.2836,
      "theoretical_loss": 3.3554642747167387,
      "tokens_seen": 2927886336
    },
    {
      "epoch": 0.77,
      "objective/train/docs_used": 1064246,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3533320426940918,
      "objective/train/theoretical_loss": 3.3554585207539236,
      "objective/train/tokens_used": 1298410976,
      "theoretical_loss": 3.3554585207539236,
      "tokens_seen": 2927951872
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011398194583751254,
      "loss": 1.2954,
      "theoretical_loss": 3.3554585207539236,
      "tokens_seen": 2927951872
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011396188565697091,
      "loss": 1.3291,
      "theoretical_loss": 3.3554527669559575,
      "tokens_seen": 2928017408
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011394182547642929,
      "loss": 1.2692,
      "theoretical_loss": 3.355447013322832,
      "tokens_seen": 2928082944
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011392176529588768,
      "loss": 1.1988,
      "theoretical_loss": 3.355441259854539,
      "tokens_seen": 2928148480
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011390170511534604,
      "loss": 1.4107,
      "theoretical_loss": 3.3554355065510704,
      "tokens_seen": 2928214016
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011388164493480442,
      "loss": 1.2076,
      "theoretical_loss": 3.355429753412417,
      "tokens_seen": 2928279552
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011386158475426279,
      "loss": 1.1879,
      "theoretical_loss": 3.3554240004385703,
      "tokens_seen": 2928345088
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011384152457372117,
      "loss": 1.271,
      "theoretical_loss": 3.3554182476295225,
      "tokens_seen": 2928410624
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011382146439317954,
      "loss": 1.305,
      "theoretical_loss": 3.3554124949852655,
      "tokens_seen": 2928476160
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011380140421263792,
      "loss": 1.2371,
      "theoretical_loss": 3.3554067425057896,
      "tokens_seen": 2928541696
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001137813440320963,
      "loss": 1.3585,
      "theoretical_loss": 3.3554009901910877,
      "tokens_seen": 2928607232
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011376128385155467,
      "loss": 1.2836,
      "theoretical_loss": 3.355395238041151,
      "tokens_seen": 2928672768
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011374122367101305,
      "loss": 1.2538,
      "theoretical_loss": 3.3553894860559708,
      "tokens_seen": 2928738304
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011372116349047141,
      "loss": 1.2818,
      "theoretical_loss": 3.355383734235539,
      "tokens_seen": 2928803840
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011370110330992979,
      "loss": 1.33,
      "theoretical_loss": 3.355377982579847,
      "tokens_seen": 2928869376
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011368104312938816,
      "loss": 1.2647,
      "theoretical_loss": 3.3553722310888863,
      "tokens_seen": 2928934912
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011366098294884654,
      "loss": 1.3394,
      "theoretical_loss": 3.355366479762649,
      "tokens_seen": 2929000448
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011364092276830491,
      "loss": 1.391,
      "theoretical_loss": 3.355360728601126,
      "tokens_seen": 2929065984
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011362086258776329,
      "loss": 1.2536,
      "theoretical_loss": 3.3553549776043097,
      "tokens_seen": 2929131520
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011360080240722166,
      "loss": 1.2874,
      "theoretical_loss": 3.3553492267721907,
      "tokens_seen": 2929197056
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011358074222668004,
      "loss": 1.1702,
      "theoretical_loss": 3.3553434761047614,
      "tokens_seen": 2929262592
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011356068204613842,
      "loss": 1.3316,
      "theoretical_loss": 3.355337725602013,
      "tokens_seen": 2929328128
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001135406218655968,
      "loss": 1.237,
      "theoretical_loss": 3.3553319752639377,
      "tokens_seen": 2929393664
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011352056168505517,
      "loss": 1.282,
      "theoretical_loss": 3.355326225090526,
      "tokens_seen": 2929459200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011350050150451354,
      "loss": 1.3515,
      "theoretical_loss": 3.3553204750817707,
      "tokens_seen": 2929524736
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011348044132397192,
      "loss": 1.192,
      "theoretical_loss": 3.3553147252376627,
      "tokens_seen": 2929590272
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001134603811434303,
      "loss": 1.2338,
      "theoretical_loss": 3.3553089755581937,
      "tokens_seen": 2929655808
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011344032096288867,
      "loss": 1.2625,
      "theoretical_loss": 3.3553032260433553,
      "tokens_seen": 2929721344
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011342026078234705,
      "loss": 1.3583,
      "theoretical_loss": 3.3552974766931394,
      "tokens_seen": 2929786880
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011340020060180542,
      "loss": 1.263,
      "theoretical_loss": 3.355291727507537,
      "tokens_seen": 2929852416
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001133801404212638,
      "loss": 1.2492,
      "theoretical_loss": 3.35528597848654,
      "tokens_seen": 2929917952
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011336008024072217,
      "loss": 1.1665,
      "theoretical_loss": 3.3552802296301407,
      "tokens_seen": 2929983488
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011334002006018055,
      "loss": 1.3263,
      "theoretical_loss": 3.3552744809383297,
      "tokens_seen": 2930049024
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011331995987963891,
      "loss": 1.326,
      "theoretical_loss": 3.355268732411099,
      "tokens_seen": 2930114560
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011329989969909729,
      "loss": 1.2541,
      "theoretical_loss": 3.3552629840484403,
      "tokens_seen": 2930180096
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011327983951855566,
      "loss": 1.3078,
      "theoretical_loss": 3.355257235850345,
      "tokens_seen": 2930245632
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011325977933801404,
      "loss": 1.1916,
      "theoretical_loss": 3.3552514878168047,
      "tokens_seen": 2930311168
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011323971915747241,
      "loss": 1.2518,
      "theoretical_loss": 3.3552457399478115,
      "tokens_seen": 2930376704
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011321965897693079,
      "loss": 1.2269,
      "theoretical_loss": 3.355239992243356,
      "tokens_seen": 2930442240
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011319959879638917,
      "loss": 1.231,
      "theoretical_loss": 3.355234244703431,
      "tokens_seen": 2930507776
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011317953861584754,
      "loss": 1.395,
      "theoretical_loss": 3.3552284973280275,
      "tokens_seen": 2930573312
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011315947843530593,
      "loss": 1.1969,
      "theoretical_loss": 3.355222750117137,
      "tokens_seen": 2930638848
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011313941825476431,
      "loss": 1.3898,
      "theoretical_loss": 3.3552170030707513,
      "tokens_seen": 2930704384
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011311935807422267,
      "loss": 1.3245,
      "theoretical_loss": 3.3552112561888623,
      "tokens_seen": 2930769920
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011309929789368105,
      "loss": 1.2349,
      "theoretical_loss": 3.355205509471461,
      "tokens_seen": 2930835456
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011307923771313942,
      "loss": 1.3437,
      "theoretical_loss": 3.3551997629185393,
      "tokens_seen": 2930900992
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001130591775325978,
      "loss": 1.271,
      "theoretical_loss": 3.355194016530089,
      "tokens_seen": 2930966528
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011303911735205617,
      "loss": 1.2776,
      "theoretical_loss": 3.355188270306101,
      "tokens_seen": 2931032064
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011301905717151455,
      "loss": 1.2051,
      "theoretical_loss": 3.355182524246568,
      "tokens_seen": 2931097600
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011299899699097292,
      "loss": 1.3442,
      "theoretical_loss": 3.355176778351481,
      "tokens_seen": 2931163136
    },
    {
      "epoch": 0.78,
      "objective/train/docs_used": 1065435,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4551390409469604,
      "objective/train/theoretical_loss": 3.3551710326208317,
      "objective/train/tokens_used": 1301687776,
      "theoretical_loss": 3.3551710326208317,
      "tokens_seen": 2931228672
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001129789368104313,
      "loss": 1.4143,
      "theoretical_loss": 3.3551710326208317,
      "tokens_seen": 2931228672
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011295887662988968,
      "loss": 1.248,
      "theoretical_loss": 3.3551652870546116,
      "tokens_seen": 2931294208
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011293881644934804,
      "loss": 1.3355,
      "theoretical_loss": 3.355159541652813,
      "tokens_seen": 2931359744
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011291875626880641,
      "loss": 1.2533,
      "theoretical_loss": 3.355153796415426,
      "tokens_seen": 2931425280
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011289869608826479,
      "loss": 1.3328,
      "theoretical_loss": 3.355148051342444,
      "tokens_seen": 2931490816
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011287863590772317,
      "loss": 1.2277,
      "theoretical_loss": 3.3551423064338572,
      "tokens_seen": 2931556352
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011285857572718154,
      "loss": 1.1976,
      "theoretical_loss": 3.3551365616896582,
      "tokens_seen": 2931621888
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011283851554663992,
      "loss": 1.3851,
      "theoretical_loss": 3.355130817109838,
      "tokens_seen": 2931687424
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011281845536609829,
      "loss": 1.2478,
      "theoretical_loss": 3.3551250726943884,
      "tokens_seen": 2931752960
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011279839518555668,
      "loss": 1.2291,
      "theoretical_loss": 3.355119328443301,
      "tokens_seen": 2931818496
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011277833500501506,
      "loss": 1.289,
      "theoretical_loss": 3.3551135843565683,
      "tokens_seen": 2931884032
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011275827482447343,
      "loss": 1.3137,
      "theoretical_loss": 3.3551078404341803,
      "tokens_seen": 2931949568
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001127382146439318,
      "loss": 1.395,
      "theoretical_loss": 3.35510209667613,
      "tokens_seen": 2932015104
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011271815446339017,
      "loss": 1.1337,
      "theoretical_loss": 3.3550963530824083,
      "tokens_seen": 2932080640
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011269809428284855,
      "loss": 1.2596,
      "theoretical_loss": 3.355090609653007,
      "tokens_seen": 2932146176
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011267803410230692,
      "loss": 1.2674,
      "theoretical_loss": 3.355084866387917,
      "tokens_seen": 2932211712
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001126579739217653,
      "loss": 1.1698,
      "theoretical_loss": 3.3550791232871315,
      "tokens_seen": 2932277248
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011263791374122368,
      "loss": 1.2975,
      "theoretical_loss": 3.355073380350641,
      "tokens_seen": 2932342784
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011261785356068205,
      "loss": 1.2116,
      "theoretical_loss": 3.3550676375784376,
      "tokens_seen": 2932408320
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011259779338014043,
      "loss": 1.2861,
      "theoretical_loss": 3.3550618949705124,
      "tokens_seen": 2932473856
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001125777331995988,
      "loss": 1.1947,
      "theoretical_loss": 3.3550561525268576,
      "tokens_seen": 2932539392
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011255767301905718,
      "loss": 1.2127,
      "theoretical_loss": 3.3550504102474648,
      "tokens_seen": 2932604928
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011253761283851554,
      "loss": 1.2556,
      "theoretical_loss": 3.355044668132325,
      "tokens_seen": 2932670464
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011251755265797392,
      "loss": 1.3877,
      "theoretical_loss": 3.3550389261814306,
      "tokens_seen": 2932736000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011249749247743229,
      "loss": 1.3004,
      "theoretical_loss": 3.355033184394773,
      "tokens_seen": 2932801536
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011247743229689067,
      "loss": 1.3006,
      "theoretical_loss": 3.355027442772343,
      "tokens_seen": 2932867072
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011245737211634904,
      "loss": 1.1752,
      "theoretical_loss": 3.3550217013141337,
      "tokens_seen": 2932932608
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011243731193580742,
      "loss": 1.2681,
      "theoretical_loss": 3.3550159600201357,
      "tokens_seen": 2932998144
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011241725175526581,
      "loss": 1.1922,
      "theoretical_loss": 3.355010218890341,
      "tokens_seen": 2933063680
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011239719157472418,
      "loss": 1.273,
      "theoretical_loss": 3.355004477924741,
      "tokens_seen": 2933129216
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011237713139418256,
      "loss": 1.2449,
      "theoretical_loss": 3.354998737123328,
      "tokens_seen": 2933194752
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011235707121364094,
      "loss": 1.2818,
      "theoretical_loss": 3.3549929964860925,
      "tokens_seen": 2933260288
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001123370110330993,
      "loss": 1.1807,
      "theoretical_loss": 3.354987256013027,
      "tokens_seen": 2933325824
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011231695085255767,
      "loss": 1.3404,
      "theoretical_loss": 3.354981515704123,
      "tokens_seen": 2933391360
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011229689067201605,
      "loss": 1.3541,
      "theoretical_loss": 3.354975775559372,
      "tokens_seen": 2933456896
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011227683049147443,
      "loss": 1.2335,
      "theoretical_loss": 3.3549700355787655,
      "tokens_seen": 2933522432
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001122567703109328,
      "loss": 1.348,
      "theoretical_loss": 3.3549642957622954,
      "tokens_seen": 2933587968
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011223671013039118,
      "loss": 1.2824,
      "theoretical_loss": 3.3549585561099535,
      "tokens_seen": 2933653504
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011221664994984955,
      "loss": 1.2805,
      "theoretical_loss": 3.354952816621731,
      "tokens_seen": 2933719040
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011219658976930793,
      "loss": 1.2578,
      "theoretical_loss": 3.35494707729762,
      "tokens_seen": 2933784576
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001121765295887663,
      "loss": 1.2495,
      "theoretical_loss": 3.3549413381376114,
      "tokens_seen": 2933850112
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011215646940822467,
      "loss": 1.2604,
      "theoretical_loss": 3.354935599141698,
      "tokens_seen": 2933915648
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011213640922768304,
      "loss": 1.2886,
      "theoretical_loss": 3.3549298603098703,
      "tokens_seen": 2933981184
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011211634904714142,
      "loss": 1.2876,
      "theoretical_loss": 3.3549241216421204,
      "tokens_seen": 2934046720
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001120962888665998,
      "loss": 1.3899,
      "theoretical_loss": 3.35491838313844,
      "tokens_seen": 2934112256
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011207622868605817,
      "loss": 1.3004,
      "theoretical_loss": 3.354912644798821,
      "tokens_seen": 2934177792
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011205616850551655,
      "loss": 1.3066,
      "theoretical_loss": 3.3549069066232544,
      "tokens_seen": 2934243328
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011203610832497494,
      "loss": 1.2423,
      "theoretical_loss": 3.3549011686117325,
      "tokens_seen": 2934308864
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011201604814443331,
      "loss": 1.2383,
      "theoretical_loss": 3.3548954307642465,
      "tokens_seen": 2934374400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011199598796389169,
      "loss": 1.2403,
      "theoretical_loss": 3.3548896930807883,
      "tokens_seen": 2934439936
    },
    {
      "epoch": 0.78,
      "objective/train/docs_used": 1066769,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3130096197128296,
      "objective/train/theoretical_loss": 3.354883955561349,
      "objective/train/tokens_used": 1304964576,
      "theoretical_loss": 3.354883955561349,
      "tokens_seen": 2934505472
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011197592778335006,
      "loss": 1.328,
      "theoretical_loss": 3.354883955561349,
      "tokens_seen": 2934505472
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011195586760280843,
      "loss": 1.3026,
      "theoretical_loss": 3.354878218205921,
      "tokens_seen": 2934571008
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001119358074222668,
      "loss": 1.1792,
      "theoretical_loss": 3.3548724810144956,
      "tokens_seen": 2934636544
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011191574724172518,
      "loss": 1.2153,
      "theoretical_loss": 3.3548667439870643,
      "tokens_seen": 2934702080
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011189568706118355,
      "loss": 1.2689,
      "theoretical_loss": 3.3548610071236196,
      "tokens_seen": 2934767616
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011187562688064193,
      "loss": 1.3382,
      "theoretical_loss": 3.3548552704241517,
      "tokens_seen": 2934833152
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001118555667001003,
      "loss": 1.2594,
      "theoretical_loss": 3.3548495338886535,
      "tokens_seen": 2934898688
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011183550651955868,
      "loss": 1.354,
      "theoretical_loss": 3.354843797517116,
      "tokens_seen": 2934964224
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011181544633901706,
      "loss": 1.2339,
      "theoretical_loss": 3.354838061309531,
      "tokens_seen": 2935029760
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011179538615847543,
      "loss": 1.2885,
      "theoretical_loss": 3.3548323252658903,
      "tokens_seen": 2935095296
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011177532597793381,
      "loss": 1.2015,
      "theoretical_loss": 3.3548265893861853,
      "tokens_seen": 2935160832
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011175526579739217,
      "loss": 1.1984,
      "theoretical_loss": 3.3548208536704083,
      "tokens_seen": 2935226368
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011173520561685055,
      "loss": 1.3061,
      "theoretical_loss": 3.35481511811855,
      "tokens_seen": 2935291904
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011171514543630892,
      "loss": 1.2024,
      "theoretical_loss": 3.3548093827306027,
      "tokens_seen": 2935357440
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001116950852557673,
      "loss": 1.2201,
      "theoretical_loss": 3.3548036475065577,
      "tokens_seen": 2935422976
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011167502507522567,
      "loss": 1.1942,
      "theoretical_loss": 3.354797912446407,
      "tokens_seen": 2935488512
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011165496489468406,
      "loss": 1.1872,
      "theoretical_loss": 3.354792177550142,
      "tokens_seen": 2935554048
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011163490471414244,
      "loss": 1.456,
      "theoretical_loss": 3.3547864428177543,
      "tokens_seen": 2935619584
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011161484453360081,
      "loss": 1.2579,
      "theoretical_loss": 3.354780708249236,
      "tokens_seen": 2935685120
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011159478435305919,
      "loss": 1.3248,
      "theoretical_loss": 3.3547749738445782,
      "tokens_seen": 2935750656
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011157472417251755,
      "loss": 1.3145,
      "theoretical_loss": 3.354769239603773,
      "tokens_seen": 2935816192
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011155466399197593,
      "loss": 1.2613,
      "theoretical_loss": 3.3547635055268117,
      "tokens_seen": 2935881728
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001115346038114343,
      "loss": 1.2289,
      "theoretical_loss": 3.354757771613686,
      "tokens_seen": 2935947264
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011151454363089268,
      "loss": 1.1793,
      "theoretical_loss": 3.354752037864388,
      "tokens_seen": 2936012800
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011149448345035105,
      "loss": 1.205,
      "theoretical_loss": 3.354746304278909,
      "tokens_seen": 2936078336
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011147442326980943,
      "loss": 1.2304,
      "theoretical_loss": 3.354740570857241,
      "tokens_seen": 2936143872
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001114543630892678,
      "loss": 1.1876,
      "theoretical_loss": 3.354734837599375,
      "tokens_seen": 2936209408
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011143430290872618,
      "loss": 1.2994,
      "theoretical_loss": 3.3547291045053034,
      "tokens_seen": 2936274944
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011141424272818456,
      "loss": 1.2276,
      "theoretical_loss": 3.3547233715750173,
      "tokens_seen": 2936340480
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011139418254764293,
      "loss": 1.1639,
      "theoretical_loss": 3.3547176388085083,
      "tokens_seen": 2936406016
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001113741223671013,
      "loss": 1.2459,
      "theoretical_loss": 3.354711906205769,
      "tokens_seen": 2936471552
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011135406218655967,
      "loss": 1.1971,
      "theoretical_loss": 3.35470617376679,
      "tokens_seen": 2936537088
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011133400200601805,
      "loss": 1.2492,
      "theoretical_loss": 3.3547004414915635,
      "tokens_seen": 2936602624
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011131394182547642,
      "loss": 1.3323,
      "theoretical_loss": 3.354694709380081,
      "tokens_seen": 2936668160
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011129388164493481,
      "loss": 1.2684,
      "theoretical_loss": 3.3546889774323345,
      "tokens_seen": 2936733696
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011127382146439319,
      "loss": 1.2434,
      "theoretical_loss": 3.3546832456483147,
      "tokens_seen": 2936799232
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011125376128385156,
      "loss": 1.2675,
      "theoretical_loss": 3.3546775140280145,
      "tokens_seen": 2936864768
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011123370110330994,
      "loss": 1.1926,
      "theoretical_loss": 3.354671782571425,
      "tokens_seen": 2936930304
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011121364092276832,
      "loss": 1.2278,
      "theoretical_loss": 3.354666051278538,
      "tokens_seen": 2936995840
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011119358074222669,
      "loss": 1.2411,
      "theoretical_loss": 3.3546603201493452,
      "tokens_seen": 2937061376
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011117352056168505,
      "loss": 1.3097,
      "theoretical_loss": 3.3546545891838377,
      "tokens_seen": 2937126912
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011115346038114343,
      "loss": 1.2361,
      "theoretical_loss": 3.3546488583820078,
      "tokens_seen": 2937192448
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001111334002006018,
      "loss": 1.2464,
      "theoretical_loss": 3.354643127743847,
      "tokens_seen": 2937257984
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011111334002006018,
      "loss": 1.2934,
      "theoretical_loss": 3.354637397269347,
      "tokens_seen": 2937323520
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011109327983951856,
      "loss": 1.3192,
      "theoretical_loss": 3.3546316669584995,
      "tokens_seen": 2937389056
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011107321965897693,
      "loss": 1.2517,
      "theoretical_loss": 3.354625936811296,
      "tokens_seen": 2937454592
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011105315947843531,
      "loss": 1.3322,
      "theoretical_loss": 3.3546202068277284,
      "tokens_seen": 2937520128
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011103309929789368,
      "loss": 1.3894,
      "theoretical_loss": 3.3546144770077886,
      "tokens_seen": 2937585664
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011101303911735206,
      "loss": 1.2307,
      "theoretical_loss": 3.3546087473514676,
      "tokens_seen": 2937651200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011099297893681042,
      "loss": 1.2067,
      "theoretical_loss": 3.354603017858757,
      "tokens_seen": 2937716736
    },
    {
      "epoch": 0.78,
      "objective/train/docs_used": 1068027,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3222405910491943,
      "objective/train/theoretical_loss": 3.3545972885296496,
      "objective/train/tokens_used": 1308241376,
      "theoretical_loss": 3.3545972885296496,
      "tokens_seen": 2937782272
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001109729187562688,
      "loss": 1.288,
      "theoretical_loss": 3.3545972885296496,
      "tokens_seen": 2937782272
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011095285857572717,
      "loss": 1.3147,
      "theoretical_loss": 3.354591559364136,
      "tokens_seen": 2937847808
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011093279839518555,
      "loss": 1.1894,
      "theoretical_loss": 3.3545858303622085,
      "tokens_seen": 2937913344
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011091273821464394,
      "loss": 1.2842,
      "theoretical_loss": 3.354580101523858,
      "tokens_seen": 2937978880
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011089267803410231,
      "loss": 1.2746,
      "theoretical_loss": 3.3545743728490773,
      "tokens_seen": 2938044416
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011087261785356069,
      "loss": 1.2047,
      "theoretical_loss": 3.3545686443378573,
      "tokens_seen": 2938109952
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011085255767301907,
      "loss": 1.3801,
      "theoretical_loss": 3.35456291599019,
      "tokens_seen": 2938175488
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011083249749247744,
      "loss": 1.2663,
      "theoretical_loss": 3.354557187806067,
      "tokens_seen": 2938241024
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011081243731193582,
      "loss": 1.4094,
      "theoretical_loss": 3.3545514597854798,
      "tokens_seen": 2938306560
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011079237713139418,
      "loss": 1.2556,
      "theoretical_loss": 3.3545457319284204,
      "tokens_seen": 2938372096
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011077231695085256,
      "loss": 1.4124,
      "theoretical_loss": 3.35454000423488,
      "tokens_seen": 2938437632
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011075225677031093,
      "loss": 1.2328,
      "theoretical_loss": 3.354534276704851,
      "tokens_seen": 2938503168
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011073219658976931,
      "loss": 1.313,
      "theoretical_loss": 3.354528549338325,
      "tokens_seen": 2938568704
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011071213640922768,
      "loss": 1.2572,
      "theoretical_loss": 3.3545228221352925,
      "tokens_seen": 2938634240
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011069207622868606,
      "loss": 1.2647,
      "theoretical_loss": 3.354517095095747,
      "tokens_seen": 2938699776
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011067201604814444,
      "loss": 1.3265,
      "theoretical_loss": 3.3545113682196783,
      "tokens_seen": 2938765312
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011065195586760281,
      "loss": 1.2034,
      "theoretical_loss": 3.3545056415070795,
      "tokens_seen": 2938830848
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011063189568706119,
      "loss": 1.2738,
      "theoretical_loss": 3.354499914957942,
      "tokens_seen": 2938896384
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011061183550651956,
      "loss": 1.2061,
      "theoretical_loss": 3.354494188572257,
      "tokens_seen": 2938961920
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011059177532597792,
      "loss": 1.3495,
      "theoretical_loss": 3.354488462350017,
      "tokens_seen": 2939027456
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001105717151454363,
      "loss": 1.2654,
      "theoretical_loss": 3.3544827362912124,
      "tokens_seen": 2939092992
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011055165496489468,
      "loss": 1.2972,
      "theoretical_loss": 3.3544770103958363,
      "tokens_seen": 2939158528
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011053159478435307,
      "loss": 1.1829,
      "theoretical_loss": 3.3544712846638793,
      "tokens_seen": 2939224064
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011051153460381144,
      "loss": 1.2808,
      "theoretical_loss": 3.354465559095334,
      "tokens_seen": 2939289600
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011049147442326982,
      "loss": 1.3403,
      "theoretical_loss": 3.354459833690192,
      "tokens_seen": 2939355136
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011047141424272819,
      "loss": 1.2992,
      "theoretical_loss": 3.3544541084484436,
      "tokens_seen": 2939420672
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011045135406218657,
      "loss": 1.3119,
      "theoretical_loss": 3.3544483833700824,
      "tokens_seen": 2939486208
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011043129388164494,
      "loss": 1.2986,
      "theoretical_loss": 3.354442658455099,
      "tokens_seen": 2939551744
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011041123370110332,
      "loss": 1.2654,
      "theoretical_loss": 3.3544369337034854,
      "tokens_seen": 2939617280
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011039117352056168,
      "loss": 1.3601,
      "theoretical_loss": 3.354431209115233,
      "tokens_seen": 2939682816
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011037111334002006,
      "loss": 1.3127,
      "theoretical_loss": 3.354425484690334,
      "tokens_seen": 2939748352
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011035105315947843,
      "loss": 1.2018,
      "theoretical_loss": 3.35441976042878,
      "tokens_seen": 2939813888
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011033099297893681,
      "loss": 1.3327,
      "theoretical_loss": 3.3544140363305623,
      "tokens_seen": 2939879424
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011031093279839519,
      "loss": 1.2339,
      "theoretical_loss": 3.354408312395673,
      "tokens_seen": 2939944960
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011029087261785356,
      "loss": 1.302,
      "theoretical_loss": 3.3544025886241036,
      "tokens_seen": 2940010496
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011027081243731194,
      "loss": 1.3366,
      "theoretical_loss": 3.3543968650158456,
      "tokens_seen": 2940076032
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011025075225677031,
      "loss": 1.1958,
      "theoretical_loss": 3.354391141570891,
      "tokens_seen": 2940141568
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011023069207622869,
      "loss": 1.2889,
      "theoretical_loss": 3.3543854182892314,
      "tokens_seen": 2940207104
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011021063189568705,
      "loss": 1.2625,
      "theoretical_loss": 3.354379695170859,
      "tokens_seen": 2940272640
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011019057171514543,
      "loss": 1.1784,
      "theoretical_loss": 3.3543739722157646,
      "tokens_seen": 2940338176
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011017051153460382,
      "loss": 1.3156,
      "theoretical_loss": 3.35436824942394,
      "tokens_seen": 2940403712
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011015045135406219,
      "loss": 1.2898,
      "theoretical_loss": 3.354362526795378,
      "tokens_seen": 2940469248
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011013039117352057,
      "loss": 1.1606,
      "theoretical_loss": 3.3543568043300693,
      "tokens_seen": 2940534784
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011011033099297894,
      "loss": 1.2644,
      "theoretical_loss": 3.3543510820280056,
      "tokens_seen": 2940600320
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011009027081243732,
      "loss": 1.2664,
      "theoretical_loss": 3.354345359889179,
      "tokens_seen": 2940665856
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001100702106318957,
      "loss": 1.2643,
      "theoretical_loss": 3.354339637913581,
      "tokens_seen": 2940731392
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011005015045135407,
      "loss": 1.2751,
      "theoretical_loss": 3.3543339161012034,
      "tokens_seen": 2940796928
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011003009027081245,
      "loss": 1.2449,
      "theoretical_loss": 3.354328194452038,
      "tokens_seen": 2940862464
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011001003009027081,
      "loss": 1.2916,
      "theoretical_loss": 3.354322472966076,
      "tokens_seen": 2940928000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010998996990972919,
      "loss": 1.2894,
      "theoretical_loss": 3.35431675164331,
      "tokens_seen": 2940993536
    },
    {
      "epoch": 0.78,
      "objective/train/docs_used": 1069310,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2102913856506348,
      "objective/train/theoretical_loss": 3.354311030483731,
      "objective/train/tokens_used": 1311518176,
      "theoretical_loss": 3.354311030483731,
      "tokens_seen": 2941059072
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010996990972918756,
      "loss": 1.2982,
      "theoretical_loss": 3.354311030483731,
      "tokens_seen": 2941059072
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010994984954864594,
      "loss": 1.2014,
      "theoretical_loss": 3.354305309487331,
      "tokens_seen": 2941124608
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010992978936810431,
      "loss": 1.232,
      "theoretical_loss": 3.354299588654101,
      "tokens_seen": 2941190144
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010990972918756269,
      "loss": 1.3297,
      "theoretical_loss": 3.354293867984034,
      "tokens_seen": 2941255680
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010988966900702106,
      "loss": 1.2724,
      "theoretical_loss": 3.3542881474771207,
      "tokens_seen": 2941321216
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010986960882647944,
      "loss": 1.2637,
      "theoretical_loss": 3.3542824271333536,
      "tokens_seen": 2941386752
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010984954864593782,
      "loss": 1.2578,
      "theoretical_loss": 3.3542767069527235,
      "tokens_seen": 2941452288
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010982948846539619,
      "loss": 1.3669,
      "theoretical_loss": 3.354270986935223,
      "tokens_seen": 2941517824
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010980942828485455,
      "loss": 1.2827,
      "theoretical_loss": 3.354265267080843,
      "tokens_seen": 2941583360
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010978936810431294,
      "loss": 1.3365,
      "theoretical_loss": 3.3542595473895758,
      "tokens_seen": 2941648896
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010976930792377132,
      "loss": 1.255,
      "theoretical_loss": 3.3542538278614127,
      "tokens_seen": 2941714432
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001097492477432297,
      "loss": 1.2034,
      "theoretical_loss": 3.354248108496346,
      "tokens_seen": 2941779968
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010972918756268807,
      "loss": 1.1882,
      "theoretical_loss": 3.3542423892943667,
      "tokens_seen": 2941845504
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010970912738214645,
      "loss": 1.2152,
      "theoretical_loss": 3.354236670255467,
      "tokens_seen": 2941911040
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010968906720160482,
      "loss": 1.3105,
      "theoretical_loss": 3.3542309513796384,
      "tokens_seen": 2941976576
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001096690070210632,
      "loss": 1.3013,
      "theoretical_loss": 3.354225232666873,
      "tokens_seen": 2942042112
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010964894684052157,
      "loss": 1.3088,
      "theoretical_loss": 3.354219514117162,
      "tokens_seen": 2942107648
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010962888665997995,
      "loss": 1.3586,
      "theoretical_loss": 3.354213795730497,
      "tokens_seen": 2942173184
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010960882647943831,
      "loss": 1.3207,
      "theoretical_loss": 3.3542080775068706,
      "tokens_seen": 2942238720
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010958876629889669,
      "loss": 1.2336,
      "theoretical_loss": 3.354202359446274,
      "tokens_seen": 2942304256
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010956870611835506,
      "loss": 1.3287,
      "theoretical_loss": 3.3541966415486986,
      "tokens_seen": 2942369792
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010954864593781344,
      "loss": 1.1451,
      "theoretical_loss": 3.3541909238141363,
      "tokens_seen": 2942435328
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010952858575727181,
      "loss": 1.251,
      "theoretical_loss": 3.354185206242579,
      "tokens_seen": 2942500864
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010950852557673019,
      "loss": 1.2135,
      "theoretical_loss": 3.354179488834019,
      "tokens_seen": 2942566400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010948846539618857,
      "loss": 1.2855,
      "theoretical_loss": 3.3541737715884468,
      "tokens_seen": 2942631936
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010946840521564694,
      "loss": 1.3289,
      "theoretical_loss": 3.3541680545058545,
      "tokens_seen": 2942697472
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010944834503510532,
      "loss": 1.237,
      "theoretical_loss": 3.3541623375862346,
      "tokens_seen": 2942763008
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010942828485456368,
      "loss": 1.302,
      "theoretical_loss": 3.354156620829578,
      "tokens_seen": 2942828544
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010940822467402207,
      "loss": 1.2203,
      "theoretical_loss": 3.3541509042358766,
      "tokens_seen": 2942894080
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010938816449348045,
      "loss": 1.2499,
      "theoretical_loss": 3.3541451878051225,
      "tokens_seen": 2942959616
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010936810431293882,
      "loss": 1.183,
      "theoretical_loss": 3.354139471537307,
      "tokens_seen": 2943025152
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001093480441323972,
      "loss": 1.2639,
      "theoretical_loss": 3.354133755432422,
      "tokens_seen": 2943090688
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010932798395185557,
      "loss": 1.2803,
      "theoretical_loss": 3.354128039490459,
      "tokens_seen": 2943156224
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010930792377131395,
      "loss": 1.2005,
      "theoretical_loss": 3.35412232371141,
      "tokens_seen": 2943221760
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010928786359077232,
      "loss": 1.3162,
      "theoretical_loss": 3.354116608095267,
      "tokens_seen": 2943287296
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001092678034102307,
      "loss": 1.4113,
      "theoretical_loss": 3.3541108926420207,
      "tokens_seen": 2943352832
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010924774322968908,
      "loss": 1.2496,
      "theoretical_loss": 3.354105177351664,
      "tokens_seen": 2943418368
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010922768304914744,
      "loss": 1.201,
      "theoretical_loss": 3.354099462224188,
      "tokens_seen": 2943483904
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010920762286860581,
      "loss": 1.3392,
      "theoretical_loss": 3.3540937472595846,
      "tokens_seen": 2943549440
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010918756268806419,
      "loss": 1.3501,
      "theoretical_loss": 3.3540880324578453,
      "tokens_seen": 2943614976
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010916750250752257,
      "loss": 1.2198,
      "theoretical_loss": 3.3540823178189623,
      "tokens_seen": 2943680512
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010914744232698094,
      "loss": 1.2893,
      "theoretical_loss": 3.3540766033429272,
      "tokens_seen": 2943746048
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010912738214643932,
      "loss": 1.235,
      "theoretical_loss": 3.3540708890297317,
      "tokens_seen": 2943811584
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010910732196589769,
      "loss": 1.2933,
      "theoretical_loss": 3.3540651748793673,
      "tokens_seen": 2943877120
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010908726178535607,
      "loss": 1.3418,
      "theoretical_loss": 3.3540594608918255,
      "tokens_seen": 2943942656
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010906720160481444,
      "loss": 1.4109,
      "theoretical_loss": 3.354053747067099,
      "tokens_seen": 2944008192
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010904714142427282,
      "loss": 1.3483,
      "theoretical_loss": 3.3540480334051783,
      "tokens_seen": 2944073728
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001090270812437312,
      "loss": 1.3063,
      "theoretical_loss": 3.354042319906056,
      "tokens_seen": 2944139264
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010900702106318957,
      "loss": 1.2426,
      "theoretical_loss": 3.354036606569724,
      "tokens_seen": 2944204800
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010898696088264795,
      "loss": 1.281,
      "theoretical_loss": 3.3540308933961733,
      "tokens_seen": 2944270336
    },
    {
      "epoch": 0.78,
      "objective/train/docs_used": 1070538,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4315128326416016,
      "objective/train/theoretical_loss": 3.3540251803853964,
      "objective/train/tokens_used": 1314794976,
      "theoretical_loss": 3.3540251803853964,
      "tokens_seen": 2944335872
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010896690070210632,
      "loss": 1.4296,
      "theoretical_loss": 3.3540251803853964,
      "tokens_seen": 2944335872
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001089468405215647,
      "loss": 1.3117,
      "theoretical_loss": 3.354019467537384,
      "tokens_seen": 2944401408
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010892678034102307,
      "loss": 1.3662,
      "theoretical_loss": 3.354013754852129,
      "tokens_seen": 2944466944
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010890672016048145,
      "loss": 1.4459,
      "theoretical_loss": 3.354008042329623,
      "tokens_seen": 2944532480
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010888665997993983,
      "loss": 1.2789,
      "theoretical_loss": 3.3540023299698563,
      "tokens_seen": 2944598016
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001088665997993982,
      "loss": 1.2728,
      "theoretical_loss": 3.353996617772822,
      "tokens_seen": 2944663552
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010884653961885658,
      "loss": 1.371,
      "theoretical_loss": 3.353990905738512,
      "tokens_seen": 2944729088
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010882647943831494,
      "loss": 1.2207,
      "theoretical_loss": 3.353985193866918,
      "tokens_seen": 2944794624
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010880641925777332,
      "loss": 1.2757,
      "theoretical_loss": 3.3539794821580307,
      "tokens_seen": 2944860160
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010878635907723169,
      "loss": 1.3555,
      "theoretical_loss": 3.353973770611842,
      "tokens_seen": 2944925696
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010876629889669007,
      "loss": 1.218,
      "theoretical_loss": 3.353968059228345,
      "tokens_seen": 2944991232
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010874623871614844,
      "loss": 1.2394,
      "theoretical_loss": 3.35396234800753,
      "tokens_seen": 2945056768
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010872617853560682,
      "loss": 1.3426,
      "theoretical_loss": 3.3539566369493894,
      "tokens_seen": 2945122304
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001087061183550652,
      "loss": 1.303,
      "theoretical_loss": 3.3539509260539155,
      "tokens_seen": 2945187840
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010868605817452357,
      "loss": 1.3722,
      "theoretical_loss": 3.3539452153210987,
      "tokens_seen": 2945253376
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010866599799398196,
      "loss": 1.3092,
      "theoretical_loss": 3.3539395047509317,
      "tokens_seen": 2945318912
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010864593781344032,
      "loss": 1.3246,
      "theoretical_loss": 3.3539337943434058,
      "tokens_seen": 2945384448
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001086258776328987,
      "loss": 1.2342,
      "theoretical_loss": 3.3539280840985133,
      "tokens_seen": 2945449984
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010860581745235707,
      "loss": 1.2874,
      "theoretical_loss": 3.3539223740162454,
      "tokens_seen": 2945515520
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010858575727181545,
      "loss": 1.2455,
      "theoretical_loss": 3.353916664096594,
      "tokens_seen": 2945581056
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010856569709127383,
      "loss": 1.2538,
      "theoretical_loss": 3.3539109543395513,
      "tokens_seen": 2945646592
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001085456369107322,
      "loss": 1.3126,
      "theoretical_loss": 3.3539052447451083,
      "tokens_seen": 2945712128
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010852557673019058,
      "loss": 1.2985,
      "theoretical_loss": 3.353899535313257,
      "tokens_seen": 2945777664
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010850551654964895,
      "loss": 1.1916,
      "theoretical_loss": 3.35389382604399,
      "tokens_seen": 2945843200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010848545636910733,
      "loss": 1.179,
      "theoretical_loss": 3.3538881169372976,
      "tokens_seen": 2945908736
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001084653961885657,
      "loss": 1.2687,
      "theoretical_loss": 3.353882407993172,
      "tokens_seen": 2945974272
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010844533600802407,
      "loss": 1.3906,
      "theoretical_loss": 3.353876699211606,
      "tokens_seen": 2946039808
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010842527582748244,
      "loss": 1.2713,
      "theoretical_loss": 3.3538709905925903,
      "tokens_seen": 2946105344
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010840521564694082,
      "loss": 1.2755,
      "theoretical_loss": 3.353865282136117,
      "tokens_seen": 2946170880
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001083851554663992,
      "loss": 1.2639,
      "theoretical_loss": 3.3538595738421777,
      "tokens_seen": 2946236416
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010836509528585757,
      "loss": 1.3521,
      "theoretical_loss": 3.3538538657107644,
      "tokens_seen": 2946301952
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010834503510531595,
      "loss": 1.2594,
      "theoretical_loss": 3.3538481577418686,
      "tokens_seen": 2946367488
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010832497492477432,
      "loss": 1.3191,
      "theoretical_loss": 3.353842449935482,
      "tokens_seen": 2946433024
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001083049147442327,
      "loss": 1.2137,
      "theoretical_loss": 3.3538367422915965,
      "tokens_seen": 2946498560
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010828485456369109,
      "loss": 1.247,
      "theoretical_loss": 3.353831034810204,
      "tokens_seen": 2946564096
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010826479438314946,
      "loss": 1.2901,
      "theoretical_loss": 3.3538253274912964,
      "tokens_seen": 2946629632
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010824473420260782,
      "loss": 1.2819,
      "theoretical_loss": 3.353819620334865,
      "tokens_seen": 2946695168
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001082246740220662,
      "loss": 1.3852,
      "theoretical_loss": 3.3538139133409017,
      "tokens_seen": 2946760704
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010820461384152458,
      "loss": 1.2125,
      "theoretical_loss": 3.3538082065093984,
      "tokens_seen": 2946826240
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010818455366098295,
      "loss": 1.3135,
      "theoretical_loss": 3.353802499840347,
      "tokens_seen": 2946891776
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010816449348044133,
      "loss": 1.1829,
      "theoretical_loss": 3.3537967933337387,
      "tokens_seen": 2946957312
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001081444332998997,
      "loss": 1.3341,
      "theoretical_loss": 3.353791086989566,
      "tokens_seen": 2947022848
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010812437311935808,
      "loss": 1.2486,
      "theoretical_loss": 3.3537853808078197,
      "tokens_seen": 2947088384
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010810431293881646,
      "loss": 1.2929,
      "theoretical_loss": 3.353779674788493,
      "tokens_seen": 2947153920
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010808425275827483,
      "loss": 1.2557,
      "theoretical_loss": 3.353773968931576,
      "tokens_seen": 2947219456
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010806419257773321,
      "loss": 1.1678,
      "theoretical_loss": 3.3537682632370616,
      "tokens_seen": 2947284992
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010804413239719157,
      "loss": 1.2282,
      "theoretical_loss": 3.3537625577049415,
      "tokens_seen": 2947350528
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010802407221664995,
      "loss": 1.2543,
      "theoretical_loss": 3.3537568523352066,
      "tokens_seen": 2947416064
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010800401203610832,
      "loss": 1.2732,
      "theoretical_loss": 3.35375114712785,
      "tokens_seen": 2947481600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001079839518555667,
      "loss": 1.2233,
      "theoretical_loss": 3.353745442082862,
      "tokens_seen": 2947547136
    },
    {
      "epoch": 0.79,
      "objective/train/docs_used": 1071710,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.390841007232666,
      "objective/train/theoretical_loss": 3.3537397372002355,
      "objective/train/tokens_used": 1318071776,
      "theoretical_loss": 3.3537397372002355,
      "tokens_seen": 2947612672
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010796389167502507,
      "loss": 1.374,
      "theoretical_loss": 3.3537397372002355,
      "tokens_seen": 2947612672
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010794383149448345,
      "loss": 1.1638,
      "theoretical_loss": 3.353734032479962,
      "tokens_seen": 2947678208
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010792377131394182,
      "loss": 1.2187,
      "theoretical_loss": 3.353728327922033,
      "tokens_seen": 2947743744
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010790371113340021,
      "loss": 1.2204,
      "theoretical_loss": 3.3537226235264406,
      "tokens_seen": 2947809280
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010788365095285859,
      "loss": 1.3467,
      "theoretical_loss": 3.353716919293176,
      "tokens_seen": 2947874816
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010786359077231695,
      "loss": 1.2073,
      "theoretical_loss": 3.3537112152222317,
      "tokens_seen": 2947940352
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010784353059177533,
      "loss": 1.2159,
      "theoretical_loss": 3.353705511313599,
      "tokens_seen": 2948005888
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001078234704112337,
      "loss": 1.285,
      "theoretical_loss": 3.35369980756727,
      "tokens_seen": 2948071424
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010780341023069208,
      "loss": 1.2636,
      "theoretical_loss": 3.3536941039832366,
      "tokens_seen": 2948136960
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010778335005015045,
      "loss": 1.3616,
      "theoretical_loss": 3.3536884005614898,
      "tokens_seen": 2948202496
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010776328986960883,
      "loss": 1.2919,
      "theoretical_loss": 3.353682697302022,
      "tokens_seen": 2948268032
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001077432296890672,
      "loss": 1.3484,
      "theoretical_loss": 3.3536769942048243,
      "tokens_seen": 2948333568
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010772316950852558,
      "loss": 1.2979,
      "theoretical_loss": 3.35367129126989,
      "tokens_seen": 2948399104
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010770310932798396,
      "loss": 1.2565,
      "theoretical_loss": 3.353665588497209,
      "tokens_seen": 2948464640
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010768304914744233,
      "loss": 1.2197,
      "theoretical_loss": 3.353659885886774,
      "tokens_seen": 2948530176
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001076629889669007,
      "loss": 1.3412,
      "theoretical_loss": 3.3536541834385774,
      "tokens_seen": 2948595712
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010764292878635907,
      "loss": 1.2059,
      "theoretical_loss": 3.35364848115261,
      "tokens_seen": 2948661248
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010762286860581745,
      "loss": 1.2645,
      "theoretical_loss": 3.3536427790288634,
      "tokens_seen": 2948726784
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010760280842527582,
      "loss": 1.2753,
      "theoretical_loss": 3.3536370770673303,
      "tokens_seen": 2948792320
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001075827482447342,
      "loss": 1.2142,
      "theoretical_loss": 3.353631375268002,
      "tokens_seen": 2948857856
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010756268806419257,
      "loss": 1.1472,
      "theoretical_loss": 3.3536256736308703,
      "tokens_seen": 2948923392
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010754262788365096,
      "loss": 1.3431,
      "theoretical_loss": 3.353619972155927,
      "tokens_seen": 2948988928
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010752256770310934,
      "loss": 1.2154,
      "theoretical_loss": 3.3536142708431638,
      "tokens_seen": 2949054464
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010750250752256772,
      "loss": 1.5054,
      "theoretical_loss": 3.3536085696925726,
      "tokens_seen": 2949120000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010748244734202609,
      "loss": 1.3101,
      "theoretical_loss": 3.3536028687041455,
      "tokens_seen": 2949185536
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010746238716148445,
      "loss": 1.2374,
      "theoretical_loss": 3.3535971678778735,
      "tokens_seen": 2949251072
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010744232698094283,
      "loss": 1.1507,
      "theoretical_loss": 3.3535914672137492,
      "tokens_seen": 2949316608
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001074222668004012,
      "loss": 1.38,
      "theoretical_loss": 3.3535857667117637,
      "tokens_seen": 2949382144
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010740220661985958,
      "loss": 1.2136,
      "theoretical_loss": 3.353580066371909,
      "tokens_seen": 2949447680
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010738214643931796,
      "loss": 1.192,
      "theoretical_loss": 3.3535743661941773,
      "tokens_seen": 2949513216
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010736208625877633,
      "loss": 1.3322,
      "theoretical_loss": 3.35356866617856,
      "tokens_seen": 2949578752
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010734202607823471,
      "loss": 1.1165,
      "theoretical_loss": 3.3535629663250486,
      "tokens_seen": 2949644288
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010732196589769308,
      "loss": 1.3117,
      "theoretical_loss": 3.3535572666336355,
      "tokens_seen": 2949709824
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010730190571715146,
      "loss": 1.3075,
      "theoretical_loss": 3.3535515671043123,
      "tokens_seen": 2949775360
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010728184553660982,
      "loss": 1.3242,
      "theoretical_loss": 3.353545867737071,
      "tokens_seen": 2949840896
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001072617853560682,
      "loss": 1.1797,
      "theoretical_loss": 3.353540168531903,
      "tokens_seen": 2949906432
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010724172517552657,
      "loss": 1.2302,
      "theoretical_loss": 3.3535344694887996,
      "tokens_seen": 2949971968
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010722166499498495,
      "loss": 1.2498,
      "theoretical_loss": 3.353528770607754,
      "tokens_seen": 2950037504
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010720160481444333,
      "loss": 1.1837,
      "theoretical_loss": 3.3535230718887563,
      "tokens_seen": 2950103040
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001071815446339017,
      "loss": 1.1967,
      "theoretical_loss": 3.3535173733318,
      "tokens_seen": 2950168576
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010716148445336009,
      "loss": 1.2354,
      "theoretical_loss": 3.3535116749368754,
      "tokens_seen": 2950234112
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010714142427281847,
      "loss": 1.3625,
      "theoretical_loss": 3.3535059767039757,
      "tokens_seen": 2950299648
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010712136409227684,
      "loss": 1.243,
      "theoretical_loss": 3.353500278633091,
      "tokens_seen": 2950365184
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010710130391173522,
      "loss": 1.2154,
      "theoretical_loss": 3.353494580724215,
      "tokens_seen": 2950430720
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010708124373119358,
      "loss": 1.2859,
      "theoretical_loss": 3.353488882977338,
      "tokens_seen": 2950496256
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010706118355065196,
      "loss": 1.3003,
      "theoretical_loss": 3.3534831853924527,
      "tokens_seen": 2950561792
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010704112337011033,
      "loss": 1.2969,
      "theoretical_loss": 3.3534774879695504,
      "tokens_seen": 2950627328
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010702106318956871,
      "loss": 1.2291,
      "theoretical_loss": 3.353471790708623,
      "tokens_seen": 2950692864
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010700100300902708,
      "loss": 1.2578,
      "theoretical_loss": 3.3534660936096623,
      "tokens_seen": 2950758400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010698094282848546,
      "loss": 1.2862,
      "theoretical_loss": 3.35346039667266,
      "tokens_seen": 2950823936
    },
    {
      "epoch": 0.79,
      "objective/train/docs_used": 1072835,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2957708835601807,
      "objective/train/theoretical_loss": 3.3534546998976085,
      "objective/train/tokens_used": 1321348576,
      "theoretical_loss": 3.3534546998976085,
      "tokens_seen": 2950889472
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010696088264794383,
      "loss": 1.3347,
      "theoretical_loss": 3.3534546998976085,
      "tokens_seen": 2950889472
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010694082246740221,
      "loss": 1.2847,
      "theoretical_loss": 3.3534490032844984,
      "tokens_seen": 2950955008
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010692076228686059,
      "loss": 1.2933,
      "theoretical_loss": 3.3534433068333227,
      "tokens_seen": 2951020544
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010690070210631896,
      "loss": 1.2119,
      "theoretical_loss": 3.3534376105440726,
      "tokens_seen": 2951086080
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010688064192577732,
      "loss": 1.3079,
      "theoretical_loss": 3.35343191441674,
      "tokens_seen": 2951151616
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001068605817452357,
      "loss": 1.2319,
      "theoretical_loss": 3.3534262184513173,
      "tokens_seen": 2951217152
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010684052156469408,
      "loss": 1.2377,
      "theoretical_loss": 3.353420522647795,
      "tokens_seen": 2951282688
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010682046138415245,
      "loss": 1.311,
      "theoretical_loss": 3.353414827006166,
      "tokens_seen": 2951348224
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010680040120361083,
      "loss": 1.3401,
      "theoretical_loss": 3.353409131526422,
      "tokens_seen": 2951413760
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010678034102306922,
      "loss": 1.2281,
      "theoretical_loss": 3.353403436208554,
      "tokens_seen": 2951479296
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010676028084252759,
      "loss": 1.2802,
      "theoretical_loss": 3.3533977410525546,
      "tokens_seen": 2951544832
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010674022066198597,
      "loss": 1.2435,
      "theoretical_loss": 3.3533920460584152,
      "tokens_seen": 2951610368
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010672016048144434,
      "loss": 1.3664,
      "theoretical_loss": 3.353386351226128,
      "tokens_seen": 2951675904
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010670010030090272,
      "loss": 1.2067,
      "theoretical_loss": 3.3533806565556845,
      "tokens_seen": 2951741440
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010668004012036108,
      "loss": 1.2811,
      "theoretical_loss": 3.3533749620470767,
      "tokens_seen": 2951806976
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010665997993981946,
      "loss": 1.2928,
      "theoretical_loss": 3.3533692677002964,
      "tokens_seen": 2951872512
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010663991975927783,
      "loss": 1.2573,
      "theoretical_loss": 3.3533635735153347,
      "tokens_seen": 2951938048
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010661985957873621,
      "loss": 1.2763,
      "theoretical_loss": 3.3533578794921843,
      "tokens_seen": 2952003584
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010659979939819459,
      "loss": 1.3431,
      "theoretical_loss": 3.353352185630837,
      "tokens_seen": 2952069120
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010657973921765296,
      "loss": 1.2606,
      "theoretical_loss": 3.3533464919312843,
      "tokens_seen": 2952134656
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010655967903711134,
      "loss": 1.203,
      "theoretical_loss": 3.3533407983935177,
      "tokens_seen": 2952200192
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010653961885656971,
      "loss": 1.3478,
      "theoretical_loss": 3.3533351050175293,
      "tokens_seen": 2952265728
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010651955867602809,
      "loss": 1.3264,
      "theoretical_loss": 3.3533294118033115,
      "tokens_seen": 2952331264
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010649949849548645,
      "loss": 1.2222,
      "theoretical_loss": 3.353323718750855,
      "tokens_seen": 2952396800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010647943831494483,
      "loss": 1.188,
      "theoretical_loss": 3.3533180258601525,
      "tokens_seen": 2952462336
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001064593781344032,
      "loss": 1.2573,
      "theoretical_loss": 3.3533123331311954,
      "tokens_seen": 2952527872
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010643931795386158,
      "loss": 1.3127,
      "theoretical_loss": 3.3533066405639755,
      "tokens_seen": 2952593408
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010641925777331995,
      "loss": 1.2041,
      "theoretical_loss": 3.353300948158485,
      "tokens_seen": 2952658944
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010639919759277834,
      "loss": 1.2584,
      "theoretical_loss": 3.3532952559147153,
      "tokens_seen": 2952724480
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010637913741223672,
      "loss": 1.3646,
      "theoretical_loss": 3.3532895638326585,
      "tokens_seen": 2952790016
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001063590772316951,
      "loss": 1.3156,
      "theoretical_loss": 3.353283871912306,
      "tokens_seen": 2952855552
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010633901705115347,
      "loss": 1.2683,
      "theoretical_loss": 3.35327818015365,
      "tokens_seen": 2952921088
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010631895687061185,
      "loss": 1.3719,
      "theoretical_loss": 3.3532724885566827,
      "tokens_seen": 2952986624
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010629889669007021,
      "loss": 1.3105,
      "theoretical_loss": 3.3532667971213947,
      "tokens_seen": 2953052160
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010627883650952858,
      "loss": 1.3171,
      "theoretical_loss": 3.353261105847779,
      "tokens_seen": 2953117696
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010625877632898696,
      "loss": 1.1772,
      "theoretical_loss": 3.353255414735827,
      "tokens_seen": 2953183232
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010623871614844534,
      "loss": 1.2061,
      "theoretical_loss": 3.35324972378553,
      "tokens_seen": 2953248768
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010621865596790371,
      "loss": 1.2777,
      "theoretical_loss": 3.353244032996881,
      "tokens_seen": 2953314304
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010619859578736209,
      "loss": 1.3758,
      "theoretical_loss": 3.3532383423698704,
      "tokens_seen": 2953379840
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010617853560682046,
      "loss": 1.155,
      "theoretical_loss": 3.3532326519044915,
      "tokens_seen": 2953445376
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010615847542627884,
      "loss": 1.3682,
      "theoretical_loss": 3.353226961600735,
      "tokens_seen": 2953510912
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010613841524573722,
      "loss": 1.2182,
      "theoretical_loss": 3.353221271458593,
      "tokens_seen": 2953576448
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010611835506519559,
      "loss": 1.3615,
      "theoretical_loss": 3.3532155814780573,
      "tokens_seen": 2953641984
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010609829488465395,
      "loss": 1.2202,
      "theoretical_loss": 3.3532098916591204,
      "tokens_seen": 2953707520
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010607823470411233,
      "loss": 1.3027,
      "theoretical_loss": 3.353204202001773,
      "tokens_seen": 2953773056
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001060581745235707,
      "loss": 1.2408,
      "theoretical_loss": 3.3531985125060078,
      "tokens_seen": 2953838592
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001060381143430291,
      "loss": 1.426,
      "theoretical_loss": 3.3531928231718164,
      "tokens_seen": 2953904128
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010601805416248747,
      "loss": 1.3704,
      "theoretical_loss": 3.3531871339991906,
      "tokens_seen": 2953969664
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010599799398194585,
      "loss": 1.3166,
      "theoretical_loss": 3.353181444988122,
      "tokens_seen": 2954035200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010597793380140422,
      "loss": 1.285,
      "theoretical_loss": 3.3531757561386026,
      "tokens_seen": 2954100736
    },
    {
      "epoch": 0.79,
      "objective/train/docs_used": 1074062,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.253406047821045,
      "objective/train/theoretical_loss": 3.353170067450624,
      "objective/train/tokens_used": 1324625376,
      "theoretical_loss": 3.353170067450624,
      "tokens_seen": 2954166272
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001059578736208626,
      "loss": 1.2778,
      "theoretical_loss": 3.353170067450624,
      "tokens_seen": 2954166272
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010593781344032097,
      "loss": 1.2449,
      "theoretical_loss": 3.3531643789241787,
      "tokens_seen": 2954231808
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010591775325977935,
      "loss": 1.2636,
      "theoretical_loss": 3.3531586905592583,
      "tokens_seen": 2954297344
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010589769307923771,
      "loss": 1.2372,
      "theoretical_loss": 3.353153002355854,
      "tokens_seen": 2954362880
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010587763289869609,
      "loss": 1.2658,
      "theoretical_loss": 3.3531473143139583,
      "tokens_seen": 2954428416
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010585757271815446,
      "loss": 1.3216,
      "theoretical_loss": 3.3531416264335627,
      "tokens_seen": 2954493952
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010583751253761284,
      "loss": 1.1897,
      "theoretical_loss": 3.353135938714659,
      "tokens_seen": 2954559488
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010581745235707121,
      "loss": 1.3416,
      "theoretical_loss": 3.353130251157239,
      "tokens_seen": 2954625024
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010579739217652959,
      "loss": 1.2818,
      "theoretical_loss": 3.353124563761295,
      "tokens_seen": 2954690560
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010577733199598797,
      "loss": 1.3515,
      "theoretical_loss": 3.3531188765268185,
      "tokens_seen": 2954756096
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010575727181544634,
      "loss": 1.2602,
      "theoretical_loss": 3.3531131894538015,
      "tokens_seen": 2954821632
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010573721163490472,
      "loss": 1.3204,
      "theoretical_loss": 3.353107502542236,
      "tokens_seen": 2954887168
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010571715145436308,
      "loss": 1.2992,
      "theoretical_loss": 3.3531018157921126,
      "tokens_seen": 2954952704
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010569709127382146,
      "loss": 1.2965,
      "theoretical_loss": 3.3530961292034247,
      "tokens_seen": 2955018240
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010567703109327983,
      "loss": 1.2707,
      "theoretical_loss": 3.353090442776163,
      "tokens_seen": 2955083776
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010565697091273822,
      "loss": 1.2556,
      "theoretical_loss": 3.3530847565103206,
      "tokens_seen": 2955149312
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001056369107321966,
      "loss": 1.332,
      "theoretical_loss": 3.3530790704058884,
      "tokens_seen": 2955214848
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010561685055165497,
      "loss": 1.1985,
      "theoretical_loss": 3.3530733844628577,
      "tokens_seen": 2955280384
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010559679037111335,
      "loss": 1.3327,
      "theoretical_loss": 3.3530676986812216,
      "tokens_seen": 2955345920
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010557673019057172,
      "loss": 1.2746,
      "theoretical_loss": 3.3530620130609714,
      "tokens_seen": 2955411456
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001055566700100301,
      "loss": 1.2918,
      "theoretical_loss": 3.353056327602099,
      "tokens_seen": 2955476992
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010553660982948848,
      "loss": 1.1558,
      "theoretical_loss": 3.3530506423045963,
      "tokens_seen": 2955542528
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010551654964894684,
      "loss": 1.2668,
      "theoretical_loss": 3.3530449571684544,
      "tokens_seen": 2955608064
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010549648946840521,
      "loss": 1.222,
      "theoretical_loss": 3.3530392721936666,
      "tokens_seen": 2955673600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010547642928786359,
      "loss": 1.2035,
      "theoretical_loss": 3.3530335873802235,
      "tokens_seen": 2955739136
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010545636910732197,
      "loss": 1.2112,
      "theoretical_loss": 3.353027902728117,
      "tokens_seen": 2955804672
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010543630892678034,
      "loss": 1.2836,
      "theoretical_loss": 3.3530222182373395,
      "tokens_seen": 2955870208
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010541624874623872,
      "loss": 1.2345,
      "theoretical_loss": 3.353016533907883,
      "tokens_seen": 2955935744
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010539618856569709,
      "loss": 1.2494,
      "theoretical_loss": 3.353010849739739,
      "tokens_seen": 2956001280
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010537612838515547,
      "loss": 1.2144,
      "theoretical_loss": 3.353005165732899,
      "tokens_seen": 2956066816
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010535606820461384,
      "loss": 1.3161,
      "theoretical_loss": 3.3529994818873554,
      "tokens_seen": 2956132352
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010533600802407222,
      "loss": 1.1748,
      "theoretical_loss": 3.3529937982030997,
      "tokens_seen": 2956197888
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010531594784353058,
      "loss": 1.2701,
      "theoretical_loss": 3.352988114680124,
      "tokens_seen": 2956263424
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010529588766298896,
      "loss": 1.2717,
      "theoretical_loss": 3.3529824313184196,
      "tokens_seen": 2956328960
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010527582748244735,
      "loss": 1.4392,
      "theoretical_loss": 3.352976748117979,
      "tokens_seen": 2956394496
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010525576730190572,
      "loss": 1.3403,
      "theoretical_loss": 3.352971065078794,
      "tokens_seen": 2956460032
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001052357071213641,
      "loss": 1.326,
      "theoretical_loss": 3.352965382200856,
      "tokens_seen": 2956525568
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010521564694082247,
      "loss": 1.4231,
      "theoretical_loss": 3.3529596994841575,
      "tokens_seen": 2956591104
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010519558676028085,
      "loss": 1.1832,
      "theoretical_loss": 3.35295401692869,
      "tokens_seen": 2956656640
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010517552657973923,
      "loss": 1.21,
      "theoretical_loss": 3.352948334534445,
      "tokens_seen": 2956722176
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001051554663991976,
      "loss": 1.3016,
      "theoretical_loss": 3.3529426523014147,
      "tokens_seen": 2956787712
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010513540621865598,
      "loss": 1.2884,
      "theoretical_loss": 3.352936970229591,
      "tokens_seen": 2956853248
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010511534603811434,
      "loss": 1.3146,
      "theoretical_loss": 3.3529312883189655,
      "tokens_seen": 2956918784
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010509528585757272,
      "loss": 1.2065,
      "theoretical_loss": 3.3529256065695305,
      "tokens_seen": 2956984320
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010507522567703109,
      "loss": 1.3611,
      "theoretical_loss": 3.3529199249812778,
      "tokens_seen": 2957049856
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010505516549648947,
      "loss": 1.2852,
      "theoretical_loss": 3.3529142435541988,
      "tokens_seen": 2957115392
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010503510531594784,
      "loss": 1.3212,
      "theoretical_loss": 3.3529085622882855,
      "tokens_seen": 2957180928
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010501504513540622,
      "loss": 1.3165,
      "theoretical_loss": 3.35290288118353,
      "tokens_seen": 2957246464
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001049949849548646,
      "loss": 1.3091,
      "theoretical_loss": 3.3528972002399238,
      "tokens_seen": 2957312000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010497492477432297,
      "loss": 1.3333,
      "theoretical_loss": 3.3528915194574593,
      "tokens_seen": 2957377536
    },
    {
      "epoch": 0.79,
      "objective/train/docs_used": 1075270,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3097165822982788,
      "objective/train/theoretical_loss": 3.352885838836128,
      "objective/train/tokens_used": 1327902176,
      "theoretical_loss": 3.352885838836128,
      "tokens_seen": 2957443072
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010495486459378135,
      "loss": 1.2758,
      "theoretical_loss": 3.352885838836128,
      "tokens_seen": 2957443072
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010493480441323971,
      "loss": 1.2561,
      "theoretical_loss": 3.3528801583759216,
      "tokens_seen": 2957508608
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001049147442326981,
      "loss": 1.267,
      "theoretical_loss": 3.3528744780768323,
      "tokens_seen": 2957574144
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010489468405215647,
      "loss": 1.2733,
      "theoretical_loss": 3.352868797938852,
      "tokens_seen": 2957639680
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010487462387161485,
      "loss": 1.2519,
      "theoretical_loss": 3.352863117961972,
      "tokens_seen": 2957705216
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010485456369107323,
      "loss": 1.2237,
      "theoretical_loss": 3.3528574381461844,
      "tokens_seen": 2957770752
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001048345035105316,
      "loss": 1.2137,
      "theoretical_loss": 3.3528517584914814,
      "tokens_seen": 2957836288
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010481444332998998,
      "loss": 1.2308,
      "theoretical_loss": 3.352846078997855,
      "tokens_seen": 2957901824
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010479438314944835,
      "loss": 1.3492,
      "theoretical_loss": 3.352840399665296,
      "tokens_seen": 2957967360
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010477432296890673,
      "loss": 1.2196,
      "theoretical_loss": 3.3528347204937976,
      "tokens_seen": 2958032896
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001047542627883651,
      "loss": 1.1902,
      "theoretical_loss": 3.352829041483351,
      "tokens_seen": 2958098432
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010473420260782347,
      "loss": 1.2267,
      "theoretical_loss": 3.352823362633948,
      "tokens_seen": 2958163968
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010471414242728184,
      "loss": 1.2216,
      "theoretical_loss": 3.3528176839455806,
      "tokens_seen": 2958229504
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010469408224674022,
      "loss": 1.1906,
      "theoretical_loss": 3.3528120054182406,
      "tokens_seen": 2958295040
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001046740220661986,
      "loss": 1.1864,
      "theoretical_loss": 3.35280632705192,
      "tokens_seen": 2958360576
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010465396188565697,
      "loss": 1.3177,
      "theoretical_loss": 3.3528006488466104,
      "tokens_seen": 2958426112
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010463390170511535,
      "loss": 1.308,
      "theoretical_loss": 3.352794970802304,
      "tokens_seen": 2958491648
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010461384152457372,
      "loss": 1.3224,
      "theoretical_loss": 3.3527892929189926,
      "tokens_seen": 2958557184
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001045937813440321,
      "loss": 1.2478,
      "theoretical_loss": 3.352783615196668,
      "tokens_seen": 2958622720
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010457372116349047,
      "loss": 1.3254,
      "theoretical_loss": 3.352777937635322,
      "tokens_seen": 2958688256
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010455366098294885,
      "loss": 1.3037,
      "theoretical_loss": 3.3527722602349463,
      "tokens_seen": 2958753792
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010453360080240722,
      "loss": 1.2391,
      "theoretical_loss": 3.3527665829955335,
      "tokens_seen": 2958819328
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001045135406218656,
      "loss": 1.2978,
      "theoretical_loss": 3.3527609059170747,
      "tokens_seen": 2958884864
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010449348044132398,
      "loss": 1.3058,
      "theoretical_loss": 3.3527552289995617,
      "tokens_seen": 2958950400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010447342026078235,
      "loss": 1.2384,
      "theoretical_loss": 3.3527495522429875,
      "tokens_seen": 2959015936
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010445336008024073,
      "loss": 1.2911,
      "theoretical_loss": 3.3527438756473424,
      "tokens_seen": 2959081472
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001044332998996991,
      "loss": 1.1744,
      "theoretical_loss": 3.3527381992126193,
      "tokens_seen": 2959147008
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010441323971915748,
      "loss": 1.3145,
      "theoretical_loss": 3.35273252293881,
      "tokens_seen": 2959212544
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010439317953861586,
      "loss": 1.2959,
      "theoretical_loss": 3.3527268468259064,
      "tokens_seen": 2959278080
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010437311935807423,
      "loss": 1.3442,
      "theoretical_loss": 3.3527211708738998,
      "tokens_seen": 2959343616
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010435305917753259,
      "loss": 1.193,
      "theoretical_loss": 3.3527154950827827,
      "tokens_seen": 2959409152
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010433299899699097,
      "loss": 1.3113,
      "theoretical_loss": 3.3527098194525466,
      "tokens_seen": 2959474688
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010431293881644934,
      "loss": 1.3615,
      "theoretical_loss": 3.3527041439831833,
      "tokens_seen": 2959540224
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010429287863590772,
      "loss": 1.2596,
      "theoretical_loss": 3.352698468674685,
      "tokens_seen": 2959605760
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001042728184553661,
      "loss": 1.265,
      "theoretical_loss": 3.352692793527044,
      "tokens_seen": 2959671296
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010425275827482447,
      "loss": 1.2134,
      "theoretical_loss": 3.352687118540251,
      "tokens_seen": 2959736832
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010423269809428285,
      "loss": 1.2859,
      "theoretical_loss": 3.352681443714299,
      "tokens_seen": 2959802368
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010421263791374122,
      "loss": 1.3616,
      "theoretical_loss": 3.352675769049179,
      "tokens_seen": 2959867904
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001041925777331996,
      "loss": 1.4126,
      "theoretical_loss": 3.3526700945448837,
      "tokens_seen": 2959933440
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010417251755265798,
      "loss": 1.2584,
      "theoretical_loss": 3.352664420201404,
      "tokens_seen": 2959998976
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010415245737211635,
      "loss": 1.3204,
      "theoretical_loss": 3.352658746018733,
      "tokens_seen": 2960064512
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010413239719157473,
      "loss": 1.1653,
      "theoretical_loss": 3.352653071996862,
      "tokens_seen": 2960130048
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001041123370110331,
      "loss": 1.3276,
      "theoretical_loss": 3.3526473981357823,
      "tokens_seen": 2960195584
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010409227683049148,
      "loss": 1.3512,
      "theoretical_loss": 3.3526417244354865,
      "tokens_seen": 2960261120
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010407221664994985,
      "loss": 1.2048,
      "theoretical_loss": 3.352636050895966,
      "tokens_seen": 2960326656
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010405215646940823,
      "loss": 1.3013,
      "theoretical_loss": 3.3526303775172135,
      "tokens_seen": 2960392192
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001040320962888666,
      "loss": 1.2994,
      "theoretical_loss": 3.3526247042992203,
      "tokens_seen": 2960457728
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010401203610832498,
      "loss": 1.2428,
      "theoretical_loss": 3.352619031241978,
      "tokens_seen": 2960523264
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010399197592778336,
      "loss": 1.2342,
      "theoretical_loss": 3.3526133583454794,
      "tokens_seen": 2960588800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010397191574724173,
      "loss": 1.278,
      "theoretical_loss": 3.352607685609715,
      "tokens_seen": 2960654336
    },
    {
      "debugging/Compilability": 1.0,
      "debugging/distinct-1-grams": 0.7304384858452192,
      "debugging/entropy-1-grams": 4.9022521807961565,
      "debugging/length": 488.7142857142857,
      "debugging/num_segments": 7,
      "debugging/score": 0.0068462501065870455,
      "debugging/score_std": 0.004935555558817252,
      "epoch": 0.79,
      "objective/train/docs_used": 1076444,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2278602123260498,
      "objective/train/theoretical_loss": 3.3526020130346783,
      "objective/train/tokens_used": 1331178976,
      "theoretical_loss": 3.3526020130346783,
      "tokens_seen": 2960719872
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001039518555667001,
      "loss": 1.2271,
      "theoretical_loss": 3.3526020130346783,
      "tokens_seen": 2960719872
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010393179538615847,
      "loss": 1.3717,
      "theoretical_loss": 3.3525963406203596,
      "tokens_seen": 2960785408
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010391173520561685,
      "loss": 1.162,
      "theoretical_loss": 3.3525906683667523,
      "tokens_seen": 2960850944
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010389167502507522,
      "loss": 1.2339,
      "theoretical_loss": 3.352584996273847,
      "tokens_seen": 2960916480
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001038716148445336,
      "loss": 1.3635,
      "theoretical_loss": 3.3525793243416366,
      "tokens_seen": 2960982016
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010385155466399197,
      "loss": 1.2894,
      "theoretical_loss": 3.3525736525701126,
      "tokens_seen": 2961047552
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010383149448345035,
      "loss": 1.3101,
      "theoretical_loss": 3.352567980959267,
      "tokens_seen": 2961113088
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010381143430290873,
      "loss": 1.3295,
      "theoretical_loss": 3.352562309509091,
      "tokens_seen": 2961178624
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001037913741223671,
      "loss": 1.2459,
      "theoretical_loss": 3.3525566382195775,
      "tokens_seen": 2961244160
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010377131394182549,
      "loss": 1.1554,
      "theoretical_loss": 3.3525509670907176,
      "tokens_seen": 2961309696
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010375125376128385,
      "loss": 1.1917,
      "theoretical_loss": 3.3525452961225035,
      "tokens_seen": 2961375232
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010373119358074223,
      "loss": 1.3334,
      "theoretical_loss": 3.352539625314927,
      "tokens_seen": 2961440768
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001037111334002006,
      "loss": 1.3269,
      "theoretical_loss": 3.352533954667981,
      "tokens_seen": 2961506304
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010369107321965898,
      "loss": 1.2182,
      "theoretical_loss": 3.3525282841816555,
      "tokens_seen": 2961571840
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010367101303911736,
      "loss": 1.2387,
      "theoretical_loss": 3.352522613855944,
      "tokens_seen": 2961637376
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010365095285857573,
      "loss": 1.2952,
      "theoretical_loss": 3.3525169436908375,
      "tokens_seen": 2961702912
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010363089267803411,
      "loss": 1.3007,
      "theoretical_loss": 3.352511273686328,
      "tokens_seen": 2961768448
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010361083249749248,
      "loss": 1.2958,
      "theoretical_loss": 3.352505603842408,
      "tokens_seen": 2961833984
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010359077231695086,
      "loss": 1.1777,
      "theoretical_loss": 3.352499934159069,
      "tokens_seen": 2961899520
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010357071213640922,
      "loss": 1.3482,
      "theoretical_loss": 3.3524942646363027,
      "tokens_seen": 2961965056
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001035506519558676,
      "loss": 1.2846,
      "theoretical_loss": 3.3524885952741013,
      "tokens_seen": 2962030592
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010353059177532597,
      "loss": 1.3025,
      "theoretical_loss": 3.3524829260724562,
      "tokens_seen": 2962096128
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010351053159478435,
      "loss": 1.3466,
      "theoretical_loss": 3.35247725703136,
      "tokens_seen": 2962161664
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010349047141424273,
      "loss": 1.2362,
      "theoretical_loss": 3.3524715881508045,
      "tokens_seen": 2962227200
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001034704112337011,
      "loss": 1.1246,
      "theoretical_loss": 3.3524659194307813,
      "tokens_seen": 2962292736
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010345035105315948,
      "loss": 1.265,
      "theoretical_loss": 3.352460250871282,
      "tokens_seen": 2962358272
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010343029087261785,
      "loss": 1.4261,
      "theoretical_loss": 3.3524545824722995,
      "tokens_seen": 2962423808
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010341023069207624,
      "loss": 1.2877,
      "theoretical_loss": 3.352448914233825,
      "tokens_seen": 2962489344
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010339017051153462,
      "loss": 1.1853,
      "theoretical_loss": 3.3524432461558504,
      "tokens_seen": 2962554880
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010337011033099298,
      "loss": 1.1856,
      "theoretical_loss": 3.3524375782383675,
      "tokens_seen": 2962620416
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010335005015045136,
      "loss": 1.3495,
      "theoretical_loss": 3.3524319104813687,
      "tokens_seen": 2962685952
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010332998996990973,
      "loss": 1.19,
      "theoretical_loss": 3.3524262428848455,
      "tokens_seen": 2962751488
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010330992978936811,
      "loss": 1.1854,
      "theoretical_loss": 3.35242057544879,
      "tokens_seen": 2962817024
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010328986960882648,
      "loss": 1.2579,
      "theoretical_loss": 3.352414908173194,
      "tokens_seen": 2962882560
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010326980942828486,
      "loss": 1.27,
      "theoretical_loss": 3.3524092410580493,
      "tokens_seen": 2962948096
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010324974924774323,
      "loss": 1.2207,
      "theoretical_loss": 3.352403574103348,
      "tokens_seen": 2963013632
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010322968906720161,
      "loss": 1.1163,
      "theoretical_loss": 3.352397907309082,
      "tokens_seen": 2963079168
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010320962888665999,
      "loss": 1.2509,
      "theoretical_loss": 3.3523922406752433,
      "tokens_seen": 2963144704
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010318956870611836,
      "loss": 1.146,
      "theoretical_loss": 3.352386574201824,
      "tokens_seen": 2963210240
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010316950852557672,
      "loss": 1.2404,
      "theoretical_loss": 3.352380907888815,
      "tokens_seen": 2963275776
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001031494483450351,
      "loss": 1.2701,
      "theoretical_loss": 3.352375241736209,
      "tokens_seen": 2963341312
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010312938816449348,
      "loss": 1.2068,
      "theoretical_loss": 3.3523695757439986,
      "tokens_seen": 2963406848
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010310932798395185,
      "loss": 1.3703,
      "theoretical_loss": 3.3523639099121745,
      "tokens_seen": 2963472384
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010308926780341023,
      "loss": 1.1609,
      "theoretical_loss": 3.3523582442407287,
      "tokens_seen": 2963537920
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001030692076228686,
      "loss": 1.299,
      "theoretical_loss": 3.3523525787296538,
      "tokens_seen": 2963603456
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010304914744232698,
      "loss": 1.33,
      "theoretical_loss": 3.352346913378941,
      "tokens_seen": 2963668992
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010302908726178537,
      "loss": 1.1599,
      "theoretical_loss": 3.352341248188583,
      "tokens_seen": 2963734528
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010300902708124374,
      "loss": 1.1693,
      "theoretical_loss": 3.3523355831585713,
      "tokens_seen": 2963800064
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010298896690070212,
      "loss": 1.3254,
      "theoretical_loss": 3.3523299182888975,
      "tokens_seen": 2963865600
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010296890672016048,
      "loss": 1.3193,
      "theoretical_loss": 3.352324253579554,
      "tokens_seen": 2963931136
    },
    {
      "epoch": 0.8,
      "objective/train/docs_used": 1077722,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1816346645355225,
      "objective/train/theoretical_loss": 3.3523185890305327,
      "objective/train/tokens_used": 1334455776,
      "theoretical_loss": 3.3523185890305327,
      "tokens_seen": 2963996672
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010294884653961886,
      "loss": 1.1716,
      "theoretical_loss": 3.3523185890305327,
      "tokens_seen": 2963996672
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010292878635907723,
      "loss": 1.2886,
      "theoretical_loss": 3.3523129246418253,
      "tokens_seen": 2964062208
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010290872617853561,
      "loss": 1.2381,
      "theoretical_loss": 3.352307260413424,
      "tokens_seen": 2964127744
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010288866599799399,
      "loss": 1.288,
      "theoretical_loss": 3.3523015963453204,
      "tokens_seen": 2964193280
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010286860581745236,
      "loss": 1.2356,
      "theoretical_loss": 3.352295932437506,
      "tokens_seen": 2964258816
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010284854563691074,
      "loss": 1.3147,
      "theoretical_loss": 3.352290268689974,
      "tokens_seen": 2964324352
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010282848545636911,
      "loss": 1.3789,
      "theoretical_loss": 3.352284605102715,
      "tokens_seen": 2964389888
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010280842527582749,
      "loss": 1.1977,
      "theoretical_loss": 3.352278941675722,
      "tokens_seen": 2964455424
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010278836509528585,
      "loss": 1.2529,
      "theoretical_loss": 3.352273278408986,
      "tokens_seen": 2964520960
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010276830491474423,
      "loss": 1.2755,
      "theoretical_loss": 3.3522676153024995,
      "tokens_seen": 2964586496
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001027482447342026,
      "loss": 1.2476,
      "theoretical_loss": 3.3522619523562542,
      "tokens_seen": 2964652032
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010272818455366098,
      "loss": 1.2611,
      "theoretical_loss": 3.352256289570242,
      "tokens_seen": 2964717568
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010270812437311935,
      "loss": 1.2286,
      "theoretical_loss": 3.352250626944455,
      "tokens_seen": 2964783104
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010268806419257773,
      "loss": 1.2784,
      "theoretical_loss": 3.3522449644788854,
      "tokens_seen": 2964848640
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001026680040120361,
      "loss": 1.1412,
      "theoretical_loss": 3.3522393021735244,
      "tokens_seen": 2964914176
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001026479438314945,
      "loss": 1.2137,
      "theoretical_loss": 3.352233640028364,
      "tokens_seen": 2964979712
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010262788365095287,
      "loss": 1.2401,
      "theoretical_loss": 3.352227978043397,
      "tokens_seen": 2965045248
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010260782347041125,
      "loss": 1.32,
      "theoretical_loss": 3.3522223162186147,
      "tokens_seen": 2965110784
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010258776328986961,
      "loss": 1.3425,
      "theoretical_loss": 3.3522166545540086,
      "tokens_seen": 2965176320
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010256770310932798,
      "loss": 1.2181,
      "theoretical_loss": 3.3522109930495714,
      "tokens_seen": 2965241856
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010254764292878636,
      "loss": 1.2091,
      "theoretical_loss": 3.352205331705295,
      "tokens_seen": 2965307392
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010252758274824474,
      "loss": 1.1886,
      "theoretical_loss": 3.3521996705211707,
      "tokens_seen": 2965372928
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010250752256770311,
      "loss": 1.1221,
      "theoretical_loss": 3.3521940094971905,
      "tokens_seen": 2965438464
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010248746238716149,
      "loss": 1.2789,
      "theoretical_loss": 3.352188348633347,
      "tokens_seen": 2965504000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010246740220661986,
      "loss": 1.2697,
      "theoretical_loss": 3.3521826879296315,
      "tokens_seen": 2965569536
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010244734202607824,
      "loss": 1.1643,
      "theoretical_loss": 3.3521770273860367,
      "tokens_seen": 2965635072
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010242728184553662,
      "loss": 1.2152,
      "theoretical_loss": 3.3521713670025535,
      "tokens_seen": 2965700608
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010240722166499499,
      "loss": 1.1597,
      "theoretical_loss": 3.3521657067791746,
      "tokens_seen": 2965766144
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010238716148445335,
      "loss": 1.3076,
      "theoretical_loss": 3.3521600467158916,
      "tokens_seen": 2965831680
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010236710130391173,
      "loss": 1.2866,
      "theoretical_loss": 3.3521543868126966,
      "tokens_seen": 2965897216
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001023470411233701,
      "loss": 1.1956,
      "theoretical_loss": 3.3521487270695816,
      "tokens_seen": 2965962752
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010232698094282848,
      "loss": 1.1166,
      "theoretical_loss": 3.352143067486538,
      "tokens_seen": 2966028288
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010230692076228686,
      "loss": 1.2588,
      "theoretical_loss": 3.3521374080635584,
      "tokens_seen": 2966093824
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010228686058174525,
      "loss": 1.1963,
      "theoretical_loss": 3.352131748800634,
      "tokens_seen": 2966159360
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010226680040120362,
      "loss": 1.2627,
      "theoretical_loss": 3.352126089697758,
      "tokens_seen": 2966224896
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000102246740220662,
      "loss": 1.4016,
      "theoretical_loss": 3.352120430754921,
      "tokens_seen": 2966290432
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010222668004012037,
      "loss": 1.3104,
      "theoretical_loss": 3.3521147719721154,
      "tokens_seen": 2966355968
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010220661985957875,
      "loss": 1.1939,
      "theoretical_loss": 3.3521091133493335,
      "tokens_seen": 2966421504
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010218655967903711,
      "loss": 1.274,
      "theoretical_loss": 3.3521034548865667,
      "tokens_seen": 2966487040
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010216649949849549,
      "loss": 1.2374,
      "theoretical_loss": 3.352097796583807,
      "tokens_seen": 2966552576
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010214643931795386,
      "loss": 1.2543,
      "theoretical_loss": 3.352092138441047,
      "tokens_seen": 2966618112
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010212637913741224,
      "loss": 1.2639,
      "theoretical_loss": 3.352086480458278,
      "tokens_seen": 2966683648
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010210631895687061,
      "loss": 1.3438,
      "theoretical_loss": 3.352080822635492,
      "tokens_seen": 2966749184
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010208625877632899,
      "loss": 1.2041,
      "theoretical_loss": 3.352075164972681,
      "tokens_seen": 2966814720
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010206619859578737,
      "loss": 1.3666,
      "theoretical_loss": 3.3520695074698375,
      "tokens_seen": 2966880256
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010204613841524574,
      "loss": 1.3108,
      "theoretical_loss": 3.352063850126952,
      "tokens_seen": 2966945792
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010202607823470412,
      "loss": 1.3297,
      "theoretical_loss": 3.3520581929440185,
      "tokens_seen": 2967011328
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010200601805416248,
      "loss": 1.2337,
      "theoretical_loss": 3.352052535921027,
      "tokens_seen": 2967076864
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010198595787362086,
      "loss": 1.3376,
      "theoretical_loss": 3.352046879057971,
      "tokens_seen": 2967142400
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010196589769307923,
      "loss": 1.3452,
      "theoretical_loss": 3.352041222354841,
      "tokens_seen": 2967207936
    },
    {
      "epoch": 0.8,
      "objective/train/docs_used": 1078969,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3053348064422607,
      "objective/train/theoretical_loss": 3.3520355658116303,
      "objective/train/tokens_used": 1337732576,
      "theoretical_loss": 3.3520355658116303,
      "tokens_seen": 2967273472
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010194583751253761,
      "loss": 1.278,
      "theoretical_loss": 3.3520355658116303,
      "tokens_seen": 2967273472
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010192577733199598,
      "loss": 1.2441,
      "theoretical_loss": 3.35202990942833,
      "tokens_seen": 2967339008
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010190571715145437,
      "loss": 1.327,
      "theoretical_loss": 3.352024253204932,
      "tokens_seen": 2967404544
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010188565697091275,
      "loss": 1.2416,
      "theoretical_loss": 3.3520185971414285,
      "tokens_seen": 2967470080
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010186559679037112,
      "loss": 1.2221,
      "theoretical_loss": 3.3520129412378115,
      "tokens_seen": 2967535616
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001018455366098295,
      "loss": 1.2241,
      "theoretical_loss": 3.3520072854940732,
      "tokens_seen": 2967601152
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010182547642928788,
      "loss": 1.259,
      "theoretical_loss": 3.352001629910205,
      "tokens_seen": 2967666688
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010180541624874624,
      "loss": 1.1815,
      "theoretical_loss": 3.3519959744861993,
      "tokens_seen": 2967732224
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010178535606820461,
      "loss": 1.2014,
      "theoretical_loss": 3.3519903192220477,
      "tokens_seen": 2967797760
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010176529588766299,
      "loss": 1.2987,
      "theoretical_loss": 3.3519846641177424,
      "tokens_seen": 2967863296
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010174523570712136,
      "loss": 1.318,
      "theoretical_loss": 3.3519790091732755,
      "tokens_seen": 2967928832
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010172517552657974,
      "loss": 1.3367,
      "theoretical_loss": 3.3519733543886385,
      "tokens_seen": 2967994368
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010170511534603812,
      "loss": 1.2238,
      "theoretical_loss": 3.351967699763823,
      "tokens_seen": 2968059904
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010168505516549649,
      "loss": 1.1826,
      "theoretical_loss": 3.3519620452988224,
      "tokens_seen": 2968125440
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010166499498495487,
      "loss": 1.2077,
      "theoretical_loss": 3.3519563909936276,
      "tokens_seen": 2968190976
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010164493480441324,
      "loss": 1.2648,
      "theoretical_loss": 3.3519507368482304,
      "tokens_seen": 2968256512
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010162487462387162,
      "loss": 1.2105,
      "theoretical_loss": 3.351945082862623,
      "tokens_seen": 2968322048
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010160481444332998,
      "loss": 1.2881,
      "theoretical_loss": 3.351939429036798,
      "tokens_seen": 2968387584
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010158475426278836,
      "loss": 1.3104,
      "theoretical_loss": 3.3519337753707465,
      "tokens_seen": 2968453120
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010156469408224673,
      "loss": 1.2829,
      "theoretical_loss": 3.3519281218644608,
      "tokens_seen": 2968518656
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010154463390170511,
      "loss": 1.2961,
      "theoretical_loss": 3.351922468517933,
      "tokens_seen": 2968584192
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001015245737211635,
      "loss": 1.295,
      "theoretical_loss": 3.3519168153311547,
      "tokens_seen": 2968649728
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010150451354062187,
      "loss": 1.2108,
      "theoretical_loss": 3.351911162304118,
      "tokens_seen": 2968715264
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010148445336008025,
      "loss": 1.1974,
      "theoretical_loss": 3.351905509436815,
      "tokens_seen": 2968780800
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010146439317953863,
      "loss": 1.2778,
      "theoretical_loss": 3.3518998567292373,
      "tokens_seen": 2968846336
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000101444332998997,
      "loss": 1.2704,
      "theoretical_loss": 3.3518942041813777,
      "tokens_seen": 2968911872
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010142427281845536,
      "loss": 1.1382,
      "theoretical_loss": 3.351888551793227,
      "tokens_seen": 2968977408
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010140421263791374,
      "loss": 1.2578,
      "theoretical_loss": 3.351882899564778,
      "tokens_seen": 2969042944
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010138415245737212,
      "loss": 1.2709,
      "theoretical_loss": 3.3518772474960223,
      "tokens_seen": 2969108480
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010136409227683049,
      "loss": 1.2953,
      "theoretical_loss": 3.3518715955869522,
      "tokens_seen": 2969174016
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010134403209628887,
      "loss": 1.2752,
      "theoretical_loss": 3.3518659438375593,
      "tokens_seen": 2969239552
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010132397191574724,
      "loss": 1.2509,
      "theoretical_loss": 3.3518602922478355,
      "tokens_seen": 2969305088
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010130391173520562,
      "loss": 1.2648,
      "theoretical_loss": 3.3518546408177734,
      "tokens_seen": 2969370624
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000101283851554664,
      "loss": 1.2878,
      "theoretical_loss": 3.3518489895473644,
      "tokens_seen": 2969436160
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010126379137412237,
      "loss": 1.2142,
      "theoretical_loss": 3.3518433384366,
      "tokens_seen": 2969501696
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010124373119358075,
      "loss": 1.1704,
      "theoretical_loss": 3.3518376874854736,
      "tokens_seen": 2969567232
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010122367101303911,
      "loss": 1.3463,
      "theoretical_loss": 3.351832036693976,
      "tokens_seen": 2969632768
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010120361083249748,
      "loss": 1.2971,
      "theoretical_loss": 3.3518263860620996,
      "tokens_seen": 2969698304
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010118355065195586,
      "loss": 1.2164,
      "theoretical_loss": 3.351820735589836,
      "tokens_seen": 2969763840
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010116349047141425,
      "loss": 1.2387,
      "theoretical_loss": 3.3518150852771775,
      "tokens_seen": 2969829376
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010114343029087263,
      "loss": 1.2313,
      "theoretical_loss": 3.351809435124116,
      "tokens_seen": 2969894912
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000101123370110331,
      "loss": 1.3482,
      "theoretical_loss": 3.351803785130644,
      "tokens_seen": 2969960448
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010110330992978938,
      "loss": 1.3213,
      "theoretical_loss": 3.351798135296752,
      "tokens_seen": 2970025984
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010108324974924775,
      "loss": 1.1861,
      "theoretical_loss": 3.3517924856224335,
      "tokens_seen": 2970091520
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010106318956870613,
      "loss": 1.3543,
      "theoretical_loss": 3.35178683610768,
      "tokens_seen": 2970157056
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001010431293881645,
      "loss": 1.2061,
      "theoretical_loss": 3.351781186752483,
      "tokens_seen": 2970222592
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010102306920762287,
      "loss": 1.248,
      "theoretical_loss": 3.3517755375568354,
      "tokens_seen": 2970288128
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010100300902708124,
      "loss": 1.2039,
      "theoretical_loss": 3.351769888520728,
      "tokens_seen": 2970353664
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010098294884653962,
      "loss": 1.2298,
      "theoretical_loss": 3.351764239644154,
      "tokens_seen": 2970419200
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000100962888665998,
      "loss": 1.2233,
      "theoretical_loss": 3.3517585909271044,
      "tokens_seen": 2970484736
    },
    {
      "epoch": 0.8,
      "objective/train/docs_used": 1080058,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3455166816711426,
      "objective/train/theoretical_loss": 3.3517529423695716,
      "objective/train/tokens_used": 1341009376,
      "theoretical_loss": 3.3517529423695716,
      "tokens_seen": 2970550272
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010094282848545637,
      "loss": 1.2756,
      "theoretical_loss": 3.3517529423695716,
      "tokens_seen": 2970550272
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010092276830491475,
      "loss": 1.2529,
      "theoretical_loss": 3.351747293971547,
      "tokens_seen": 2970615808
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010090270812437312,
      "loss": 1.204,
      "theoretical_loss": 3.351741645733024,
      "tokens_seen": 2970681344
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001008826479438315,
      "loss": 1.2084,
      "theoretical_loss": 3.3517359976539933,
      "tokens_seen": 2970746880
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010086258776328987,
      "loss": 1.2278,
      "theoretical_loss": 3.351730349734447,
      "tokens_seen": 2970812416
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010084252758274825,
      "loss": 1.3062,
      "theoretical_loss": 3.3517247019743777,
      "tokens_seen": 2970877952
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010082246740220661,
      "loss": 1.2295,
      "theoretical_loss": 3.3517190543737767,
      "tokens_seen": 2970943488
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010080240722166499,
      "loss": 1.2044,
      "theoretical_loss": 3.351713406932636,
      "tokens_seen": 2971009024
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010078234704112338,
      "loss": 1.1457,
      "theoretical_loss": 3.3517077596509486,
      "tokens_seen": 2971074560
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010076228686058175,
      "loss": 1.2549,
      "theoretical_loss": 3.3517021125287054,
      "tokens_seen": 2971140096
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010074222668004013,
      "loss": 1.2474,
      "theoretical_loss": 3.3516964655658987,
      "tokens_seen": 2971205632
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001007221664994985,
      "loss": 1.2753,
      "theoretical_loss": 3.3516908187625205,
      "tokens_seen": 2971271168
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010070210631895688,
      "loss": 1.3213,
      "theoretical_loss": 3.351685172118563,
      "tokens_seen": 2971336704
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010068204613841525,
      "loss": 1.2769,
      "theoretical_loss": 3.351679525634018,
      "tokens_seen": 2971402240
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010066198595787363,
      "loss": 1.2987,
      "theoretical_loss": 3.351673879308877,
      "tokens_seen": 2971467776
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010064192577733199,
      "loss": 1.3087,
      "theoretical_loss": 3.351668233143133,
      "tokens_seen": 2971533312
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010062186559679037,
      "loss": 1.3229,
      "theoretical_loss": 3.351662587136777,
      "tokens_seen": 2971598848
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010060180541624874,
      "loss": 1.216,
      "theoretical_loss": 3.351656941289802,
      "tokens_seen": 2971664384
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010058174523570712,
      "loss": 1.258,
      "theoretical_loss": 3.3516512956021987,
      "tokens_seen": 2971729920
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001005616850551655,
      "loss": 1.2629,
      "theoretical_loss": 3.35164565007396,
      "tokens_seen": 2971795456
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010054162487462387,
      "loss": 1.2367,
      "theoretical_loss": 3.351640004705078,
      "tokens_seen": 2971860992
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010052156469408225,
      "loss": 1.1479,
      "theoretical_loss": 3.351634359495544,
      "tokens_seen": 2971926528
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010050150451354062,
      "loss": 1.3501,
      "theoretical_loss": 3.3516287144453503,
      "tokens_seen": 2971992064
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000100481444332999,
      "loss": 1.2462,
      "theoretical_loss": 3.3516230695544893,
      "tokens_seen": 2972057600
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010046138415245738,
      "loss": 1.2546,
      "theoretical_loss": 3.351617424822952,
      "tokens_seen": 2972123136
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010044132397191574,
      "loss": 1.2418,
      "theoretical_loss": 3.3516117802507317,
      "tokens_seen": 2972188672
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010042126379137411,
      "loss": 1.1772,
      "theoretical_loss": 3.3516061358378195,
      "tokens_seen": 2972254208
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001004012036108325,
      "loss": 1.3895,
      "theoretical_loss": 3.3516004915842075,
      "tokens_seen": 2972319744
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010038114343029088,
      "loss": 1.2457,
      "theoretical_loss": 3.3515948474898876,
      "tokens_seen": 2972385280
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010036108324974925,
      "loss": 1.1557,
      "theoretical_loss": 3.3515892035548522,
      "tokens_seen": 2972450816
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010034102306920763,
      "loss": 1.2511,
      "theoretical_loss": 3.351583559779093,
      "tokens_seen": 2972516352
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000100320962888666,
      "loss": 1.2101,
      "theoretical_loss": 3.3515779161626025,
      "tokens_seen": 2972581888
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010030090270812438,
      "loss": 1.3762,
      "theoretical_loss": 3.351572272705372,
      "tokens_seen": 2972647424
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010028084252758276,
      "loss": 1.3009,
      "theoretical_loss": 3.351566629407394,
      "tokens_seen": 2972712960
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010026078234704113,
      "loss": 1.238,
      "theoretical_loss": 3.3515609862686597,
      "tokens_seen": 2972778496
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001002407221664995,
      "loss": 1.3088,
      "theoretical_loss": 3.351555343289162,
      "tokens_seen": 2972844032
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010022066198595787,
      "loss": 1.3399,
      "theoretical_loss": 3.351549700468892,
      "tokens_seen": 2972909568
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010020060180541625,
      "loss": 1.244,
      "theoretical_loss": 3.351544057807843,
      "tokens_seen": 2972975104
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010018054162487462,
      "loss": 1.2284,
      "theoretical_loss": 3.351538415306006,
      "tokens_seen": 2973040640
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000100160481444333,
      "loss": 1.1566,
      "theoretical_loss": 3.351532772963373,
      "tokens_seen": 2973106176
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010014042126379137,
      "loss": 1.2468,
      "theoretical_loss": 3.3515271307799366,
      "tokens_seen": 2973171712
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010012036108324975,
      "loss": 1.2895,
      "theoretical_loss": 3.3515214887556883,
      "tokens_seen": 2973237248
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010010030090270813,
      "loss": 1.302,
      "theoretical_loss": 3.3515158468906203,
      "tokens_seen": 2973302784
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001000802407221665,
      "loss": 1.4016,
      "theoretical_loss": 3.3515102051847245,
      "tokens_seen": 2973368320
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010006018054162486,
      "loss": 1.2979,
      "theoretical_loss": 3.3515045636379925,
      "tokens_seen": 2973433856
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010004012036108324,
      "loss": 1.3343,
      "theoretical_loss": 3.3514989222504177,
      "tokens_seen": 2973499392
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010002006018054163,
      "loss": 1.167,
      "theoretical_loss": 3.35149328102199,
      "tokens_seen": 2973564928
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001,
      "loss": 1.1974,
      "theoretical_loss": 3.3514876399527034,
      "tokens_seen": 2973630464
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.997993981945838e-05,
      "loss": 1.24,
      "theoretical_loss": 3.351481999042549,
      "tokens_seen": 2973696000
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.995987963891676e-05,
      "loss": 1.3974,
      "theoretical_loss": 3.3514763582915186,
      "tokens_seen": 2973761536
    },
    {
      "epoch": 0.8,
      "objective/train/docs_used": 1081215,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2280032634735107,
      "objective/train/theoretical_loss": 3.3514707176996046,
      "objective/train/tokens_used": 1344286176,
      "theoretical_loss": 3.3514707176996046,
      "tokens_seen": 2973827072
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.993981945837513e-05,
      "loss": 1.2228,
      "theoretical_loss": 3.3514707176996046,
      "tokens_seen": 2973827072
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.991975927783351e-05,
      "loss": 1.1668,
      "theoretical_loss": 3.3514650772667984,
      "tokens_seen": 2973892608
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.989969909729188e-05,
      "loss": 1.2275,
      "theoretical_loss": 3.351459436993093,
      "tokens_seen": 2973958144
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.987963891675026e-05,
      "loss": 1.1917,
      "theoretical_loss": 3.35145379687848,
      "tokens_seen": 2974023680
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.985957873620862e-05,
      "loss": 1.3146,
      "theoretical_loss": 3.3514481569229506,
      "tokens_seen": 2974089216
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.9839518555667e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.351442517126498,
      "tokens_seen": 2974154752
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.981945837512537e-05,
      "loss": 1.2594,
      "theoretical_loss": 3.351436877489114,
      "tokens_seen": 2974220288
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.979939819458375e-05,
      "loss": 1.2331,
      "theoretical_loss": 3.3514312380107896,
      "tokens_seen": 2974285824
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.977933801404212e-05,
      "loss": 1.3198,
      "theoretical_loss": 3.351425598691518,
      "tokens_seen": 2974351360
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.97592778335005e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.3514199595312903,
      "tokens_seen": 2974416896
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.973921765295888e-05,
      "loss": 1.17,
      "theoretical_loss": 3.3514143205300995,
      "tokens_seen": 2974482432
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.971915747241725e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.3514086816879365,
      "tokens_seen": 2974547968
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.969909729187563e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.3514030430047943,
      "tokens_seen": 2974613504
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.9679037111334e-05,
      "loss": 1.2277,
      "theoretical_loss": 3.3513974044806645,
      "tokens_seen": 2974679040
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.965897693079238e-05,
      "loss": 1.3057,
      "theoretical_loss": 3.3513917661155386,
      "tokens_seen": 2974744576
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.963891675025076e-05,
      "loss": 1.2152,
      "theoretical_loss": 3.3513861279094095,
      "tokens_seen": 2974810112
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.961885656970913e-05,
      "loss": 1.2329,
      "theoretical_loss": 3.3513804898622688,
      "tokens_seen": 2974875648
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.959879638916751e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.3513748519741084,
      "tokens_seen": 2974941184
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.957873620862588e-05,
      "loss": 1.1847,
      "theoretical_loss": 3.3513692142449205,
      "tokens_seen": 2975006720
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.955867602808426e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.351363576674697,
      "tokens_seen": 2975072256
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.953861584754263e-05,
      "loss": 1.354,
      "theoretical_loss": 3.35135793926343,
      "tokens_seen": 2975137792
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.951855566700101e-05,
      "loss": 1.3846,
      "theoretical_loss": 3.3513523020111116,
      "tokens_seen": 2975203328
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.949849548645939e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.3513466649177337,
      "tokens_seen": 2975268864
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.947843530591776e-05,
      "loss": 1.1916,
      "theoretical_loss": 3.3513410279832883,
      "tokens_seen": 2975334400
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.945837512537612e-05,
      "loss": 1.136,
      "theoretical_loss": 3.3513353912077672,
      "tokens_seen": 2975399936
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.94383149448345e-05,
      "loss": 1.2404,
      "theoretical_loss": 3.3513297545911627,
      "tokens_seen": 2975465472
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.941825476429288e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.351324118133467,
      "tokens_seen": 2975531008
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.939819458375125e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.3513184818346717,
      "tokens_seen": 2975596544
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.937813440320963e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.3513128456947694,
      "tokens_seen": 2975662080
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.9358074222668e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.3513072097137515,
      "tokens_seen": 2975727616
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.933801404212638e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.35130157389161,
      "tokens_seen": 2975793152
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.931795386158475e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.3512959382283376,
      "tokens_seen": 2975858688
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.929789368104313e-05,
      "loss": 1.3555,
      "theoretical_loss": 3.351290302723926,
      "tokens_seen": 2975924224
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.927783350050152e-05,
      "loss": 1.2865,
      "theoretical_loss": 3.351284667378367,
      "tokens_seen": 2975989760
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.925777331995988e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.3512790321916524,
      "tokens_seen": 2976055296
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.923771313941826e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.3512733971637747,
      "tokens_seen": 2976120832
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.921765295887663e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.3512677622947264,
      "tokens_seen": 2976186368
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.919759277833501e-05,
      "loss": 1.1753,
      "theoretical_loss": 3.3512621275844983,
      "tokens_seen": 2976251904
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.917753259779339e-05,
      "loss": 1.2229,
      "theoretical_loss": 3.3512564930330835,
      "tokens_seen": 2976317440
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.915747241725176e-05,
      "loss": 1.3348,
      "theoretical_loss": 3.3512508586404732,
      "tokens_seen": 2976382976
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.913741223671014e-05,
      "loss": 1.2053,
      "theoretical_loss": 3.35124522440666,
      "tokens_seen": 2976448512
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.911735205616851e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.3512395903316357,
      "tokens_seen": 2976514048
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.909729187562689e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.3512339564153923,
      "tokens_seen": 2976579584
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.907723169508525e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.3512283226579225,
      "tokens_seen": 2976645120
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.905717151454363e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.3512226890592167,
      "tokens_seen": 2976710656
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.9037111334002e-05,
      "loss": 1.1878,
      "theoretical_loss": 3.351217055619269,
      "tokens_seen": 2976776192
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.901705115346038e-05,
      "loss": 1.3495,
      "theoretical_loss": 3.3512114223380696,
      "tokens_seen": 2976841728
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.899699097291875e-05,
      "loss": 1.1998,
      "theoretical_loss": 3.3512057892156117,
      "tokens_seen": 2976907264
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.897693079237713e-05,
      "loss": 1.2351,
      "theoretical_loss": 3.3512001562518874,
      "tokens_seen": 2976972800
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.89568706118355e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.3511945234468876,
      "tokens_seen": 2977038336
    },
    {
      "epoch": 0.8,
      "objective/train/docs_used": 1082519,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1863994598388672,
      "objective/train/theoretical_loss": 3.3511888908006053,
      "objective/train/tokens_used": 1347562976,
      "theoretical_loss": 3.3511888908006053,
      "tokens_seen": 2977103872
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.893681043129388e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.3511888908006053,
      "tokens_seen": 2977103872
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.891675025075226e-05,
      "loss": 1.2231,
      "theoretical_loss": 3.351183258313032,
      "tokens_seen": 2977169408
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.889669007021065e-05,
      "loss": 1.1329,
      "theoretical_loss": 3.3511776259841604,
      "tokens_seen": 2977234944
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.887662988966901e-05,
      "loss": 1.3585,
      "theoretical_loss": 3.3511719938139817,
      "tokens_seen": 2977300480
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.885656970912738e-05,
      "loss": 1.2183,
      "theoretical_loss": 3.3511663618024885,
      "tokens_seen": 2977366016
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.883650952858576e-05,
      "loss": 1.2542,
      "theoretical_loss": 3.351160729949673,
      "tokens_seen": 2977431552
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.881644934804414e-05,
      "loss": 1.2012,
      "theoretical_loss": 3.3511550982555267,
      "tokens_seen": 2977497088
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.879638916750251e-05,
      "loss": 1.35,
      "theoretical_loss": 3.351149466720042,
      "tokens_seen": 2977562624
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.877632898696089e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.3511438353432106,
      "tokens_seen": 2977628160
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.875626880641926e-05,
      "loss": 1.2489,
      "theoretical_loss": 3.351138204125025,
      "tokens_seen": 2977693696
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.873620862587764e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.351132573065477,
      "tokens_seen": 2977759232
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.871614844533601e-05,
      "loss": 1.1962,
      "theoretical_loss": 3.3511269421645586,
      "tokens_seen": 2977824768
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.869608826479439e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.3511213114222618,
      "tokens_seen": 2977890304
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.867602808425275e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.3511156808385785,
      "tokens_seen": 2977955840
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.865596790371113e-05,
      "loss": 1.1939,
      "theoretical_loss": 3.351110050413501,
      "tokens_seen": 2978021376
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.86359077231695e-05,
      "loss": 1.2191,
      "theoretical_loss": 3.351104420147022,
      "tokens_seen": 2978086912
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.861584754262788e-05,
      "loss": 1.3502,
      "theoretical_loss": 3.3510987900391322,
      "tokens_seen": 2978152448
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.859578736208626e-05,
      "loss": 1.1563,
      "theoretical_loss": 3.3510931600898246,
      "tokens_seen": 2978217984
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.857572718154463e-05,
      "loss": 1.3839,
      "theoretical_loss": 3.3510875302990906,
      "tokens_seen": 2978283520
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.855566700100301e-05,
      "loss": 1.244,
      "theoretical_loss": 3.3510819006669226,
      "tokens_seen": 2978349056
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.85356068204614e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3510762711933126,
      "tokens_seen": 2978414592
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.851554663991977e-05,
      "loss": 1.263,
      "theoretical_loss": 3.3510706418782528,
      "tokens_seen": 2978480128
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.849548645937815e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.3510650127217354,
      "tokens_seen": 2978545664
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.847542627883651e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.3510593837237517,
      "tokens_seen": 2978611200
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.845536609829489e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.3510537548842945,
      "tokens_seen": 2978676736
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.843530591775326e-05,
      "loss": 1.1418,
      "theoretical_loss": 3.3510481262033553,
      "tokens_seen": 2978742272
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.841524573721164e-05,
      "loss": 1.3473,
      "theoretical_loss": 3.3510424976809263,
      "tokens_seen": 2978807808
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.839518555667001e-05,
      "loss": 1.2249,
      "theoretical_loss": 3.3510368693170003,
      "tokens_seen": 2978873344
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.837512537612839e-05,
      "loss": 1.1512,
      "theoretical_loss": 3.351031241111568,
      "tokens_seen": 2978938880
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.835506519558677e-05,
      "loss": 1.2328,
      "theoretical_loss": 3.3510256130646225,
      "tokens_seen": 2979004416
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.833500501504514e-05,
      "loss": 1.3418,
      "theoretical_loss": 3.3510199851761553,
      "tokens_seen": 2979069952
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.831494483450352e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.351014357446159,
      "tokens_seen": 2979135488
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.829488465396188e-05,
      "loss": 1.2039,
      "theoretical_loss": 3.351008729874625,
      "tokens_seen": 2979201024
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.827482447342026e-05,
      "loss": 1.3313,
      "theoretical_loss": 3.3510031024615454,
      "tokens_seen": 2979266560
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.825476429287863e-05,
      "loss": 1.2903,
      "theoretical_loss": 3.350997475206913,
      "tokens_seen": 2979332096
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.823470411233701e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.350991848110719,
      "tokens_seen": 2979397632
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.821464393179538e-05,
      "loss": 1.2009,
      "theoretical_loss": 3.350986221172956,
      "tokens_seen": 2979463168
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.819458375125376e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.350980594393616,
      "tokens_seen": 2979528704
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.817452357071213e-05,
      "loss": 1.3119,
      "theoretical_loss": 3.3509749677726908,
      "tokens_seen": 2979594240
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.815446339017052e-05,
      "loss": 1.1786,
      "theoretical_loss": 3.3509693413101727,
      "tokens_seen": 2979659776
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.81344032096289e-05,
      "loss": 1.3683,
      "theoretical_loss": 3.3509637150060536,
      "tokens_seen": 2979725312
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.811434302908728e-05,
      "loss": 1.3236,
      "theoretical_loss": 3.3509580888603256,
      "tokens_seen": 2979790848
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.809428284854564e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.3509524628729803,
      "tokens_seen": 2979856384
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.807422266800401e-05,
      "loss": 1.3067,
      "theoretical_loss": 3.3509468370440105,
      "tokens_seen": 2979921920
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.805416248746239e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.3509412113734083,
      "tokens_seen": 2979987456
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.803410230692076e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.350935585861165,
      "tokens_seen": 2980052992
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.801404212637914e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.3509299605072735,
      "tokens_seen": 2980118528
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.799398194583752e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.350924335311725,
      "tokens_seen": 2980184064
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.797392176529589e-05,
      "loss": 1.3422,
      "theoretical_loss": 3.350918710274512,
      "tokens_seen": 2980249600
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.795386158475427e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.3509130853956273,
      "tokens_seen": 2980315136
    },
    {
      "epoch": 0.81,
      "objective/train/docs_used": 1083727,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2798722982406616,
      "objective/train/theoretical_loss": 3.3509074606750615,
      "objective/train/tokens_used": 1350839776,
      "theoretical_loss": 3.3509074606750615,
      "tokens_seen": 2980380672
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.793380140421264e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.3509074606750615,
      "tokens_seen": 2980380672
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.791374122367102e-05,
      "loss": 1.2086,
      "theoretical_loss": 3.350901836112808,
      "tokens_seen": 2980446208
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.789368104312938e-05,
      "loss": 1.287,
      "theoretical_loss": 3.3508962117088577,
      "tokens_seen": 2980511744
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.787362086258776e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.350890587463203,
      "tokens_seen": 2980577280
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.785356068204613e-05,
      "loss": 1.3216,
      "theoretical_loss": 3.350884963375837,
      "tokens_seen": 2980642816
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.783350050150451e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.3508793394467506,
      "tokens_seen": 2980708352
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.781344032096288e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.3508737156759363,
      "tokens_seen": 2980773888
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.779338014042126e-05,
      "loss": 1.2229,
      "theoretical_loss": 3.3508680920633855,
      "tokens_seen": 2980839424
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.777331995987965e-05,
      "loss": 1.2024,
      "theoretical_loss": 3.3508624686090913,
      "tokens_seen": 2980904960
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.775325977933803e-05,
      "loss": 1.1896,
      "theoretical_loss": 3.3508568453130456,
      "tokens_seen": 2980970496
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.77331995987964e-05,
      "loss": 1.2317,
      "theoretical_loss": 3.3508512221752396,
      "tokens_seen": 2981036032
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.771313941825476e-05,
      "loss": 1.1943,
      "theoretical_loss": 3.3508455991956665,
      "tokens_seen": 2981101568
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.769307923771314e-05,
      "loss": 1.195,
      "theoretical_loss": 3.3508399763743175,
      "tokens_seen": 2981167104
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.767301905717152e-05,
      "loss": 1.2353,
      "theoretical_loss": 3.350834353711185,
      "tokens_seen": 2981232640
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.765295887662989e-05,
      "loss": 1.3386,
      "theoretical_loss": 3.3508287312062612,
      "tokens_seen": 2981298176
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.763289869608827e-05,
      "loss": 1.278,
      "theoretical_loss": 3.350823108859538,
      "tokens_seen": 2981363712
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.761283851554664e-05,
      "loss": 1.332,
      "theoretical_loss": 3.3508174866710077,
      "tokens_seen": 2981429248
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.759277833500502e-05,
      "loss": 1.2715,
      "theoretical_loss": 3.3508118646406615,
      "tokens_seen": 2981494784
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.75727181544634e-05,
      "loss": 1.3201,
      "theoretical_loss": 3.350806242768493,
      "tokens_seen": 2981560320
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.755265797392177e-05,
      "loss": 1.3149,
      "theoretical_loss": 3.3508006210544927,
      "tokens_seen": 2981625856
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.753259779338015e-05,
      "loss": 1.1822,
      "theoretical_loss": 3.350794999498654,
      "tokens_seen": 2981691392
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.751253761283851e-05,
      "loss": 1.3545,
      "theoretical_loss": 3.350789378100968,
      "tokens_seen": 2981756928
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.749247743229688e-05,
      "loss": 1.2418,
      "theoretical_loss": 3.350783756861427,
      "tokens_seen": 2981822464
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.747241725175526e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.3507781357800237,
      "tokens_seen": 2981888000
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.745235707121364e-05,
      "loss": 1.2042,
      "theoretical_loss": 3.3507725148567493,
      "tokens_seen": 2981953536
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.743229689067201e-05,
      "loss": 1.3373,
      "theoretical_loss": 3.3507668940915964,
      "tokens_seen": 2982019072
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.741223671013039e-05,
      "loss": 1.3524,
      "theoretical_loss": 3.350761273484557,
      "tokens_seen": 2982084608
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.739217652958878e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.3507556530356233,
      "tokens_seen": 2982150144
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.737211634904715e-05,
      "loss": 1.311,
      "theoretical_loss": 3.350750032744787,
      "tokens_seen": 2982215680
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.735205616850553e-05,
      "loss": 1.277,
      "theoretical_loss": 3.3507444126120403,
      "tokens_seen": 2982281216
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.73319959879639e-05,
      "loss": 1.2577,
      "theoretical_loss": 3.3507387926373755,
      "tokens_seen": 2982346752
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.731193580742227e-05,
      "loss": 1.1941,
      "theoretical_loss": 3.3507331728207843,
      "tokens_seen": 2982412288
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.729187562688064e-05,
      "loss": 1.1898,
      "theoretical_loss": 3.3507275531622596,
      "tokens_seen": 2982477824
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.727181544633902e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.3507219336617924,
      "tokens_seen": 2982543360
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.72517552657974e-05,
      "loss": 1.2434,
      "theoretical_loss": 3.350716314319375,
      "tokens_seen": 2982608896
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.723169508525577e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.3507106951350005,
      "tokens_seen": 2982674432
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.721163490471415e-05,
      "loss": 1.3555,
      "theoretical_loss": 3.3507050761086603,
      "tokens_seen": 2982739968
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.719157472417252e-05,
      "loss": 1.2167,
      "theoretical_loss": 3.3506994572403457,
      "tokens_seen": 2982805504
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.71715145436309e-05,
      "loss": 1.2628,
      "theoretical_loss": 3.35069383853005,
      "tokens_seen": 2982871040
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.715145436308927e-05,
      "loss": 1.1934,
      "theoretical_loss": 3.350688219977765,
      "tokens_seen": 2982936576
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.713139418254765e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.350682601583482,
      "tokens_seen": 2983002112
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.711133400200601e-05,
      "loss": 1.1631,
      "theoretical_loss": 3.350676983347194,
      "tokens_seen": 2983067648
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.709127382146439e-05,
      "loss": 1.3327,
      "theoretical_loss": 3.3506713652688926,
      "tokens_seen": 2983133184
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.707121364092276e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.3506657473485704,
      "tokens_seen": 2983198720
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.705115346038114e-05,
      "loss": 1.1561,
      "theoretical_loss": 3.3506601295862186,
      "tokens_seen": 2983264256
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.703109327983953e-05,
      "loss": 1.2109,
      "theoretical_loss": 3.3506545119818303,
      "tokens_seen": 2983329792
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.70110330992979e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.350648894535397,
      "tokens_seen": 2983395328
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.699097291875628e-05,
      "loss": 1.2409,
      "theoretical_loss": 3.3506432772469106,
      "tokens_seen": 2983460864
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.697091273821465e-05,
      "loss": 1.1304,
      "theoretical_loss": 3.350637660116364,
      "tokens_seen": 2983526400
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.695085255767303e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.350632043143748,
      "tokens_seen": 2983591936
    },
    {
      "epoch": 0.81,
      "objective/train/docs_used": 1084846,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2808384895324707,
      "objective/train/theoretical_loss": 3.3506264263290566,
      "objective/train/tokens_used": 1354116576,
      "theoretical_loss": 3.3506264263290566,
      "tokens_seen": 2983657472
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.693079237713139e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.3506264263290566,
      "tokens_seen": 2983657472
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.691073219658977e-05,
      "loss": 1.2182,
      "theoretical_loss": 3.35062080967228,
      "tokens_seen": 2983723008
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.689067201604814e-05,
      "loss": 1.2261,
      "theoretical_loss": 3.350615193173411,
      "tokens_seen": 2983788544
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.687061183550652e-05,
      "loss": 1.3611,
      "theoretical_loss": 3.3506095768324418,
      "tokens_seen": 2983854080
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.68505516549649e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.350603960649365,
      "tokens_seen": 2983919616
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.683049147442327e-05,
      "loss": 1.207,
      "theoretical_loss": 3.3505983446241716,
      "tokens_seen": 2983985152
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.681043129388165e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.350592728756854,
      "tokens_seen": 2984050688
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.679037111334002e-05,
      "loss": 1.2148,
      "theoretical_loss": 3.350587113047405,
      "tokens_seen": 2984116224
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.67703109327984e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.350581497495816,
      "tokens_seen": 2984181760
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.675025075225677e-05,
      "loss": 1.2386,
      "theoretical_loss": 3.3505758821020795,
      "tokens_seen": 2984247296
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.673019057171514e-05,
      "loss": 1.2282,
      "theoretical_loss": 3.350570266866187,
      "tokens_seen": 2984312832
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.671013039117351e-05,
      "loss": 1.2705,
      "theoretical_loss": 3.3505646517881313,
      "tokens_seen": 2984378368
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.669007021063189e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.3505590368679044,
      "tokens_seen": 2984443904
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.667001003009026e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.3505534221054982,
      "tokens_seen": 2984509440
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.664994984954865e-05,
      "loss": 1.2434,
      "theoretical_loss": 3.3505478075009045,
      "tokens_seen": 2984574976
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.662988966900703e-05,
      "loss": 1.241,
      "theoretical_loss": 3.3505421930541157,
      "tokens_seen": 2984640512
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.66098294884654e-05,
      "loss": 1.3194,
      "theoretical_loss": 3.350536578765124,
      "tokens_seen": 2984706048
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.658976930792378e-05,
      "loss": 1.2329,
      "theoretical_loss": 3.350530964633921,
      "tokens_seen": 2984771584
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.656970912738216e-05,
      "loss": 1.1831,
      "theoretical_loss": 3.3505253506605,
      "tokens_seen": 2984837120
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.654964894684053e-05,
      "loss": 1.307,
      "theoretical_loss": 3.350519736844852,
      "tokens_seen": 2984902656
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.65295887662989e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.350514123186969,
      "tokens_seen": 2984968192
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.650952858575727e-05,
      "loss": 1.2113,
      "theoretical_loss": 3.350508509686844,
      "tokens_seen": 2985033728
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.648946840521565e-05,
      "loss": 1.1169,
      "theoretical_loss": 3.3505028963444685,
      "tokens_seen": 2985099264
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.646940822467402e-05,
      "loss": 1.237,
      "theoretical_loss": 3.3504972831598345,
      "tokens_seen": 2985164800
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.64493480441324e-05,
      "loss": 1.2115,
      "theoretical_loss": 3.3504916701329344,
      "tokens_seen": 2985230336
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.642928786359077e-05,
      "loss": 1.179,
      "theoretical_loss": 3.3504860572637605,
      "tokens_seen": 2985295872
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.640922768304915e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.350480444552304,
      "tokens_seen": 2985361408
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.638916750250753e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.3504748319985582,
      "tokens_seen": 2985426944
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.63691073219659e-05,
      "loss": 1.2363,
      "theoretical_loss": 3.3504692196025143,
      "tokens_seen": 2985492480
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.634904714142426e-05,
      "loss": 1.1735,
      "theoretical_loss": 3.350463607364165,
      "tokens_seen": 2985558016
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.632898696088264e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.3504579952835023,
      "tokens_seen": 2985623552
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.630892678034102e-05,
      "loss": 1.1297,
      "theoretical_loss": 3.3504523833605178,
      "tokens_seen": 2985689088
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.628886659979939e-05,
      "loss": 1.1864,
      "theoretical_loss": 3.350446771595204,
      "tokens_seen": 2985754624
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.626880641925778e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.350441159987553,
      "tokens_seen": 2985820160
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.624874623871616e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.350435548537557,
      "tokens_seen": 2985885696
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.622868605817453e-05,
      "loss": 1.1237,
      "theoretical_loss": 3.350429937245208,
      "tokens_seen": 2985951232
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.620862587763291e-05,
      "loss": 1.357,
      "theoretical_loss": 3.350424326110498,
      "tokens_seen": 2986016768
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.618856569709128e-05,
      "loss": 1.19,
      "theoretical_loss": 3.3504187151334195,
      "tokens_seen": 2986082304
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.616850551654966e-05,
      "loss": 1.2234,
      "theoretical_loss": 3.350413104313964,
      "tokens_seen": 2986147840
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.614844533600802e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.350407493652124,
      "tokens_seen": 2986213376
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.61283851554664e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.350401883147892,
      "tokens_seen": 2986278912
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.610832497492477e-05,
      "loss": 1.297,
      "theoretical_loss": 3.3503962728012593,
      "tokens_seen": 2986344448
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.608826479438315e-05,
      "loss": 1.1948,
      "theoretical_loss": 3.3503906626122184,
      "tokens_seen": 2986409984
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.606820461384152e-05,
      "loss": 1.209,
      "theoretical_loss": 3.3503850525807612,
      "tokens_seen": 2986475520
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.60481444332999e-05,
      "loss": 1.1895,
      "theoretical_loss": 3.3503794427068803,
      "tokens_seen": 2986541056
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.602808425275828e-05,
      "loss": 1.3712,
      "theoretical_loss": 3.350373832990568,
      "tokens_seen": 2986606592
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.600802407221665e-05,
      "loss": 1.222,
      "theoretical_loss": 3.3503682234318153,
      "tokens_seen": 2986672128
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.598796389167503e-05,
      "loss": 1.2072,
      "theoretical_loss": 3.3503626140306153,
      "tokens_seen": 2986737664
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.59679037111334e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.3503570047869595,
      "tokens_seen": 2986803200
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.594784353059177e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.3503513957008404,
      "tokens_seen": 2986868736
    },
    {
      "epoch": 0.81,
      "objective/train/docs_used": 1086010,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2536039352416992,
      "objective/train/theoretical_loss": 3.35034578677225,
      "objective/train/tokens_used": 1357393376,
      "theoretical_loss": 3.35034578677225,
      "tokens_seen": 2986934272
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.592778335005014e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.35034578677225,
      "tokens_seen": 2986934272
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.590772316950853e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.3503401780011806,
      "tokens_seen": 2986999808
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.588766298896691e-05,
      "loss": 1.2164,
      "theoretical_loss": 3.3503345693876243,
      "tokens_seen": 2987065344
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.586760280842528e-05,
      "loss": 1.3118,
      "theoretical_loss": 3.3503289609315727,
      "tokens_seen": 2987130880
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.584754262788366e-05,
      "loss": 1.3303,
      "theoretical_loss": 3.3503233526330183,
      "tokens_seen": 2987196416
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.582748244734203e-05,
      "loss": 1.3471,
      "theoretical_loss": 3.350317744491954,
      "tokens_seen": 2987261952
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.580742226680041e-05,
      "loss": 1.1706,
      "theoretical_loss": 3.35031213650837,
      "tokens_seen": 2987327488
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.578736208625879e-05,
      "loss": 1.1709,
      "theoretical_loss": 3.3503065286822604,
      "tokens_seen": 2987393024
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.576730190571716e-05,
      "loss": 1.2586,
      "theoretical_loss": 3.3503009210136163,
      "tokens_seen": 2987458560
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.574724172517552e-05,
      "loss": 1.3649,
      "theoretical_loss": 3.35029531350243,
      "tokens_seen": 2987524096
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.57271815446339e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.3502897061486934,
      "tokens_seen": 2987589632
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.570712136409228e-05,
      "loss": 1.2332,
      "theoretical_loss": 3.3502840989523994,
      "tokens_seen": 2987655168
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.568706118355065e-05,
      "loss": 1.275,
      "theoretical_loss": 3.3502784919135395,
      "tokens_seen": 2987720704
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.566700100300903e-05,
      "loss": 1.324,
      "theoretical_loss": 3.3502728850321057,
      "tokens_seen": 2987786240
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.56469408224674e-05,
      "loss": 1.221,
      "theoretical_loss": 3.3502672783080905,
      "tokens_seen": 2987851776
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.562688064192578e-05,
      "loss": 1.2474,
      "theoretical_loss": 3.3502616717414857,
      "tokens_seen": 2987917312
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.560682046138415e-05,
      "loss": 1.2294,
      "theoretical_loss": 3.3502560653322835,
      "tokens_seen": 2987982848
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.558676028084253e-05,
      "loss": 1.2732,
      "theoretical_loss": 3.3502504590804767,
      "tokens_seen": 2988048384
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.556670010030089e-05,
      "loss": 1.2183,
      "theoretical_loss": 3.3502448529860565,
      "tokens_seen": 2988113920
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.554663991975927e-05,
      "loss": 1.2175,
      "theoretical_loss": 3.350239247049015,
      "tokens_seen": 2988179456
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.552657973921766e-05,
      "loss": 1.2108,
      "theoretical_loss": 3.3502336412693454,
      "tokens_seen": 2988244992
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.550651955867603e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.350228035647039,
      "tokens_seen": 2988310528
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.548645937813441e-05,
      "loss": 1.306,
      "theoretical_loss": 3.350222430182088,
      "tokens_seen": 2988376064
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.546639919759278e-05,
      "loss": 1.1677,
      "theoretical_loss": 3.350216824874485,
      "tokens_seen": 2988441600
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.544633901705116e-05,
      "loss": 1.1502,
      "theoretical_loss": 3.3502112197242213,
      "tokens_seen": 2988507136
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.542627883650954e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.3502056147312897,
      "tokens_seen": 2988572672
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.540621865596791e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.3502000098956817,
      "tokens_seen": 2988638208
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.538615847542629e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.3501944052173904,
      "tokens_seen": 2988703744
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.536609829488465e-05,
      "loss": 1.1738,
      "theoretical_loss": 3.350188800696407,
      "tokens_seen": 2988769280
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.534603811434303e-05,
      "loss": 1.3936,
      "theoretical_loss": 3.350183196332724,
      "tokens_seen": 2988834816
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.53259779338014e-05,
      "loss": 1.1619,
      "theoretical_loss": 3.350177592126334,
      "tokens_seen": 2988900352
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.530591775325978e-05,
      "loss": 1.1933,
      "theoretical_loss": 3.3501719880772285,
      "tokens_seen": 2988965888
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.528585757271815e-05,
      "loss": 1.2011,
      "theoretical_loss": 3.3501663841854,
      "tokens_seen": 2989031424
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.526579739217653e-05,
      "loss": 1.2437,
      "theoretical_loss": 3.3501607804508406,
      "tokens_seen": 2989096960
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.52457372116349e-05,
      "loss": 1.2139,
      "theoretical_loss": 3.3501551768735416,
      "tokens_seen": 2989162496
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.522567703109328e-05,
      "loss": 1.2065,
      "theoretical_loss": 3.3501495734534963,
      "tokens_seen": 2989228032
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.520561685055166e-05,
      "loss": 1.3012,
      "theoretical_loss": 3.3501439701906968,
      "tokens_seen": 2989293568
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.518555667001003e-05,
      "loss": 1.374,
      "theoretical_loss": 3.350138367085134,
      "tokens_seen": 2989359104
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.51654964894684e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.3501327641368013,
      "tokens_seen": 2989424640
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.514543630892678e-05,
      "loss": 1.1764,
      "theoretical_loss": 3.3501271613456907,
      "tokens_seen": 2989490176
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.512537612838516e-05,
      "loss": 1.1968,
      "theoretical_loss": 3.350121558711794,
      "tokens_seen": 2989555712
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.510531594784354e-05,
      "loss": 1.2142,
      "theoretical_loss": 3.350115956235103,
      "tokens_seen": 2989621248
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.508525576730191e-05,
      "loss": 1.343,
      "theoretical_loss": 3.3501103539156105,
      "tokens_seen": 2989686784
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.506519558676029e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.350104751753308,
      "tokens_seen": 2989752320
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.504513540621866e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.3500991497481887,
      "tokens_seen": 2989817856
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.502507522567704e-05,
      "loss": 1.241,
      "theoretical_loss": 3.3500935479002436,
      "tokens_seen": 2989883392
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.500501504513541e-05,
      "loss": 1.2046,
      "theoretical_loss": 3.3500879462094657,
      "tokens_seen": 2989948928
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.498495486459379e-05,
      "loss": 1.2391,
      "theoretical_loss": 3.3500823446758465,
      "tokens_seen": 2990014464
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.496489468405215e-05,
      "loss": 1.246,
      "theoretical_loss": 3.350076743299378,
      "tokens_seen": 2990080000
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.494483450351053e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.3500711420800533,
      "tokens_seen": 2990145536
    },
    {
      "epoch": 0.81,
      "objective/train/docs_used": 1087207,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.278800368309021,
      "objective/train/theoretical_loss": 3.350065541017864,
      "objective/train/tokens_used": 1360670176,
      "theoretical_loss": 3.350065541017864,
      "tokens_seen": 2990211072
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.49247743229689e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.350065541017864,
      "tokens_seen": 2990211072
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.490471414242728e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.350059940112802,
      "tokens_seen": 2990276608
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.488465396188566e-05,
      "loss": 1.247,
      "theoretical_loss": 3.35005433936486,
      "tokens_seen": 2990342144
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.486459378134403e-05,
      "loss": 1.2122,
      "theoretical_loss": 3.3500487387740296,
      "tokens_seen": 2990407680
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.484453360080241e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.3500431383403035,
      "tokens_seen": 2990473216
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.482447342026078e-05,
      "loss": 1.1483,
      "theoretical_loss": 3.3500375380636735,
      "tokens_seen": 2990538752
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.480441323971916e-05,
      "loss": 1.2013,
      "theoretical_loss": 3.3500319379441317,
      "tokens_seen": 2990604288
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.478435305917752e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.3500263379816704,
      "tokens_seen": 2990669824
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.476429287863591e-05,
      "loss": 1.1882,
      "theoretical_loss": 3.3500207381762817,
      "tokens_seen": 2990735360
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.474423269809429e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.3500151385279575,
      "tokens_seen": 2990800896
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.472417251755266e-05,
      "loss": 1.4778,
      "theoretical_loss": 3.3500095390366904,
      "tokens_seen": 2990866432
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.470411233701104e-05,
      "loss": 1.2052,
      "theoretical_loss": 3.3500039397024723,
      "tokens_seen": 2990931968
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.468405215646941e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.3499983405252958,
      "tokens_seen": 2990997504
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.466399197592779e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.3499927415051522,
      "tokens_seen": 2991063040
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.464393179538617e-05,
      "loss": 1.2132,
      "theoretical_loss": 3.3499871426420347,
      "tokens_seen": 2991128576
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.462387161484454e-05,
      "loss": 1.1422,
      "theoretical_loss": 3.349981543935934,
      "tokens_seen": 2991194112
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.460381143430292e-05,
      "loss": 1.1107,
      "theoretical_loss": 3.349975945386844,
      "tokens_seen": 2991259648
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.458375125376128e-05,
      "loss": 1.1999,
      "theoretical_loss": 3.3499703469947555,
      "tokens_seen": 2991325184
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.456369107321966e-05,
      "loss": 1.2622,
      "theoretical_loss": 3.3499647487596613,
      "tokens_seen": 2991390720
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.454363089267803e-05,
      "loss": 1.2387,
      "theoretical_loss": 3.3499591506815536,
      "tokens_seen": 2991456256
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.45235707121364e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.3499535527604243,
      "tokens_seen": 2991521792
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.450351053159478e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.3499479549962654,
      "tokens_seen": 2991587328
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.448345035105316e-05,
      "loss": 1.108,
      "theoretical_loss": 3.3499423573890694,
      "tokens_seen": 2991652864
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.446339017051153e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.349936759938829,
      "tokens_seen": 2991718400
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.444332998996991e-05,
      "loss": 1.3502,
      "theoretical_loss": 3.3499311626455346,
      "tokens_seen": 2991783936
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.442326980942829e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.34992556550918,
      "tokens_seen": 2991849472
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.440320962888667e-05,
      "loss": 1.2301,
      "theoretical_loss": 3.3499199685297567,
      "tokens_seen": 2991915008
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.438314944834504e-05,
      "loss": 1.1726,
      "theoretical_loss": 3.349914371707257,
      "tokens_seen": 2991980544
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.436308926780341e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.3499087750416736,
      "tokens_seen": 2992046080
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.434302908726179e-05,
      "loss": 1.217,
      "theoretical_loss": 3.3499031785329976,
      "tokens_seen": 2992111616
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.432296890672016e-05,
      "loss": 1.1684,
      "theoretical_loss": 3.349897582181222,
      "tokens_seen": 2992177152
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.430290872617854e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.3498919859863383,
      "tokens_seen": 2992242688
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.428284854563692e-05,
      "loss": 1.2296,
      "theoretical_loss": 3.349886389948339,
      "tokens_seen": 2992308224
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.426278836509529e-05,
      "loss": 1.1951,
      "theoretical_loss": 3.349880794067217,
      "tokens_seen": 2992373760
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.424272818455367e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.3498751983429633,
      "tokens_seen": 2992439296
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.422266800401204e-05,
      "loss": 1.1607,
      "theoretical_loss": 3.3498696027755703,
      "tokens_seen": 2992504832
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.420260782347042e-05,
      "loss": 1.177,
      "theoretical_loss": 3.3498640073650305,
      "tokens_seen": 2992570368
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.418254764292878e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.349858412111336,
      "tokens_seen": 2992635904
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.416248746238716e-05,
      "loss": 1.1972,
      "theoretical_loss": 3.349852817014479,
      "tokens_seen": 2992701440
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.414242728184553e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3498472220744517,
      "tokens_seen": 2992766976
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.412236710130391e-05,
      "loss": 1.2316,
      "theoretical_loss": 3.349841627291246,
      "tokens_seen": 2992832512
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.410230692076228e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.349836032664854,
      "tokens_seen": 2992898048
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.408224674022066e-05,
      "loss": 1.1049,
      "theoretical_loss": 3.3498304381952684,
      "tokens_seen": 2992963584
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.406218655967904e-05,
      "loss": 1.2731,
      "theoretical_loss": 3.349824843882481,
      "tokens_seen": 2993029120
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.404212637913741e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.3498192497264845,
      "tokens_seen": 2993094656
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.40220661985958e-05,
      "loss": 1.1476,
      "theoretical_loss": 3.34981365572727,
      "tokens_seen": 2993160192
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.400200601805416e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.3498080618848305,
      "tokens_seen": 2993225728
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.398194583751254e-05,
      "loss": 1.21,
      "theoretical_loss": 3.349802468199158,
      "tokens_seen": 2993291264
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.396188565697092e-05,
      "loss": 1.196,
      "theoretical_loss": 3.349796874670245,
      "tokens_seen": 2993356800
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.394182547642929e-05,
      "loss": 1.4338,
      "theoretical_loss": 3.349791281298083,
      "tokens_seen": 2993422336
    },
    {
      "epoch": 0.81,
      "objective/train/docs_used": 1088408,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.320845365524292,
      "objective/train/theoretical_loss": 3.3497856880826644,
      "objective/train/tokens_used": 1363946976,
      "theoretical_loss": 3.3497856880826644,
      "tokens_seen": 2993487872
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.392176529588767e-05,
      "loss": 1.284,
      "theoretical_loss": 3.3497856880826644,
      "tokens_seen": 2993487872
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.390170511534604e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.3497800950239816,
      "tokens_seen": 2993553408
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.388164493480442e-05,
      "loss": 1.3187,
      "theoretical_loss": 3.3497745021220267,
      "tokens_seen": 2993618944
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.38615847542628e-05,
      "loss": 1.3896,
      "theoretical_loss": 3.3497689093767917,
      "tokens_seen": 2993684480
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.384152457372117e-05,
      "loss": 1.2537,
      "theoretical_loss": 3.3497633167882688,
      "tokens_seen": 2993750016
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.382146439317955e-05,
      "loss": 1.2819,
      "theoretical_loss": 3.3497577243564507,
      "tokens_seen": 2993815552
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.380140421263791e-05,
      "loss": 1.2074,
      "theoretical_loss": 3.3497521320813286,
      "tokens_seen": 2993881088
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.378134403209628e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.3497465399628954,
      "tokens_seen": 2993946624
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.376128385155466e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.3497409480011435,
      "tokens_seen": 2994012160
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.374122367101304e-05,
      "loss": 1.1889,
      "theoretical_loss": 3.3497353561960646,
      "tokens_seen": 2994077696
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.372116349047141e-05,
      "loss": 1.298,
      "theoretical_loss": 3.3497297645476505,
      "tokens_seen": 2994143232
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.370110330992979e-05,
      "loss": 1.3405,
      "theoretical_loss": 3.349724173055894,
      "tokens_seen": 2994208768
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.368104312938816e-05,
      "loss": 1.3388,
      "theoretical_loss": 3.3497185817207877,
      "tokens_seen": 2994274304
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.366098294884654e-05,
      "loss": 1.1593,
      "theoretical_loss": 3.349712990542323,
      "tokens_seen": 2994339840
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.364092276830493e-05,
      "loss": 1.1546,
      "theoretical_loss": 3.3497073995204922,
      "tokens_seen": 2994405376
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.36208625877633e-05,
      "loss": 1.287,
      "theoretical_loss": 3.3497018086552877,
      "tokens_seen": 2994470912
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.360080240722167e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.3496962179467014,
      "tokens_seen": 2994536448
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.358074222668004e-05,
      "loss": 1.2287,
      "theoretical_loss": 3.349690627394726,
      "tokens_seen": 2994601984
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.356068204613842e-05,
      "loss": 1.2284,
      "theoretical_loss": 3.349685036999353,
      "tokens_seen": 2994667520
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.35406218655968e-05,
      "loss": 1.3419,
      "theoretical_loss": 3.349679446760575,
      "tokens_seen": 2994733056
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.352056168505517e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.3496738566783844,
      "tokens_seen": 2994798592
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.350050150451354e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.3496682667527726,
      "tokens_seen": 2994864128
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.348044132397192e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.3496626769837325,
      "tokens_seen": 2994929664
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.34603811434303e-05,
      "loss": 1.4221,
      "theoretical_loss": 3.3496570873712566,
      "tokens_seen": 2994995200
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.344032096288867e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.3496514979153362,
      "tokens_seen": 2995060736
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.342026078234703e-05,
      "loss": 1.1576,
      "theoretical_loss": 3.3496459086159636,
      "tokens_seen": 2995126272
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.340020060180541e-05,
      "loss": 1.4138,
      "theoretical_loss": 3.3496403194731315,
      "tokens_seen": 2995191808
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.338014042126379e-05,
      "loss": 1.3926,
      "theoretical_loss": 3.349634730486832,
      "tokens_seen": 2995257344
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.336008024072216e-05,
      "loss": 1.2614,
      "theoretical_loss": 3.349629141657057,
      "tokens_seen": 2995322880
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.334002006018054e-05,
      "loss": 1.1885,
      "theoretical_loss": 3.349623552983799,
      "tokens_seen": 2995388416
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.331995987963891e-05,
      "loss": 1.2733,
      "theoretical_loss": 3.3496179644670496,
      "tokens_seen": 2995453952
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.329989969909729e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.3496123761068017,
      "tokens_seen": 2995519488
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.327983951855568e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.3496067879030473,
      "tokens_seen": 2995585024
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.325977933801405e-05,
      "loss": 1.246,
      "theoretical_loss": 3.3496011998557784,
      "tokens_seen": 2995650560
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.323971915747243e-05,
      "loss": 1.2428,
      "theoretical_loss": 3.3495956119649875,
      "tokens_seen": 2995716096
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.321965897693079e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.3495900242306664,
      "tokens_seen": 2995781632
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.319959879638917e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.3495844366528074,
      "tokens_seen": 2995847168
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.317953861584754e-05,
      "loss": 1.2095,
      "theoretical_loss": 3.349578849231403,
      "tokens_seen": 2995912704
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.315947843530592e-05,
      "loss": 1.2499,
      "theoretical_loss": 3.3495732619664453,
      "tokens_seen": 2995978240
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.31394182547643e-05,
      "loss": 1.2268,
      "theoretical_loss": 3.349567674857926,
      "tokens_seen": 2996043776
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.311935807422267e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.349562087905838,
      "tokens_seen": 2996109312
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.309929789368105e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.3495565011101727,
      "tokens_seen": 2996174848
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.307923771313942e-05,
      "loss": 1.1999,
      "theoretical_loss": 3.3495509144709232,
      "tokens_seen": 2996240384
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.30591775325978e-05,
      "loss": 1.1939,
      "theoretical_loss": 3.349545327988081,
      "tokens_seen": 2996305920
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.303911735205617e-05,
      "loss": 1.2167,
      "theoretical_loss": 3.3495397416616393,
      "tokens_seen": 2996371456
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.301905717151454e-05,
      "loss": 1.3351,
      "theoretical_loss": 3.3495341554915887,
      "tokens_seen": 2996436992
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.299899699097291e-05,
      "loss": 1.1796,
      "theoretical_loss": 3.349528569477923,
      "tokens_seen": 2996502528
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.297893681043129e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.3495229836206333,
      "tokens_seen": 2996568064
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.295887662988966e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.349517397919712,
      "tokens_seen": 2996633600
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.293881644934804e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.3495118123751517,
      "tokens_seen": 2996699136
    },
    {
      "epoch": 0.82,
      "objective/train/docs_used": 1089453,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2000184059143066,
      "objective/train/theoretical_loss": 3.3495062269869447,
      "objective/train/tokens_used": 1367223776,
      "theoretical_loss": 3.3495062269869447,
      "tokens_seen": 2996764672
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.291875626880642e-05,
      "loss": 1.254,
      "theoretical_loss": 3.3495062269869447,
      "tokens_seen": 2996764672
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.28986960882648e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.3495006417550823,
      "tokens_seen": 2996830208
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.287863590772318e-05,
      "loss": 1.4098,
      "theoretical_loss": 3.3494950566795576,
      "tokens_seen": 2996895744
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.285857572718156e-05,
      "loss": 1.1957,
      "theoretical_loss": 3.3494894717603625,
      "tokens_seen": 2996961280
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.283851554663993e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.3494838869974894,
      "tokens_seen": 2997026816
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.28184553660983e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.34947830239093,
      "tokens_seen": 2997092352
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.279839518555667e-05,
      "loss": 1.187,
      "theoretical_loss": 3.349472717940677,
      "tokens_seen": 2997157888
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.277833500501505e-05,
      "loss": 1.2107,
      "theoretical_loss": 3.3494671336467223,
      "tokens_seen": 2997223424
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.275827482447342e-05,
      "loss": 1.315,
      "theoretical_loss": 3.3494615495090585,
      "tokens_seen": 2997288960
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.27382146439318e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.349455965527677,
      "tokens_seen": 2997354496
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.271815446339017e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.3494503817025714,
      "tokens_seen": 2997420032
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.269809428284855e-05,
      "loss": 1.3273,
      "theoretical_loss": 3.3494447980337325,
      "tokens_seen": 2997485568
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.267803410230693e-05,
      "loss": 1.2116,
      "theoretical_loss": 3.349439214521153,
      "tokens_seen": 2997551104
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.26579739217653e-05,
      "loss": 1.1808,
      "theoretical_loss": 3.3494336311648256,
      "tokens_seen": 2997616640
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.263791374122366e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.349428047964742,
      "tokens_seen": 2997682176
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.261785356068204e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.3494224649208943,
      "tokens_seen": 2997747712
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.259779338014041e-05,
      "loss": 1.2095,
      "theoretical_loss": 3.349416882033275,
      "tokens_seen": 2997813248
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.257773319959879e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.3494112993018765,
      "tokens_seen": 2997878784
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.255767301905717e-05,
      "loss": 1.2139,
      "theoretical_loss": 3.3494057167266904,
      "tokens_seen": 2997944320
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.253761283851554e-05,
      "loss": 1.2625,
      "theoretical_loss": 3.3494001343077096,
      "tokens_seen": 2998009856
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.251755265797393e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.3493945520449255,
      "tokens_seen": 2998075392
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.249749247743231e-05,
      "loss": 1.102,
      "theoretical_loss": 3.3493889699383312,
      "tokens_seen": 2998140928
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.247743229689068e-05,
      "loss": 1.2317,
      "theoretical_loss": 3.3493833879879182,
      "tokens_seen": 2998206464
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.245737211634906e-05,
      "loss": 1.2532,
      "theoretical_loss": 3.3493778061936794,
      "tokens_seen": 2998272000
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.243731193580742e-05,
      "loss": 1.3422,
      "theoretical_loss": 3.3493722245556063,
      "tokens_seen": 2998337536
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.24172517552658e-05,
      "loss": 1.4117,
      "theoretical_loss": 3.349366643073692,
      "tokens_seen": 2998403072
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.239719157472417e-05,
      "loss": 1.2225,
      "theoretical_loss": 3.3493610617479277,
      "tokens_seen": 2998468608
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.237713139418255e-05,
      "loss": 1.265,
      "theoretical_loss": 3.349355480578306,
      "tokens_seen": 2998534144
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.235707121364092e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.3493498995648197,
      "tokens_seen": 2998599680
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.23370110330993e-05,
      "loss": 1.3106,
      "theoretical_loss": 3.3493443187074603,
      "tokens_seen": 2998665216
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.231695085255768e-05,
      "loss": 1.2469,
      "theoretical_loss": 3.34933873800622,
      "tokens_seen": 2998730752
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.229689067201605e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.349333157461092,
      "tokens_seen": 2998796288
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.227683049147443e-05,
      "loss": 1.2008,
      "theoretical_loss": 3.349327577072067,
      "tokens_seen": 2998861824
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.22567703109328e-05,
      "loss": 1.2446,
      "theoretical_loss": 3.349321996839138,
      "tokens_seen": 2998927360
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.223671013039117e-05,
      "loss": 1.1966,
      "theoretical_loss": 3.349316416762298,
      "tokens_seen": 2998992896
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.221664994984954e-05,
      "loss": 1.26,
      "theoretical_loss": 3.3493108368415383,
      "tokens_seen": 2999058432
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.219658976930792e-05,
      "loss": 1.2055,
      "theoretical_loss": 3.3493052570768507,
      "tokens_seen": 2999123968
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.217652958876629e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.3492996774682284,
      "tokens_seen": 2999189504
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.215646940822467e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.3492940980156636,
      "tokens_seen": 2999255040
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.213640922768306e-05,
      "loss": 1.2008,
      "theoretical_loss": 3.3492885187191477,
      "tokens_seen": 2999320576
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.211634904714143e-05,
      "loss": 1.2375,
      "theoretical_loss": 3.3492829395786736,
      "tokens_seen": 2999386112
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.209628886659981e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.349277360594233,
      "tokens_seen": 2999451648
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.207622868605819e-05,
      "loss": 1.1091,
      "theoretical_loss": 3.3492717817658186,
      "tokens_seen": 2999517184
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.205616850551656e-05,
      "loss": 1.1626,
      "theoretical_loss": 3.3492662030934226,
      "tokens_seen": 2999582720
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.203610832497492e-05,
      "loss": 1.1813,
      "theoretical_loss": 3.3492606245770373,
      "tokens_seen": 2999648256
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.20160481444333e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.3492550462166544,
      "tokens_seen": 2999713792
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.199598796389168e-05,
      "loss": 1.1508,
      "theoretical_loss": 3.3492494680122666,
      "tokens_seen": 2999779328
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.197592778335005e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.3492438899638657,
      "tokens_seen": 2999844864
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.195586760280843e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.349238312071445,
      "tokens_seen": 2999910400
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.19358074222668e-05,
      "loss": 1.2251,
      "theoretical_loss": 3.3492327343349952,
      "tokens_seen": 2999975936
    },
    {
      "epoch": 0.82,
      "objective/train/docs_used": 1090669,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.322437047958374,
      "objective/train/theoretical_loss": 3.3492271567545093,
      "objective/train/tokens_used": 1370500576,
      "theoretical_loss": 3.3492271567545093,
      "tokens_seen": 3000041472
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.191574724172518e-05,
      "loss": 1.402,
      "theoretical_loss": 3.3492271567545093,
      "tokens_seen": 3000041472
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.189568706118355e-05,
      "loss": 1.1713,
      "theoretical_loss": 3.34922157932998,
      "tokens_seen": 3000107008
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.187562688064193e-05,
      "loss": 1.1902,
      "theoretical_loss": 3.3492160020613984,
      "tokens_seen": 3000172544
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.185556670010029e-05,
      "loss": 1.2054,
      "theoretical_loss": 3.349210424948758,
      "tokens_seen": 3000238080
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.183550651955867e-05,
      "loss": 1.2394,
      "theoretical_loss": 3.34920484799205,
      "tokens_seen": 3000303616
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.181544633901704e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.3491992711912673,
      "tokens_seen": 3000369152
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.179538615847542e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.3491936945464023,
      "tokens_seen": 3000434688
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.177532597793381e-05,
      "loss": 1.3449,
      "theoretical_loss": 3.349188118057446,
      "tokens_seen": 3000500224
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.175526579739218e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.349182541724392,
      "tokens_seen": 3000565760
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.173520561685056e-05,
      "loss": 1.2459,
      "theoretical_loss": 3.3491769655472314,
      "tokens_seen": 3000631296
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.171514543630894e-05,
      "loss": 1.223,
      "theoretical_loss": 3.3491713895259574,
      "tokens_seen": 3000696832
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.169508525576731e-05,
      "loss": 1.1506,
      "theoretical_loss": 3.349165813660562,
      "tokens_seen": 3000762368
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.167502507522569e-05,
      "loss": 1.1472,
      "theoretical_loss": 3.3491602379510375,
      "tokens_seen": 3000827904
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.165496489468405e-05,
      "loss": 1.2445,
      "theoretical_loss": 3.3491546623973756,
      "tokens_seen": 3000893440
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.163490471414243e-05,
      "loss": 1.3146,
      "theoretical_loss": 3.349149086999569,
      "tokens_seen": 3000958976
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.16148445336008e-05,
      "loss": 1.3349,
      "theoretical_loss": 3.3491435117576094,
      "tokens_seen": 3001024512
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.159478435305918e-05,
      "loss": 1.4089,
      "theoretical_loss": 3.34913793667149,
      "tokens_seen": 3001090048
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.157472417251755e-05,
      "loss": 1.1726,
      "theoretical_loss": 3.3491323617412023,
      "tokens_seen": 3001155584
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.155466399197593e-05,
      "loss": 1.2012,
      "theoretical_loss": 3.349126786966739,
      "tokens_seen": 3001221120
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.15346038114343e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.349121212348092,
      "tokens_seen": 3001286656
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.151454363089268e-05,
      "loss": 1.3385,
      "theoretical_loss": 3.3491156378852533,
      "tokens_seen": 3001352192
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.149448345035106e-05,
      "loss": 1.0968,
      "theoretical_loss": 3.3491100635782156,
      "tokens_seen": 3001417728
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.147442326980943e-05,
      "loss": 1.1905,
      "theoretical_loss": 3.349104489426971,
      "tokens_seen": 3001483264
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.14543630892678e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.349098915431512,
      "tokens_seen": 3001548800
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.143430290872617e-05,
      "loss": 1.1305,
      "theoretical_loss": 3.3490933415918303,
      "tokens_seen": 3001614336
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.141424272818455e-05,
      "loss": 1.1321,
      "theoretical_loss": 3.349087767907919,
      "tokens_seen": 3001679872
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.139418254764294e-05,
      "loss": 1.2404,
      "theoretical_loss": 3.3490821943797693,
      "tokens_seen": 3001745408
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.137412236710131e-05,
      "loss": 1.3083,
      "theoretical_loss": 3.349076621007374,
      "tokens_seen": 3001810944
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.135406218655969e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.3490710477907255,
      "tokens_seen": 3001876480
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.133400200601806e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.3490654747298154,
      "tokens_seen": 3001942016
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.131394182547644e-05,
      "loss": 1.2181,
      "theoretical_loss": 3.3490599018246368,
      "tokens_seen": 3002007552
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.129388164493481e-05,
      "loss": 1.3606,
      "theoretical_loss": 3.3490543290751815,
      "tokens_seen": 3002073088
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.127382146439319e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.3490487564814417,
      "tokens_seen": 3002138624
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.125376128385155e-05,
      "loss": 1.3946,
      "theoretical_loss": 3.34904318404341,
      "tokens_seen": 3002204160
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.123370110330993e-05,
      "loss": 1.2212,
      "theoretical_loss": 3.3490376117610783,
      "tokens_seen": 3002269696
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.12136409227683e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3490320396344386,
      "tokens_seen": 3002335232
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.119358074222668e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.3490264676634833,
      "tokens_seen": 3002400768
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.117352056168506e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.3490208958482057,
      "tokens_seen": 3002466304
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.115346038114343e-05,
      "loss": 1.3479,
      "theoretical_loss": 3.3490153241885965,
      "tokens_seen": 3002531840
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.113340020060181e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.349009752684649,
      "tokens_seen": 3002597376
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.111334002006018e-05,
      "loss": 1.2547,
      "theoretical_loss": 3.3490041813363547,
      "tokens_seen": 3002662912
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.109327983951856e-05,
      "loss": 1.3736,
      "theoretical_loss": 3.3489986101437066,
      "tokens_seen": 3002728448
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.107321965897692e-05,
      "loss": 1.2581,
      "theoretical_loss": 3.3489930391066967,
      "tokens_seen": 3002793984
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.10531594784353e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.348987468225317,
      "tokens_seen": 3002859520
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.103309929789367e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.34898189749956,
      "tokens_seen": 3002925056
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.101303911735206e-05,
      "loss": 1.1875,
      "theoretical_loss": 3.3489763269294177,
      "tokens_seen": 3002990592
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.099297893681044e-05,
      "loss": 1.2169,
      "theoretical_loss": 3.3489707565148827,
      "tokens_seen": 3003056128
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.097291875626881e-05,
      "loss": 1.3784,
      "theoretical_loss": 3.348965186255947,
      "tokens_seen": 3003121664
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.095285857572719e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.348959616152603,
      "tokens_seen": 3003187200
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.093279839518557e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.348954046204843,
      "tokens_seen": 3003252736
    },
    {
      "epoch": 0.82,
      "objective/train/docs_used": 1091879,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2252734899520874,
      "objective/train/theoretical_loss": 3.3489484764126587,
      "objective/train/tokens_used": 1373777376,
      "theoretical_loss": 3.3489484764126587,
      "tokens_seen": 3003318272
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.091273821464394e-05,
      "loss": 1.2283,
      "theoretical_loss": 3.3489484764126587,
      "tokens_seen": 3003318272
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.089267803410232e-05,
      "loss": 1.3223,
      "theoretical_loss": 3.3489429067760437,
      "tokens_seen": 3003383808
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.087261785356068e-05,
      "loss": 1.3736,
      "theoretical_loss": 3.3489373372949887,
      "tokens_seen": 3003449344
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.085255767301905e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.3489317679694866,
      "tokens_seen": 3003514880
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.083249749247743e-05,
      "loss": 1.2242,
      "theoretical_loss": 3.3489261987995302,
      "tokens_seen": 3003580416
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.08124373119358e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.348920629785111,
      "tokens_seen": 3003645952
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.079237713139418e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.3489150609262217,
      "tokens_seen": 3003711488
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.077231695085256e-05,
      "loss": 1.2489,
      "theoretical_loss": 3.348909492222854,
      "tokens_seen": 3003777024
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.075225677031093e-05,
      "loss": 1.1544,
      "theoretical_loss": 3.348903923675001,
      "tokens_seen": 3003842560
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.073219658976931e-05,
      "loss": 1.2372,
      "theoretical_loss": 3.348898355282654,
      "tokens_seen": 3003908096
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.071213640922769e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.3488927870458065,
      "tokens_seen": 3003973632
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.069207622868606e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.3488872189644496,
      "tokens_seen": 3004039168
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.067201604814442e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.3488816510385764,
      "tokens_seen": 3004104704
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.065195586760281e-05,
      "loss": 1.2259,
      "theoretical_loss": 3.3488760832681783,
      "tokens_seen": 3004170240
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.063189568706119e-05,
      "loss": 1.1897,
      "theoretical_loss": 3.3488705156532483,
      "tokens_seen": 3004235776
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.061183550651956e-05,
      "loss": 1.26,
      "theoretical_loss": 3.3488649481937784,
      "tokens_seen": 3004301312
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.059177532597794e-05,
      "loss": 1.1755,
      "theoretical_loss": 3.348859380889761,
      "tokens_seen": 3004366848
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.057171514543632e-05,
      "loss": 1.2211,
      "theoretical_loss": 3.3488538137411883,
      "tokens_seen": 3004432384
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.055165496489469e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.3488482467480525,
      "tokens_seen": 3004497920
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.053159478435307e-05,
      "loss": 1.209,
      "theoretical_loss": 3.3488426799103457,
      "tokens_seen": 3004563456
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.051153460381144e-05,
      "loss": 1.261,
      "theoretical_loss": 3.3488371132280603,
      "tokens_seen": 3004628992
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.04914744232698e-05,
      "loss": 1.1495,
      "theoretical_loss": 3.348831546701189,
      "tokens_seen": 3004694528
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.047141424272818e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.3488259803297233,
      "tokens_seen": 3004760064
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.045135406218656e-05,
      "loss": 1.1674,
      "theoretical_loss": 3.348820414113656,
      "tokens_seen": 3004825600
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.043129388164493e-05,
      "loss": 1.2354,
      "theoretical_loss": 3.34881484805298,
      "tokens_seen": 3004891136
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.041123370110331e-05,
      "loss": 1.1841,
      "theoretical_loss": 3.348809282147686,
      "tokens_seen": 3004956672
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.039117352056168e-05,
      "loss": 1.1809,
      "theoretical_loss": 3.348803716397767,
      "tokens_seen": 3005022208
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.037111334002006e-05,
      "loss": 1.272,
      "theoretical_loss": 3.3487981508032156,
      "tokens_seen": 3005087744
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.035105315947844e-05,
      "loss": 1.2214,
      "theoretical_loss": 3.3487925853640244,
      "tokens_seen": 3005153280
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.033099297893681e-05,
      "loss": 1.2514,
      "theoretical_loss": 3.3487870200801844,
      "tokens_seen": 3005218816
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.031093279839519e-05,
      "loss": 1.2047,
      "theoretical_loss": 3.348781454951689,
      "tokens_seen": 3005284352
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.029087261785355e-05,
      "loss": 1.1444,
      "theoretical_loss": 3.34877588997853,
      "tokens_seen": 3005349888
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.027081243731194e-05,
      "loss": 1.3247,
      "theoretical_loss": 3.3487703251606993,
      "tokens_seen": 3005415424
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.025075225677032e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.3487647604981903,
      "tokens_seen": 3005480960
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.023069207622869e-05,
      "loss": 1.248,
      "theoretical_loss": 3.3487591959909944,
      "tokens_seen": 3005546496
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.021063189568707e-05,
      "loss": 1.1967,
      "theoretical_loss": 3.348753631639104,
      "tokens_seen": 3005612032
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.019057171514544e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.3487480674425116,
      "tokens_seen": 3005677568
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.017051153460382e-05,
      "loss": 1.2585,
      "theoretical_loss": 3.3487425034012093,
      "tokens_seen": 3005743104
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.01504513540622e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.3487369395151894,
      "tokens_seen": 3005808640
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.013039117352057e-05,
      "loss": 1.2026,
      "theoretical_loss": 3.3487313757844444,
      "tokens_seen": 3005874176
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.011033099297895e-05,
      "loss": 1.148,
      "theoretical_loss": 3.348725812208966,
      "tokens_seen": 3005939712
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.009027081243731e-05,
      "loss": 1.1544,
      "theoretical_loss": 3.348720248788747,
      "tokens_seen": 3006005248
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.007021063189568e-05,
      "loss": 1.1707,
      "theoretical_loss": 3.3487146855237797,
      "tokens_seen": 3006070784
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.005015045135406e-05,
      "loss": 1.1891,
      "theoretical_loss": 3.348709122414056,
      "tokens_seen": 3006136320
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.003009027081244e-05,
      "loss": 1.1441,
      "theoretical_loss": 3.348703559459569,
      "tokens_seen": 3006201856
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.001003009027081e-05,
      "loss": 1.203,
      "theoretical_loss": 3.34869799666031,
      "tokens_seen": 3006267392
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.998996990972919e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.3486924340162716,
      "tokens_seen": 3006332928
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.996990972918756e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.3486868715274465,
      "tokens_seen": 3006398464
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.994984954864594e-05,
      "loss": 1.1836,
      "theoretical_loss": 3.3486813091938266,
      "tokens_seen": 3006464000
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.992978936810431e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.3486757470154043,
      "tokens_seen": 3006529536
    },
    {
      "epoch": 0.82,
      "objective/train/docs_used": 1093144,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2492966651916504,
      "objective/train/theoretical_loss": 3.348670184992171,
      "objective/train/tokens_used": 1377054176,
      "theoretical_loss": 3.348670184992171,
      "tokens_seen": 3006595072
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.990972918756269e-05,
      "loss": 1.2636,
      "theoretical_loss": 3.348670184992171,
      "tokens_seen": 3006595072
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.988966900702107e-05,
      "loss": 1.1011,
      "theoretical_loss": 3.348664623124121,
      "tokens_seen": 3006660608
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.986960882647944e-05,
      "loss": 1.3603,
      "theoretical_loss": 3.3486590614112446,
      "tokens_seen": 3006726144
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.984954864593782e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.3486534998535356,
      "tokens_seen": 3006791680
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.982948846539619e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.348647938450985,
      "tokens_seen": 3006857216
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.980942828485457e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.348642377203586,
      "tokens_seen": 3006922752
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.978936810431294e-05,
      "loss": 1.1692,
      "theoretical_loss": 3.3486368161113305,
      "tokens_seen": 3006988288
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.976930792377132e-05,
      "loss": 1.0817,
      "theoretical_loss": 3.348631255174211,
      "tokens_seen": 3007053824
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.97492477432297e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.3486256943922195,
      "tokens_seen": 3007119360
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.972918756268807e-05,
      "loss": 1.2218,
      "theoretical_loss": 3.348620133765348,
      "tokens_seen": 3007184896
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.970912738214643e-05,
      "loss": 1.3164,
      "theoretical_loss": 3.34861457329359,
      "tokens_seen": 3007250432
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.968906720160481e-05,
      "loss": 1.2174,
      "theoretical_loss": 3.3486090129769366,
      "tokens_seen": 3007315968
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.966900702106319e-05,
      "loss": 1.1073,
      "theoretical_loss": 3.3486034528153805,
      "tokens_seen": 3007381504
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.964894684052156e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.348597892808914,
      "tokens_seen": 3007447040
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.962888665997994e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.34859233295753,
      "tokens_seen": 3007512576
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.960882647943831e-05,
      "loss": 1.1767,
      "theoretical_loss": 3.3485867732612196,
      "tokens_seen": 3007578112
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.958876629889669e-05,
      "loss": 1.4104,
      "theoretical_loss": 3.3485812137199757,
      "tokens_seen": 3007643648
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.956870611835506e-05,
      "loss": 1.1561,
      "theoretical_loss": 3.348575654333791,
      "tokens_seen": 3007709184
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.954864593781344e-05,
      "loss": 1.2311,
      "theoretical_loss": 3.348570095102657,
      "tokens_seen": 3007774720
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.952858575727182e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.348564536026567,
      "tokens_seen": 3007840256
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.950852557673019e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.3485589771055118,
      "tokens_seen": 3007905792
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.948846539618857e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.348553418339485,
      "tokens_seen": 3007971328
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.946840521564694e-05,
      "loss": 1.1555,
      "theoretical_loss": 3.3485478597284786,
      "tokens_seen": 3008036864
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.944834503510532e-05,
      "loss": 1.1866,
      "theoretical_loss": 3.348542301272485,
      "tokens_seen": 3008102400
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.94282848545637e-05,
      "loss": 1.2488,
      "theoretical_loss": 3.348536742971496,
      "tokens_seen": 3008167936
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.940822467402207e-05,
      "loss": 1.2058,
      "theoretical_loss": 3.348531184825504,
      "tokens_seen": 3008233472
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.938816449348045e-05,
      "loss": 1.2506,
      "theoretical_loss": 3.348525626834502,
      "tokens_seen": 3008299008
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.936810431293882e-05,
      "loss": 1.1255,
      "theoretical_loss": 3.3485200689984813,
      "tokens_seen": 3008364544
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.93480441323972e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.348514511317435,
      "tokens_seen": 3008430080
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.932798395185557e-05,
      "loss": 1.1621,
      "theoretical_loss": 3.348508953791355,
      "tokens_seen": 3008495616
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.930792377131394e-05,
      "loss": 1.2395,
      "theoretical_loss": 3.3485033964202335,
      "tokens_seen": 3008561152
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.928786359077231e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.348497839204063,
      "tokens_seen": 3008626688
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.926780341023069e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.3484922821428365,
      "tokens_seen": 3008692224
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.924774322968906e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.348486725236545,
      "tokens_seen": 3008757760
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.922768304914744e-05,
      "loss": 1.311,
      "theoretical_loss": 3.3484811684851814,
      "tokens_seen": 3008823296
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.920762286860582e-05,
      "loss": 1.1931,
      "theoretical_loss": 3.3484756118887384,
      "tokens_seen": 3008888832
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.918756268806419e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.3484700554472075,
      "tokens_seen": 3008954368
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.916750250752257e-05,
      "loss": 1.1465,
      "theoretical_loss": 3.348464499160582,
      "tokens_seen": 3009019904
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.914744232698096e-05,
      "loss": 1.2023,
      "theoretical_loss": 3.348458943028853,
      "tokens_seen": 3009085440
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.912738214643933e-05,
      "loss": 1.349,
      "theoretical_loss": 3.3484533870520137,
      "tokens_seen": 3009150976
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.91073219658977e-05,
      "loss": 1.2251,
      "theoretical_loss": 3.3484478312300565,
      "tokens_seen": 3009216512
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.908726178535607e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.348442275562973,
      "tokens_seen": 3009282048
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.906720160481445e-05,
      "loss": 1.1821,
      "theoretical_loss": 3.3484367200507563,
      "tokens_seen": 3009347584
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.904714142427282e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.3484311646933977,
      "tokens_seen": 3009413120
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.90270812437312e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.3484256094908904,
      "tokens_seen": 3009478656
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.900702106318957e-05,
      "loss": 1.222,
      "theoretical_loss": 3.3484200544432263,
      "tokens_seen": 3009544192
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.898696088264795e-05,
      "loss": 1.2082,
      "theoretical_loss": 3.3484144995503984,
      "tokens_seen": 3009609728
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.896690070210633e-05,
      "loss": 1.2133,
      "theoretical_loss": 3.3484089448123977,
      "tokens_seen": 3009675264
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.89468405215647e-05,
      "loss": 1.2566,
      "theoretical_loss": 3.3484033902292176,
      "tokens_seen": 3009740800
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.892678034102306e-05,
      "loss": 1.2148,
      "theoretical_loss": 3.34839783580085,
      "tokens_seen": 3009806336
    },
    {
      "epoch": 0.82,
      "objective/train/docs_used": 1094320,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2957937717437744,
      "objective/train/theoretical_loss": 3.3483922815272873,
      "objective/train/tokens_used": 1380330976,
      "theoretical_loss": 3.3483922815272873,
      "tokens_seen": 3009871872
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.890672016048144e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.3483922815272873,
      "tokens_seen": 3009871872
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.888665997993981e-05,
      "loss": 1.218,
      "theoretical_loss": 3.348386727408522,
      "tokens_seen": 3009937408
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.886659979939819e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.3483811734445457,
      "tokens_seen": 3010002944
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.884653961885657e-05,
      "loss": 1.1732,
      "theoretical_loss": 3.348375619635352,
      "tokens_seen": 3010068480
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.882647943831494e-05,
      "loss": 1.2053,
      "theoretical_loss": 3.348370065980932,
      "tokens_seen": 3010134016
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.880641925777332e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.3483645124812784,
      "tokens_seen": 3010199552
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.87863590772317e-05,
      "loss": 1.1889,
      "theoretical_loss": 3.348358959136384,
      "tokens_seen": 3010265088
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.876629889669008e-05,
      "loss": 1.3915,
      "theoretical_loss": 3.34835340594624,
      "tokens_seen": 3010330624
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.874623871614846e-05,
      "loss": 1.171,
      "theoretical_loss": 3.3483478529108397,
      "tokens_seen": 3010396160
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.872617853560682e-05,
      "loss": 1.2031,
      "theoretical_loss": 3.3483423000301755,
      "tokens_seen": 3010461696
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.87061183550652e-05,
      "loss": 1.2315,
      "theoretical_loss": 3.348336747304239,
      "tokens_seen": 3010527232
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.868605817452357e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.348331194733023,
      "tokens_seen": 3010592768
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.866599799398195e-05,
      "loss": 1.3551,
      "theoretical_loss": 3.3483256423165195,
      "tokens_seen": 3010658304
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.864593781344032e-05,
      "loss": 1.2163,
      "theoretical_loss": 3.348320090054721,
      "tokens_seen": 3010723840
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.86258776328987e-05,
      "loss": 1.2434,
      "theoretical_loss": 3.3483145379476205,
      "tokens_seen": 3010789376
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.860581745235708e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.3483089859952093,
      "tokens_seen": 3010854912
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.858575727181545e-05,
      "loss": 1.2318,
      "theoretical_loss": 3.3483034341974798,
      "tokens_seen": 3010920448
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.856569709127383e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.3482978825544247,
      "tokens_seen": 3010985984
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.85456369107322e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.3482923310660366,
      "tokens_seen": 3011051520
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.852557673019057e-05,
      "loss": 1.209,
      "theoretical_loss": 3.348286779732307,
      "tokens_seen": 3011117056
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.850551654964894e-05,
      "loss": 1.1812,
      "theoretical_loss": 3.3482812285532293,
      "tokens_seen": 3011182592
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.848545636910732e-05,
      "loss": 1.2183,
      "theoretical_loss": 3.348275677528795,
      "tokens_seen": 3011248128
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.846539618856569e-05,
      "loss": 1.1619,
      "theoretical_loss": 3.3482701266589965,
      "tokens_seen": 3011313664
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.844533600802407e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.348264575943826,
      "tokens_seen": 3011379200
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.842527582748244e-05,
      "loss": 1.2232,
      "theoretical_loss": 3.3482590253832765,
      "tokens_seen": 3011444736
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.840521564694082e-05,
      "loss": 1.242,
      "theoretical_loss": 3.34825347497734,
      "tokens_seen": 3011510272
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.838515546639921e-05,
      "loss": 1.3083,
      "theoretical_loss": 3.3482479247260084,
      "tokens_seen": 3011575808
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.836509528585759e-05,
      "loss": 1.2316,
      "theoretical_loss": 3.348242374629274,
      "tokens_seen": 3011641344
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.834503510531596e-05,
      "loss": 1.2961,
      "theoretical_loss": 3.34823682468713,
      "tokens_seen": 3011706880
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.832497492477432e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.3482312748995686,
      "tokens_seen": 3011772416
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.83049147442327e-05,
      "loss": 1.1723,
      "theoretical_loss": 3.3482257252665817,
      "tokens_seen": 3011837952
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.828485456369108e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3482201757881613,
      "tokens_seen": 3011903488
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.826479438314945e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.3482146264643005,
      "tokens_seen": 3011969024
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.824473420260783e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.3482090772949906,
      "tokens_seen": 3012034560
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.82246740220662e-05,
      "loss": 1.1937,
      "theoretical_loss": 3.348203528280225,
      "tokens_seen": 3012100096
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.820461384152458e-05,
      "loss": 1.2166,
      "theoretical_loss": 3.348197979419996,
      "tokens_seen": 3012165632
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.818455366098295e-05,
      "loss": 1.1765,
      "theoretical_loss": 3.3481924307142954,
      "tokens_seen": 3012231168
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.816449348044133e-05,
      "loss": 1.3187,
      "theoretical_loss": 3.3481868821631156,
      "tokens_seen": 3012296704
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.814443329989969e-05,
      "loss": 1.1384,
      "theoretical_loss": 3.348181333766449,
      "tokens_seen": 3012362240
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.812437311935807e-05,
      "loss": 1.1879,
      "theoretical_loss": 3.348175785524288,
      "tokens_seen": 3012427776
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.810431293881644e-05,
      "loss": 1.179,
      "theoretical_loss": 3.3481702374366247,
      "tokens_seen": 3012493312
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.808425275827482e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.3481646895034523,
      "tokens_seen": 3012558848
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.80641925777332e-05,
      "loss": 1.3446,
      "theoretical_loss": 3.3481591417247616,
      "tokens_seen": 3012624384
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.804413239719157e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.3481535941005465,
      "tokens_seen": 3012689920
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.802407221664996e-05,
      "loss": 1.2108,
      "theoretical_loss": 3.3481480466307985,
      "tokens_seen": 3012755456
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.800401203610834e-05,
      "loss": 1.1505,
      "theoretical_loss": 3.34814249931551,
      "tokens_seen": 3012820992
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.798395185556671e-05,
      "loss": 1.3459,
      "theoretical_loss": 3.3481369521546736,
      "tokens_seen": 3012886528
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.796389167502509e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.3481314051482816,
      "tokens_seen": 3012952064
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.794383149448345e-05,
      "loss": 1.3141,
      "theoretical_loss": 3.3481258582963256,
      "tokens_seen": 3013017600
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.792377131394183e-05,
      "loss": 1.188,
      "theoretical_loss": 3.348120311598799,
      "tokens_seen": 3013083136
    },
    {
      "epoch": 0.83,
      "objective/train/docs_used": 1095529,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2712202072143555,
      "objective/train/theoretical_loss": 3.3481147650556937,
      "objective/train/tokens_used": 1383607776,
      "theoretical_loss": 3.3481147650556937,
      "tokens_seen": 3013148672
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.79037111334002e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.3481147650556937,
      "tokens_seen": 3013148672
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.788365095285858e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.348109218667002,
      "tokens_seen": 3013214208
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.786359077231695e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.3481036724327167,
      "tokens_seen": 3013279744
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.784353059177533e-05,
      "loss": 1.2016,
      "theoretical_loss": 3.348098126352829,
      "tokens_seen": 3013345280
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.78234704112337e-05,
      "loss": 1.2157,
      "theoretical_loss": 3.3480925804273323,
      "tokens_seen": 3013410816
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.780341023069208e-05,
      "loss": 1.275,
      "theoretical_loss": 3.348087034656219,
      "tokens_seen": 3013476352
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.778335005015046e-05,
      "loss": 1.1858,
      "theoretical_loss": 3.3480814890394806,
      "tokens_seen": 3013541888
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.776328986960883e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.34807594357711,
      "tokens_seen": 3013607424
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.77432296890672e-05,
      "loss": 1.2293,
      "theoretical_loss": 3.3480703982690994,
      "tokens_seen": 3013672960
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.772316950852557e-05,
      "loss": 1.2157,
      "theoretical_loss": 3.3480648531154413,
      "tokens_seen": 3013738496
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.770310932798395e-05,
      "loss": 1.2538,
      "theoretical_loss": 3.348059308116128,
      "tokens_seen": 3013804032
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.768304914744232e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.348053763271152,
      "tokens_seen": 3013869568
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.76629889669007e-05,
      "loss": 1.2555,
      "theoretical_loss": 3.348048218580505,
      "tokens_seen": 3013935104
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.764292878635909e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.34804267404418,
      "tokens_seen": 3014000640
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.762286860581746e-05,
      "loss": 1.1971,
      "theoretical_loss": 3.348037129662169,
      "tokens_seen": 3014066176
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.760280842527584e-05,
      "loss": 1.3072,
      "theoretical_loss": 3.348031585434465,
      "tokens_seen": 3014131712
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.758274824473421e-05,
      "loss": 1.285,
      "theoretical_loss": 3.3480260413610594,
      "tokens_seen": 3014197248
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.756268806419259e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.348020497441945,
      "tokens_seen": 3014262784
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.754262788365095e-05,
      "loss": 1.1665,
      "theoretical_loss": 3.348014953677114,
      "tokens_seen": 3014328320
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.752256770310933e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.3480094100665596,
      "tokens_seen": 3014393856
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.75025075225677e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.3480038666102727,
      "tokens_seen": 3014459392
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.748244734202608e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.347998323308247,
      "tokens_seen": 3014524928
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.746238716148446e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.347992780160474,
      "tokens_seen": 3014590464
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.744232698094283e-05,
      "loss": 1.2501,
      "theoretical_loss": 3.3479872371669464,
      "tokens_seen": 3014656000
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.742226680040121e-05,
      "loss": 1.2013,
      "theoretical_loss": 3.3479816943276566,
      "tokens_seen": 3014721536
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.740220661985958e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.3479761516425963,
      "tokens_seen": 3014787072
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.738214643931796e-05,
      "loss": 1.2229,
      "theoretical_loss": 3.347970609111759,
      "tokens_seen": 3014852608
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.736208625877632e-05,
      "loss": 1.2146,
      "theoretical_loss": 3.347965066735136,
      "tokens_seen": 3014918144
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.73420260782347e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.3479595245127203,
      "tokens_seen": 3014983680
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.732196589769307e-05,
      "loss": 1.3636,
      "theoretical_loss": 3.347953982444504,
      "tokens_seen": 3015049216
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.730190571715145e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.3479484405304794,
      "tokens_seen": 3015114752
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.728184553660982e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.347942898770639,
      "tokens_seen": 3015180288
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.726178535606821e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.3479373571649753,
      "tokens_seen": 3015245824
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.724172517552659e-05,
      "loss": 1.229,
      "theoretical_loss": 3.3479318157134808,
      "tokens_seen": 3015311360
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.722166499498496e-05,
      "loss": 1.2065,
      "theoretical_loss": 3.347926274416147,
      "tokens_seen": 3015376896
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.720160481444334e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.347920733272967,
      "tokens_seen": 3015442432
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.718154463390172e-05,
      "loss": 1.2117,
      "theoretical_loss": 3.3479151922839328,
      "tokens_seen": 3015507968
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.716148445336008e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.3479096514490374,
      "tokens_seen": 3015573504
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.714142427281845e-05,
      "loss": 1.367,
      "theoretical_loss": 3.347904110768272,
      "tokens_seen": 3015639040
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.712136409227683e-05,
      "loss": 1.1617,
      "theoretical_loss": 3.34789857024163,
      "tokens_seen": 3015704576
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.71013039117352e-05,
      "loss": 1.156,
      "theoretical_loss": 3.3478930298691036,
      "tokens_seen": 3015770112
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.708124373119358e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.347887489650685,
      "tokens_seen": 3015835648
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.706118355065196e-05,
      "loss": 1.1781,
      "theoretical_loss": 3.3478819495863665,
      "tokens_seen": 3015901184
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.704112337011033e-05,
      "loss": 1.1598,
      "theoretical_loss": 3.3478764096761404,
      "tokens_seen": 3015966720
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.702106318956871e-05,
      "loss": 1.2099,
      "theoretical_loss": 3.3478708699199995,
      "tokens_seen": 3016032256
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.700100300902709e-05,
      "loss": 1.1565,
      "theoretical_loss": 3.3478653303179353,
      "tokens_seen": 3016097792
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.698094282848546e-05,
      "loss": 1.2102,
      "theoretical_loss": 3.3478597908699412,
      "tokens_seen": 3016163328
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.696088264794382e-05,
      "loss": 1.227,
      "theoretical_loss": 3.347854251576009,
      "tokens_seen": 3016228864
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.69408224674022e-05,
      "loss": 1.187,
      "theoretical_loss": 3.347848712436131,
      "tokens_seen": 3016294400
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.692076228686057e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.3478431734502996,
      "tokens_seen": 3016359936
    },
    {
      "epoch": 0.83,
      "objective/train/docs_used": 1096638,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3409316539764404,
      "objective/train/theoretical_loss": 3.3478376346185077,
      "objective/train/tokens_used": 1386884576,
      "theoretical_loss": 3.3478376346185077,
      "tokens_seen": 3016425472
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.690070210631895e-05,
      "loss": 1.3796,
      "theoretical_loss": 3.3478376346185077,
      "tokens_seen": 3016425472
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.688064192577734e-05,
      "loss": 1.1939,
      "theoretical_loss": 3.347832095940747,
      "tokens_seen": 3016491008
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.686058174523572e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.3478265574170103,
      "tokens_seen": 3016556544
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.684052156469409e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.3478210190472897,
      "tokens_seen": 3016622080
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.682046138415247e-05,
      "loss": 1.2729,
      "theoretical_loss": 3.3478154808315774,
      "tokens_seen": 3016687616
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.680040120361084e-05,
      "loss": 1.1604,
      "theoretical_loss": 3.3478099427698664,
      "tokens_seen": 3016753152
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.67803410230692e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.3478044048621487,
      "tokens_seen": 3016818688
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.676028084252758e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.3477988671084167,
      "tokens_seen": 3016884224
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.674022066198596e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.3477933295086624,
      "tokens_seen": 3016949760
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.672016048144433e-05,
      "loss": 1.1738,
      "theoretical_loss": 3.3477877920628787,
      "tokens_seen": 3017015296
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.670010030090271e-05,
      "loss": 1.3217,
      "theoretical_loss": 3.347782254771058,
      "tokens_seen": 3017080832
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.668004012036108e-05,
      "loss": 1.2361,
      "theoretical_loss": 3.3477767176331925,
      "tokens_seen": 3017146368
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.665997993981946e-05,
      "loss": 1.2087,
      "theoretical_loss": 3.3477711806492745,
      "tokens_seen": 3017211904
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.663991975927784e-05,
      "loss": 1.2087,
      "theoretical_loss": 3.3477656438192964,
      "tokens_seen": 3017277440
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.661985957873621e-05,
      "loss": 1.2083,
      "theoretical_loss": 3.347760107143251,
      "tokens_seen": 3017342976
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.659979939819459e-05,
      "loss": 1.1481,
      "theoretical_loss": 3.3477545706211296,
      "tokens_seen": 3017408512
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.657973921765295e-05,
      "loss": 1.1471,
      "theoretical_loss": 3.3477490342529257,
      "tokens_seen": 3017474048
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.655967903711133e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.347743498038631,
      "tokens_seen": 3017539584
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.65396188565697e-05,
      "loss": 1.2001,
      "theoretical_loss": 3.3477379619782384,
      "tokens_seen": 3017605120
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.651955867602809e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.34773242607174,
      "tokens_seen": 3017670656
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.649949849548647e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.347726890319128,
      "tokens_seen": 3017736192
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.647943831494484e-05,
      "loss": 1.359,
      "theoretical_loss": 3.3477213547203952,
      "tokens_seen": 3017801728
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.645937813440322e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.3477158192755336,
      "tokens_seen": 3017867264
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.64393179538616e-05,
      "loss": 1.1404,
      "theoretical_loss": 3.3477102839845356,
      "tokens_seen": 3017932800
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.641925777331997e-05,
      "loss": 1.1919,
      "theoretical_loss": 3.3477047488473937,
      "tokens_seen": 3017998336
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.639919759277835e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.3476992138641006,
      "tokens_seen": 3018063872
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.637913741223671e-05,
      "loss": 1.2141,
      "theoretical_loss": 3.347693679034648,
      "tokens_seen": 3018129408
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.635907723169508e-05,
      "loss": 1.225,
      "theoretical_loss": 3.347688144359029,
      "tokens_seen": 3018194944
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.633901705115346e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.3476826098372356,
      "tokens_seen": 3018260480
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.631895687061183e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.34767707546926,
      "tokens_seen": 3018326016
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.629889669007021e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.347671541255095,
      "tokens_seen": 3018391552
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.627883650952859e-05,
      "loss": 1.2537,
      "theoretical_loss": 3.3476660071947326,
      "tokens_seen": 3018457088
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.625877632898696e-05,
      "loss": 1.269,
      "theoretical_loss": 3.347660473288166,
      "tokens_seen": 3018522624
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.623871614844534e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.3476549395353863,
      "tokens_seen": 3018588160
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.621865596790371e-05,
      "loss": 1.2069,
      "theoretical_loss": 3.347649405936387,
      "tokens_seen": 3018653696
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.619859578736209e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.34764387249116,
      "tokens_seen": 3018719232
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.617853560682045e-05,
      "loss": 1.132,
      "theoretical_loss": 3.347638339199697,
      "tokens_seen": 3018784768
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.615847542627883e-05,
      "loss": 1.2233,
      "theoretical_loss": 3.347632806061992,
      "tokens_seen": 3018850304
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.613841524573722e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.3476272730780363,
      "tokens_seen": 3018915840
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.611835506519559e-05,
      "loss": 1.3935,
      "theoretical_loss": 3.3476217402478223,
      "tokens_seen": 3018981376
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.609829488465397e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.3476162075713427,
      "tokens_seen": 3019046912
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.607823470411234e-05,
      "loss": 1.1895,
      "theoretical_loss": 3.34761067504859,
      "tokens_seen": 3019112448
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.605817452357072e-05,
      "loss": 1.1821,
      "theoretical_loss": 3.3476051426795563,
      "tokens_seen": 3019177984
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.60381143430291e-05,
      "loss": 1.266,
      "theoretical_loss": 3.347599610464234,
      "tokens_seen": 3019243520
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.601805416248747e-05,
      "loss": 1.3396,
      "theoretical_loss": 3.3475940784026155,
      "tokens_seen": 3019309056
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.599799398194583e-05,
      "loss": 1.2026,
      "theoretical_loss": 3.347588546494693,
      "tokens_seen": 3019374592
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.597793380140421e-05,
      "loss": 1.2665,
      "theoretical_loss": 3.3475830147404597,
      "tokens_seen": 3019440128
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.595787362086259e-05,
      "loss": 1.2466,
      "theoretical_loss": 3.3475774831399074,
      "tokens_seen": 3019505664
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.593781344032096e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.3475719516930287,
      "tokens_seen": 3019571200
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.591775325977934e-05,
      "loss": 1.2207,
      "theoretical_loss": 3.347566420399815,
      "tokens_seen": 3019636736
    },
    {
      "epoch": 0.83,
      "objective/train/docs_used": 1097674,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2506718635559082,
      "objective/train/theoretical_loss": 3.34756088926026,
      "objective/train/tokens_used": 1390161376,
      "theoretical_loss": 3.34756088926026,
      "tokens_seen": 3019702272
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.589769307923771e-05,
      "loss": 1.245,
      "theoretical_loss": 3.34756088926026,
      "tokens_seen": 3019702272
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.587763289869609e-05,
      "loss": 1.3835,
      "theoretical_loss": 3.347555358274356,
      "tokens_seen": 3019767808
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.585757271815446e-05,
      "loss": 1.2042,
      "theoretical_loss": 3.3475498274420947,
      "tokens_seen": 3019833344
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.583751253761284e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.347544296763469,
      "tokens_seen": 3019898880
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.581745235707122e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.347538766238471,
      "tokens_seen": 3019964416
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.579739217652958e-05,
      "loss": 1.1757,
      "theoretical_loss": 3.3475332358670933,
      "tokens_seen": 3020029952
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.577733199598795e-05,
      "loss": 1.254,
      "theoretical_loss": 3.347527705649328,
      "tokens_seen": 3020095488
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.575727181544634e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.347522175585168,
      "tokens_seen": 3020161024
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.573721163490472e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.347516645674605,
      "tokens_seen": 3020226560
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.57171514543631e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.3475111159176323,
      "tokens_seen": 3020292096
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.569709127382147e-05,
      "loss": 1.2092,
      "theoretical_loss": 3.3475055863142416,
      "tokens_seen": 3020357632
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.567703109327985e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.3475000568644258,
      "tokens_seen": 3020423168
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.565697091273822e-05,
      "loss": 1.2049,
      "theoretical_loss": 3.3474945275681764,
      "tokens_seen": 3020488704
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.56369107321966e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.347488998425487,
      "tokens_seen": 3020554240
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.561685055165497e-05,
      "loss": 1.2141,
      "theoretical_loss": 3.3474834694363493,
      "tokens_seen": 3020619776
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.559679037111334e-05,
      "loss": 1.1889,
      "theoretical_loss": 3.347477940600756,
      "tokens_seen": 3020685312
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.557673019057171e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.3474724119186994,
      "tokens_seen": 3020750848
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.555667001003009e-05,
      "loss": 1.1912,
      "theoretical_loss": 3.3474668833901715,
      "tokens_seen": 3020816384
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.553660982948846e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.3474613550151653,
      "tokens_seen": 3020881920
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.551654964894684e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.3474558267936727,
      "tokens_seen": 3020947456
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.549648946840522e-05,
      "loss": 1.1994,
      "theoretical_loss": 3.3474502987256867,
      "tokens_seen": 3021012992
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.547642928786359e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.347444770811199,
      "tokens_seen": 3021078528
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.545636910732197e-05,
      "loss": 1.3217,
      "theoretical_loss": 3.3474392430502027,
      "tokens_seen": 3021144064
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.543630892678034e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.34743371544269,
      "tokens_seen": 3021209600
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.54162487462387e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.347428187988653,
      "tokens_seen": 3021275136
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.53961885656971e-05,
      "loss": 1.3743,
      "theoretical_loss": 3.3474226606880846,
      "tokens_seen": 3021340672
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.537612838515547e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.3474171335409766,
      "tokens_seen": 3021406208
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.535606820461385e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.347411606547322,
      "tokens_seen": 3021471744
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.533600802407222e-05,
      "loss": 1.1858,
      "theoretical_loss": 3.3474060797071123,
      "tokens_seen": 3021537280
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.53159478435306e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.347400553020341,
      "tokens_seen": 3021602816
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.529588766298897e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3473950264870003,
      "tokens_seen": 3021668352
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.527582748244735e-05,
      "loss": 1.244,
      "theoretical_loss": 3.347389500107082,
      "tokens_seen": 3021733888
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.525576730190572e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.3473839738805795,
      "tokens_seen": 3021799424
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.52357071213641e-05,
      "loss": 1.2018,
      "theoretical_loss": 3.3473784478074844,
      "tokens_seen": 3021864960
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.521564694082246e-05,
      "loss": 1.3602,
      "theoretical_loss": 3.347372921887789,
      "tokens_seen": 3021930496
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.519558676028084e-05,
      "loss": 1.2057,
      "theoretical_loss": 3.347367396121486,
      "tokens_seen": 3021996032
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.517552657973921e-05,
      "loss": 1.1948,
      "theoretical_loss": 3.347361870508568,
      "tokens_seen": 3022061568
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.515546639919759e-05,
      "loss": 1.1635,
      "theoretical_loss": 3.3473563450490276,
      "tokens_seen": 3022127104
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.513540621865597e-05,
      "loss": 1.3319,
      "theoretical_loss": 3.3473508197428563,
      "tokens_seen": 3022192640
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.511534603811434e-05,
      "loss": 1.244,
      "theoretical_loss": 3.3473452945900477,
      "tokens_seen": 3022258176
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.509528585757272e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.3473397695905933,
      "tokens_seen": 3022323712
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.50752256770311e-05,
      "loss": 1.3208,
      "theoretical_loss": 3.347334244744486,
      "tokens_seen": 3022389248
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.505516549648947e-05,
      "loss": 1.1542,
      "theoretical_loss": 3.347328720051718,
      "tokens_seen": 3022454784
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.503510531594785e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.3473231955122817,
      "tokens_seen": 3022520320
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.501504513540622e-05,
      "loss": 1.1626,
      "theoretical_loss": 3.3473176711261696,
      "tokens_seen": 3022585856
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.49949849548646e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.3473121468933744,
      "tokens_seen": 3022651392
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.497492477432297e-05,
      "loss": 1.1774,
      "theoretical_loss": 3.347306622813888,
      "tokens_seen": 3022716928
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.495486459378135e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.347301098887703,
      "tokens_seen": 3022782464
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.493480441323972e-05,
      "loss": 1.1391,
      "theoretical_loss": 3.347295575114812,
      "tokens_seen": 3022848000
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.49147442326981e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.3472900514952073,
      "tokens_seen": 3022913536
    },
    {
      "epoch": 0.83,
      "objective/train/docs_used": 1098835,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1438367366790771,
      "objective/train/theoretical_loss": 3.347284528028881,
      "objective/train/tokens_used": 1393438176,
      "theoretical_loss": 3.347284528028881,
      "tokens_seen": 3022979072
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.489468405215648e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.347284528028881,
      "tokens_seen": 3022979072
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.487462387161485e-05,
      "loss": 1.1539,
      "theoretical_loss": 3.3472790047158267,
      "tokens_seen": 3023044608
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.485456369107323e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.3472734815560354,
      "tokens_seen": 3023110144
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.48345035105316e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.3472679585495,
      "tokens_seen": 3023175680
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.481444332998997e-05,
      "loss": 1.2409,
      "theoretical_loss": 3.3472624356962135,
      "tokens_seen": 3023241216
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.479438314944834e-05,
      "loss": 1.207,
      "theoretical_loss": 3.3472569129961673,
      "tokens_seen": 3023306752
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.477432296890672e-05,
      "loss": 1.3755,
      "theoretical_loss": 3.347251390449355,
      "tokens_seen": 3023372288
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.475426278836509e-05,
      "loss": 1.1983,
      "theoretical_loss": 3.347245868055768,
      "tokens_seen": 3023437824
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.473420260782347e-05,
      "loss": 1.1823,
      "theoretical_loss": 3.347240345815399,
      "tokens_seen": 3023503360
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.471414242728184e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.347234823728241,
      "tokens_seen": 3023568896
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.469408224674022e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.347229301794286,
      "tokens_seen": 3023634432
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.46740220661986e-05,
      "loss": 1.1643,
      "theoretical_loss": 3.3472237800135263,
      "tokens_seen": 3023699968
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.465396188565697e-05,
      "loss": 1.3158,
      "theoretical_loss": 3.347218258385954,
      "tokens_seen": 3023765504
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.463390170511536e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.3472127369115627,
      "tokens_seen": 3023831040
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.461384152457372e-05,
      "loss": 1.1237,
      "theoretical_loss": 3.347207215590344,
      "tokens_seen": 3023896576
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.45937813440321e-05,
      "loss": 1.1584,
      "theoretical_loss": 3.34720169442229,
      "tokens_seen": 3023962112
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.457372116349047e-05,
      "loss": 1.2243,
      "theoretical_loss": 3.347196173407394,
      "tokens_seen": 3024027648
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.455366098294885e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.347190652545648,
      "tokens_seen": 3024093184
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.453360080240723e-05,
      "loss": 1.2058,
      "theoretical_loss": 3.347185131837044,
      "tokens_seen": 3024158720
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.45135406218656e-05,
      "loss": 1.1982,
      "theoretical_loss": 3.3471796112815753,
      "tokens_seen": 3024224256
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.449348044132398e-05,
      "loss": 1.2272,
      "theoretical_loss": 3.347174090879234,
      "tokens_seen": 3024289792
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.447342026078235e-05,
      "loss": 1.1686,
      "theoretical_loss": 3.347168570630012,
      "tokens_seen": 3024355328
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.445336008024073e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.3471630505339025,
      "tokens_seen": 3024420864
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.443329989969909e-05,
      "loss": 1.3129,
      "theoretical_loss": 3.3471575305908976,
      "tokens_seen": 3024486400
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.441323971915747e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.3471520108009893,
      "tokens_seen": 3024551936
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.439317953861584e-05,
      "loss": 1.366,
      "theoretical_loss": 3.347146491164171,
      "tokens_seen": 3024617472
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.437311935807422e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.3471409716804343,
      "tokens_seen": 3024683008
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.43530591775326e-05,
      "loss": 1.1619,
      "theoretical_loss": 3.3471354523497725,
      "tokens_seen": 3024748544
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.433299899699097e-05,
      "loss": 1.1911,
      "theoretical_loss": 3.347129933172177,
      "tokens_seen": 3024814080
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.431293881644935e-05,
      "loss": 1.3469,
      "theoretical_loss": 3.347124414147641,
      "tokens_seen": 3024879616
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.429287863590772e-05,
      "loss": 1.2297,
      "theoretical_loss": 3.347118895276157,
      "tokens_seen": 3024945152
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.42728184553661e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.3471133765577163,
      "tokens_seen": 3025010688
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.425275827482449e-05,
      "loss": 1.261,
      "theoretical_loss": 3.3471078579923126,
      "tokens_seen": 3025076224
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.423269809428285e-05,
      "loss": 1.2284,
      "theoretical_loss": 3.347102339579938,
      "tokens_seen": 3025141760
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.421263791374123e-05,
      "loss": 1.2061,
      "theoretical_loss": 3.3470968213205845,
      "tokens_seen": 3025207296
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.41925777331996e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.3470913032142455,
      "tokens_seen": 3025272832
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.417251755265798e-05,
      "loss": 1.2281,
      "theoretical_loss": 3.347085785260912,
      "tokens_seen": 3025338368
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.415245737211635e-05,
      "loss": 1.2292,
      "theoretical_loss": 3.347080267460578,
      "tokens_seen": 3025403904
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.413239719157473e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.347074749813235,
      "tokens_seen": 3025469440
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.41123370110331e-05,
      "loss": 1.4134,
      "theoretical_loss": 3.3470692323188755,
      "tokens_seen": 3025534976
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.409227683049148e-05,
      "loss": 1.313,
      "theoretical_loss": 3.347063714977492,
      "tokens_seen": 3025600512
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.407221664994986e-05,
      "loss": 1.1043,
      "theoretical_loss": 3.3470581977890776,
      "tokens_seen": 3025666048
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.405215646940823e-05,
      "loss": 1.2497,
      "theoretical_loss": 3.347052680753624,
      "tokens_seen": 3025731584
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.40320962888666e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.3470471638711237,
      "tokens_seen": 3025797120
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.401203610832497e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.347041647141569,
      "tokens_seen": 3025862656
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.399197592778335e-05,
      "loss": 1.1873,
      "theoretical_loss": 3.3470361305649527,
      "tokens_seen": 3025928192
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.397191574724172e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.347030614141268,
      "tokens_seen": 3025993728
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.39518555667001e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.347025097870506,
      "tokens_seen": 3026059264
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.393179538615847e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.3470195817526593,
      "tokens_seen": 3026124800
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.391173520561685e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.347014065787721,
      "tokens_seen": 3026190336
    },
    {
      "epoch": 0.83,
      "objective/train/docs_used": 1099986,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3417669534683228,
      "objective/train/theoretical_loss": 3.3470085499756834,
      "objective/train/tokens_used": 1396714976,
      "theoretical_loss": 3.3470085499756834,
      "tokens_seen": 3026255872
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.389167502507524e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.3470085499756834,
      "tokens_seen": 3026255872
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.387161484453361e-05,
      "loss": 1.4052,
      "theoretical_loss": 3.347003034316539,
      "tokens_seen": 3026321408
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.385155466399198e-05,
      "loss": 1.2444,
      "theoretical_loss": 3.3469975188102796,
      "tokens_seen": 3026386944
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.383149448345035e-05,
      "loss": 1.3747,
      "theoretical_loss": 3.3469920034568985,
      "tokens_seen": 3026452480
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.381143430290873e-05,
      "loss": 1.2653,
      "theoretical_loss": 3.3469864882563876,
      "tokens_seen": 3026518016
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.37913741223671e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.3469809732087397,
      "tokens_seen": 3026583552
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.377131394182548e-05,
      "loss": 1.302,
      "theoretical_loss": 3.346975458313947,
      "tokens_seen": 3026649088
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.375125376128386e-05,
      "loss": 1.214,
      "theoretical_loss": 3.346969943572002,
      "tokens_seen": 3026714624
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.373119358074223e-05,
      "loss": 1.238,
      "theoretical_loss": 3.346964428982897,
      "tokens_seen": 3026780160
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.37111334002006e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.346958914546625,
      "tokens_seen": 3026845696
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.369107321965898e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.3469534002631782,
      "tokens_seen": 3026911232
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.367101303911736e-05,
      "loss": 1.2009,
      "theoretical_loss": 3.346947886132549,
      "tokens_seen": 3026976768
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.365095285857572e-05,
      "loss": 1.2155,
      "theoretical_loss": 3.3469423721547296,
      "tokens_seen": 3027042304
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.36308926780341e-05,
      "loss": 1.2345,
      "theoretical_loss": 3.3469368583297125,
      "tokens_seen": 3027107840
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.361083249749247e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.3469313446574906,
      "tokens_seen": 3027173376
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.359077231695085e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.346925831138056,
      "tokens_seen": 3027238912
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.357071213640922e-05,
      "loss": 1.3279,
      "theoretical_loss": 3.346920317771401,
      "tokens_seen": 3027304448
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.35506519558676e-05,
      "loss": 1.2119,
      "theoretical_loss": 3.3469148045575188,
      "tokens_seen": 3027369984
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.353059177532598e-05,
      "loss": 1.3831,
      "theoretical_loss": 3.3469092914964014,
      "tokens_seen": 3027435520
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.351053159478436e-05,
      "loss": 1.2556,
      "theoretical_loss": 3.346903778588041,
      "tokens_seen": 3027501056
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.349047141424274e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.34689826583243,
      "tokens_seen": 3027566592
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.347041123370112e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.346892753229562,
      "tokens_seen": 3027632128
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.345035105315948e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.3468872407794277,
      "tokens_seen": 3027697664
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.343029087261785e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.346881728482021,
      "tokens_seen": 3027763200
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.341023069207623e-05,
      "loss": 1.228,
      "theoretical_loss": 3.346876216337334,
      "tokens_seen": 3027828736
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.33901705115346e-05,
      "loss": 1.3207,
      "theoretical_loss": 3.346870704345358,
      "tokens_seen": 3027894272
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.337011033099298e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.3468651925060877,
      "tokens_seen": 3027959808
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.335005015045136e-05,
      "loss": 1.1998,
      "theoretical_loss": 3.346859680819514,
      "tokens_seen": 3028025344
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.332998996990973e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.3468541692856295,
      "tokens_seen": 3028090880
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.330992978936811e-05,
      "loss": 1.1669,
      "theoretical_loss": 3.3468486579044265,
      "tokens_seen": 3028156416
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.328986960882648e-05,
      "loss": 1.2171,
      "theoretical_loss": 3.3468431466758983,
      "tokens_seen": 3028221952
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.326980942828486e-05,
      "loss": 1.2282,
      "theoretical_loss": 3.346837635600037,
      "tokens_seen": 3028287488
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.324974924774322e-05,
      "loss": 1.138,
      "theoretical_loss": 3.3468321246768347,
      "tokens_seen": 3028353024
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.32296890672016e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.346826613906284,
      "tokens_seen": 3028418560
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.320962888665997e-05,
      "loss": 1.2174,
      "theoretical_loss": 3.346821103288378,
      "tokens_seen": 3028484096
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.318956870611835e-05,
      "loss": 1.1805,
      "theoretical_loss": 3.3468155928231083,
      "tokens_seen": 3028549632
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.316950852557673e-05,
      "loss": 1.1702,
      "theoretical_loss": 3.346810082510468,
      "tokens_seen": 3028615168
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.31494483450351e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.346804572350449,
      "tokens_seen": 3028680704
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.312938816449349e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.3467990623430444,
      "tokens_seen": 3028746240
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.310932798395187e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.346793552488246,
      "tokens_seen": 3028811776
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.308926780341024e-05,
      "loss": 1.3538,
      "theoretical_loss": 3.346788042786047,
      "tokens_seen": 3028877312
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.30692076228686e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.346782533236439,
      "tokens_seen": 3028942848
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.304914744232698e-05,
      "loss": 1.3039,
      "theoretical_loss": 3.3467770238394157,
      "tokens_seen": 3029008384
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.302908726178536e-05,
      "loss": 1.312,
      "theoretical_loss": 3.3467715145949684,
      "tokens_seen": 3029073920
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.300902708124373e-05,
      "loss": 1.2209,
      "theoretical_loss": 3.34676600550309,
      "tokens_seen": 3029139456
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.298896690070211e-05,
      "loss": 1.259,
      "theoretical_loss": 3.346760496563773,
      "tokens_seen": 3029204992
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.296890672016048e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.3467549877770097,
      "tokens_seen": 3029270528
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.294884653961886e-05,
      "loss": 1.2281,
      "theoretical_loss": 3.346749479142793,
      "tokens_seen": 3029336064
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.292878635907724e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.346743970661115,
      "tokens_seen": 3029401600
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.290872617853561e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.3467384623319685,
      "tokens_seen": 3029467136
    },
    {
      "epoch": 0.84,
      "objective/train/docs_used": 1101281,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2901899814605713,
      "objective/train/theoretical_loss": 3.3467329541553457,
      "objective/train/tokens_used": 1399991776,
      "theoretical_loss": 3.3467329541553457,
      "tokens_seen": 3029532672
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.288866599799399e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.3467329541553457,
      "tokens_seen": 3029532672
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.286860581745235e-05,
      "loss": 1.2278,
      "theoretical_loss": 3.346727446131239,
      "tokens_seen": 3029598208
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.284854563691073e-05,
      "loss": 1.2291,
      "theoretical_loss": 3.346721938259641,
      "tokens_seen": 3029663744
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.28284854563691e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.3467164305405444,
      "tokens_seen": 3029729280
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.280842527582748e-05,
      "loss": 1.1136,
      "theoretical_loss": 3.346710922973941,
      "tokens_seen": 3029794816
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.278836509528585e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.3467054155598244,
      "tokens_seen": 3029860352
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.276830491474424e-05,
      "loss": 1.2186,
      "theoretical_loss": 3.346699908298186,
      "tokens_seen": 3029925888
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.274824473420262e-05,
      "loss": 1.272,
      "theoretical_loss": 3.346694401189019,
      "tokens_seen": 3029991424
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.2728184553661e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.346688894232315,
      "tokens_seen": 3030056960
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.270812437311937e-05,
      "loss": 1.1944,
      "theoretical_loss": 3.346683387428068,
      "tokens_seen": 3030122496
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.268806419257775e-05,
      "loss": 1.245,
      "theoretical_loss": 3.3466778807762694,
      "tokens_seen": 3030188032
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.266800401203611e-05,
      "loss": 1.277,
      "theoretical_loss": 3.346672374276911,
      "tokens_seen": 3030253568
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.264794383149448e-05,
      "loss": 1.2613,
      "theoretical_loss": 3.346666867929987,
      "tokens_seen": 3030319104
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.262788365095286e-05,
      "loss": 1.3421,
      "theoretical_loss": 3.346661361735489,
      "tokens_seen": 3030384640
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.260782347041123e-05,
      "loss": 1.2026,
      "theoretical_loss": 3.3466558556934087,
      "tokens_seen": 3030450176
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.258776328986961e-05,
      "loss": 1.1613,
      "theoretical_loss": 3.34665034980374,
      "tokens_seen": 3030515712
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.256770310932799e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.346644844066475,
      "tokens_seen": 3030581248
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.254764292878636e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.346639338481605,
      "tokens_seen": 3030646784
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.252758274824474e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.3466338330491245,
      "tokens_seen": 3030712320
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.250752256770311e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.3466283277690243,
      "tokens_seen": 3030777856
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.248746238716148e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.3466228226412977,
      "tokens_seen": 3030843392
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.246740220661985e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.3466173176659373,
      "tokens_seen": 3030908928
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.244734202607823e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.346611812842935,
      "tokens_seen": 3030974464
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.24272818455366e-05,
      "loss": 1.2147,
      "theoretical_loss": 3.3466063081722837,
      "tokens_seen": 3031040000
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.240722166499498e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.3466008036539754,
      "tokens_seen": 3031105536
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.238716148445337e-05,
      "loss": 1.2167,
      "theoretical_loss": 3.3465952992880035,
      "tokens_seen": 3031171072
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.236710130391174e-05,
      "loss": 1.2379,
      "theoretical_loss": 3.3465897950743595,
      "tokens_seen": 3031236608
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.234704112337012e-05,
      "loss": 1.287,
      "theoretical_loss": 3.3465842910130363,
      "tokens_seen": 3031302144
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.23269809428285e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.346578787104027,
      "tokens_seen": 3031367680
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.230692076228687e-05,
      "loss": 1.4265,
      "theoretical_loss": 3.3465732833473227,
      "tokens_seen": 3031433216
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.228686058174523e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.3465677797429176,
      "tokens_seen": 3031498752
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.226680040120361e-05,
      "loss": 1.2391,
      "theoretical_loss": 3.346562276290803,
      "tokens_seen": 3031564288
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.224674022066199e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.3465567729909713,
      "tokens_seen": 3031629824
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.222668004012036e-05,
      "loss": 1.2308,
      "theoretical_loss": 3.346551269843416,
      "tokens_seen": 3031695360
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.220661985957874e-05,
      "loss": 1.3756,
      "theoretical_loss": 3.346545766848129,
      "tokens_seen": 3031760896
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.218655967903711e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.3465402640051023,
      "tokens_seen": 3031826432
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.216649949849549e-05,
      "loss": 1.1679,
      "theoretical_loss": 3.346534761314329,
      "tokens_seen": 3031891968
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.214643931795386e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.3465292587758015,
      "tokens_seen": 3031957504
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.212637913741224e-05,
      "loss": 1.3352,
      "theoretical_loss": 3.3465237563895123,
      "tokens_seen": 3032023040
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.210631895687062e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.346518254155454,
      "tokens_seen": 3032088576
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.208625877632898e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.346512752073619,
      "tokens_seen": 3032154112
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.206619859578735e-05,
      "loss": 1.1764,
      "theoretical_loss": 3.3465072501439996,
      "tokens_seen": 3032219648
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.204613841524573e-05,
      "loss": 1.2189,
      "theoretical_loss": 3.3465017483665886,
      "tokens_seen": 3032285184
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.20260782347041e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.3464962467413786,
      "tokens_seen": 3032350720
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.20060180541625e-05,
      "loss": 1.3172,
      "theoretical_loss": 3.3464907452683614,
      "tokens_seen": 3032416256
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.198595787362087e-05,
      "loss": 1.2168,
      "theoretical_loss": 3.34648524394753,
      "tokens_seen": 3032481792
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.196589769307925e-05,
      "loss": 1.2237,
      "theoretical_loss": 3.3464797427788775,
      "tokens_seen": 3032547328
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.194583751253762e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.346474241762395,
      "tokens_seen": 3032612864
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.1925777331996e-05,
      "loss": 1.345,
      "theoretical_loss": 3.3464687408980764,
      "tokens_seen": 3032678400
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.190571715145437e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.3464632401859133,
      "tokens_seen": 3032743936
    },
    {
      "epoch": 0.84,
      "objective/train/docs_used": 1102459,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2690143585205078,
      "objective/train/theoretical_loss": 3.3464577396258983,
      "objective/train/tokens_used": 1403268576,
      "theoretical_loss": 3.3464577396258983,
      "tokens_seen": 3032809472
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.188565697091274e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.3464577396258983,
      "tokens_seen": 3032809472
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.186559679037111e-05,
      "loss": 1.1818,
      "theoretical_loss": 3.3464522392180247,
      "tokens_seen": 3032875008
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.184553660982949e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.346446738962284,
      "tokens_seen": 3032940544
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.182547642928786e-05,
      "loss": 1.3231,
      "theoretical_loss": 3.3464412388586693,
      "tokens_seen": 3033006080
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.180541624874624e-05,
      "loss": 1.1731,
      "theoretical_loss": 3.3464357389071724,
      "tokens_seen": 3033071616
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.178535606820462e-05,
      "loss": 1.2901,
      "theoretical_loss": 3.3464302391077867,
      "tokens_seen": 3033137152
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.176529588766299e-05,
      "loss": 1.1553,
      "theoretical_loss": 3.3464247394605042,
      "tokens_seen": 3033202688
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.174523570712137e-05,
      "loss": 1.2106,
      "theoretical_loss": 3.3464192399653174,
      "tokens_seen": 3033268224
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.172517552657974e-05,
      "loss": 1.217,
      "theoretical_loss": 3.346413740622219,
      "tokens_seen": 3033333760
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.17051153460381e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.346408241431202,
      "tokens_seen": 3033399296
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.168505516549648e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.3464027423922578,
      "tokens_seen": 3033464832
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.166499498495486e-05,
      "loss": 1.2143,
      "theoretical_loss": 3.3463972435053795,
      "tokens_seen": 3033530368
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.164493480441323e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.3463917447705596,
      "tokens_seen": 3033595904
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.162487462387162e-05,
      "loss": 1.3331,
      "theoretical_loss": 3.346386246187791,
      "tokens_seen": 3033661440
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.160481444333e-05,
      "loss": 1.2306,
      "theoretical_loss": 3.346380747757065,
      "tokens_seen": 3033726976
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.158475426278837e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.346375249478376,
      "tokens_seen": 3033792512
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.156469408224675e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.3463697513517143,
      "tokens_seen": 3033858048
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.154463390170512e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.346364253377074,
      "tokens_seen": 3033923584
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.15245737211635e-05,
      "loss": 1.2392,
      "theoretical_loss": 3.346358755554447,
      "tokens_seen": 3033989120
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.150451354062186e-05,
      "loss": 1.2126,
      "theoretical_loss": 3.346353257883826,
      "tokens_seen": 3034054656
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.148445336008024e-05,
      "loss": 1.4059,
      "theoretical_loss": 3.3463477603652034,
      "tokens_seen": 3034120192
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.146439317953861e-05,
      "loss": 1.3293,
      "theoretical_loss": 3.346342262998572,
      "tokens_seen": 3034185728
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.144433299899699e-05,
      "loss": 1.2444,
      "theoretical_loss": 3.3463367657839242,
      "tokens_seen": 3034251264
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.142427281845537e-05,
      "loss": 1.2232,
      "theoretical_loss": 3.3463312687212525,
      "tokens_seen": 3034316800
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.140421263791374e-05,
      "loss": 1.2317,
      "theoretical_loss": 3.3463257718105486,
      "tokens_seen": 3034382336
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.138415245737212e-05,
      "loss": 1.3124,
      "theoretical_loss": 3.3463202750518066,
      "tokens_seen": 3034447872
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.13640922768305e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.3463147784450173,
      "tokens_seen": 3034513408
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.134403209628887e-05,
      "loss": 1.2331,
      "theoretical_loss": 3.346309281990175,
      "tokens_seen": 3034578944
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.132397191574724e-05,
      "loss": 1.1598,
      "theoretical_loss": 3.3463037856872706,
      "tokens_seen": 3034644480
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.130391173520561e-05,
      "loss": 1.3199,
      "theoretical_loss": 3.3462982895362976,
      "tokens_seen": 3034710016
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.128385155466398e-05,
      "loss": 1.331,
      "theoretical_loss": 3.3462927935372484,
      "tokens_seen": 3034775552
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.126379137412237e-05,
      "loss": 1.1731,
      "theoretical_loss": 3.346287297690115,
      "tokens_seen": 3034841088
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.124373119358075e-05,
      "loss": 1.2192,
      "theoretical_loss": 3.3462818019948903,
      "tokens_seen": 3034906624
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.122367101303912e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.3462763064515673,
      "tokens_seen": 3034972160
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.12036108324975e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.3462708110601374,
      "tokens_seen": 3035037696
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.118355065195588e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3462653158205944,
      "tokens_seen": 3035103232
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.116349047141425e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.3462598207329295,
      "tokens_seen": 3035168768
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.114343029087263e-05,
      "loss": 1.368,
      "theoretical_loss": 3.346254325797136,
      "tokens_seen": 3035234304
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.1123370110331e-05,
      "loss": 1.2334,
      "theoretical_loss": 3.346248831013207,
      "tokens_seen": 3035299840
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.110330992978937e-05,
      "loss": 1.2119,
      "theoretical_loss": 3.3462433363811335,
      "tokens_seen": 3035365376
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.108324974924774e-05,
      "loss": 1.3174,
      "theoretical_loss": 3.3462378419009093,
      "tokens_seen": 3035430912
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.106318956870612e-05,
      "loss": 1.3344,
      "theoretical_loss": 3.346232347572526,
      "tokens_seen": 3035496448
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.104312938816449e-05,
      "loss": 1.2341,
      "theoretical_loss": 3.346226853395977,
      "tokens_seen": 3035561984
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.102306920762287e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.3462213593712544,
      "tokens_seen": 3035627520
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.100300902708124e-05,
      "loss": 1.2427,
      "theoretical_loss": 3.346215865498351,
      "tokens_seen": 3035693056
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.098294884653962e-05,
      "loss": 1.3289,
      "theoretical_loss": 3.346210371777259,
      "tokens_seen": 3035758592
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.0962888665998e-05,
      "loss": 1.2463,
      "theoretical_loss": 3.346204878207971,
      "tokens_seen": 3035824128
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.094282848545637e-05,
      "loss": 1.3251,
      "theoretical_loss": 3.3461993847904794,
      "tokens_seen": 3035889664
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.092276830491473e-05,
      "loss": 1.3799,
      "theoretical_loss": 3.346193891524777,
      "tokens_seen": 3035955200
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.090270812437311e-05,
      "loss": 1.272,
      "theoretical_loss": 3.346188398410856,
      "tokens_seen": 3036020736
    },
    {
      "epoch": 0.84,
      "objective/train/docs_used": 1103322,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2968523502349854,
      "objective/train/theoretical_loss": 3.346182905448709,
      "objective/train/tokens_used": 1406545376,
      "theoretical_loss": 3.346182905448709,
      "tokens_seen": 3036086272
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.08826479438315e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.346182905448709,
      "tokens_seen": 3036086272
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.086258776328987e-05,
      "loss": 1.352,
      "theoretical_loss": 3.3461774126383292,
      "tokens_seen": 3036151808
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.084252758274825e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.3461719199797084,
      "tokens_seen": 3036217344
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.082246740220663e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.346166427472839,
      "tokens_seen": 3036282880
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.0802407221665e-05,
      "loss": 1.1911,
      "theoretical_loss": 3.346160935117714,
      "tokens_seen": 3036348416
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.078234704112338e-05,
      "loss": 1.247,
      "theoretical_loss": 3.346155442914326,
      "tokens_seen": 3036413952
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.076228686058175e-05,
      "loss": 1.3837,
      "theoretical_loss": 3.3461499508626673,
      "tokens_seen": 3036479488
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.074222668004013e-05,
      "loss": 1.3276,
      "theoretical_loss": 3.3461444589627307,
      "tokens_seen": 3036545024
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.072216649949849e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.3461389672145083,
      "tokens_seen": 3036610560
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.070210631895687e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.3461334756179926,
      "tokens_seen": 3036676096
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.068204613841524e-05,
      "loss": 1.3708,
      "theoretical_loss": 3.3461279841731764,
      "tokens_seen": 3036741632
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.066198595787362e-05,
      "loss": 1.2164,
      "theoretical_loss": 3.346122492880052,
      "tokens_seen": 3036807168
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.0641925777332e-05,
      "loss": 1.3123,
      "theoretical_loss": 3.346117001738613,
      "tokens_seen": 3036872704
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.062186559679037e-05,
      "loss": 1.204,
      "theoretical_loss": 3.3461115107488504,
      "tokens_seen": 3036938240
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.060180541624875e-05,
      "loss": 1.1967,
      "theoretical_loss": 3.3461060199107573,
      "tokens_seen": 3037003776
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.058174523570712e-05,
      "loss": 1.2383,
      "theoretical_loss": 3.3461005292243264,
      "tokens_seen": 3037069312
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.05616850551655e-05,
      "loss": 1.3572,
      "theoretical_loss": 3.3460950386895507,
      "tokens_seen": 3037134848
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.054162487462387e-05,
      "loss": 1.3348,
      "theoretical_loss": 3.3460895483064217,
      "tokens_seen": 3037200384
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.052156469408224e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.3460840580749327,
      "tokens_seen": 3037265920
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.050150451354063e-05,
      "loss": 1.2046,
      "theoretical_loss": 3.346078567995076,
      "tokens_seen": 3037331456
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.0481444332999e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.346073078066844,
      "tokens_seen": 3037396992
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.046138415245738e-05,
      "loss": 1.2237,
      "theoretical_loss": 3.3460675882902295,
      "tokens_seen": 3037462528
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.044132397191575e-05,
      "loss": 1.2374,
      "theoretical_loss": 3.346062098665225,
      "tokens_seen": 3037528064
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.042126379137413e-05,
      "loss": 1.264,
      "theoretical_loss": 3.346056609191823,
      "tokens_seen": 3037593600
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.04012036108325e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.3460511198700162,
      "tokens_seen": 3037659136
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.038114343029088e-05,
      "loss": 1.3568,
      "theoretical_loss": 3.3460456306997965,
      "tokens_seen": 3037724672
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.036108324974926e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.346040141681157,
      "tokens_seen": 3037790208
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.034102306920763e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.3460346528140903,
      "tokens_seen": 3037855744
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.0320962888666e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.3460291640985886,
      "tokens_seen": 3037921280
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.030090270812437e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.346023675534645,
      "tokens_seen": 3037986816
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.028084252758275e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.3460181871222514,
      "tokens_seen": 3038052352
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.026078234704112e-05,
      "loss": 1.217,
      "theoretical_loss": 3.3460126988614007,
      "tokens_seen": 3038117888
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.02407221664995e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.3460072107520853,
      "tokens_seen": 3038183424
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.022066198595787e-05,
      "loss": 1.1944,
      "theoretical_loss": 3.346001722794298,
      "tokens_seen": 3038248960
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.020060180541625e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.345996234988031,
      "tokens_seen": 3038314496
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.018054162487462e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.345990747333277,
      "tokens_seen": 3038380032
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.0160481444333e-05,
      "loss": 1.1755,
      "theoretical_loss": 3.3459852598300284,
      "tokens_seen": 3038445568
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.014042126379138e-05,
      "loss": 1.1279,
      "theoretical_loss": 3.3459797724782785,
      "tokens_seen": 3038511104
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.012036108324975e-05,
      "loss": 1.2043,
      "theoretical_loss": 3.3459742852780185,
      "tokens_seen": 3038576640
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.010030090270813e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.3459687982292423,
      "tokens_seen": 3038642176
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.00802407221665e-05,
      "loss": 1.1954,
      "theoretical_loss": 3.345963311331942,
      "tokens_seen": 3038707712
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.006018054162488e-05,
      "loss": 1.1452,
      "theoretical_loss": 3.3459578245861095,
      "tokens_seen": 3038773248
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.004012036108325e-05,
      "loss": 1.3296,
      "theoretical_loss": 3.345952337991738,
      "tokens_seen": 3038838784
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.002006018054163e-05,
      "loss": 1.2302,
      "theoretical_loss": 3.34594685154882,
      "tokens_seen": 3038904320
    },
    {
      "epoch": 0.84,
      "learning_rate": 8e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.345941365257348,
      "tokens_seen": 3038969856
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.997993981945838e-05,
      "loss": 1.1288,
      "theoretical_loss": 3.3459358791173144,
      "tokens_seen": 3039035392
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.995987963891676e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.345930393128712,
      "tokens_seen": 3039100928
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.993981945837512e-05,
      "loss": 1.3435,
      "theoretical_loss": 3.3459249072915336,
      "tokens_seen": 3039166464
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.99197592778335e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.345919421605771,
      "tokens_seen": 3039232000
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.989969909729187e-05,
      "loss": 1.2168,
      "theoretical_loss": 3.345913936071417,
      "tokens_seen": 3039297536
    },
    {
      "epoch": 0.84,
      "objective/train/docs_used": 1104383,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1367526054382324,
      "objective/train/theoretical_loss": 3.3459084506884644,
      "objective/train/tokens_used": 1409822176,
      "theoretical_loss": 3.3459084506884644,
      "tokens_seen": 3039363072
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.987963891675025e-05,
      "loss": 1.2062,
      "theoretical_loss": 3.3459084506884644,
      "tokens_seen": 3039363072
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.985957873620862e-05,
      "loss": 1.2118,
      "theoretical_loss": 3.345902965456906,
      "tokens_seen": 3039428608
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.9839518555667e-05,
      "loss": 1.2667,
      "theoretical_loss": 3.345897480376734,
      "tokens_seen": 3039494144
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.981945837512538e-05,
      "loss": 1.3485,
      "theoretical_loss": 3.3458919954479405,
      "tokens_seen": 3039559680
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.979939819458375e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.345886510670519,
      "tokens_seen": 3039625216
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.977933801404213e-05,
      "loss": 1.2881,
      "theoretical_loss": 3.345881026044461,
      "tokens_seen": 3039690752
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.975927783350052e-05,
      "loss": 1.2018,
      "theoretical_loss": 3.34587554156976,
      "tokens_seen": 3039756288
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.973921765295888e-05,
      "loss": 1.299,
      "theoretical_loss": 3.3458700572464086,
      "tokens_seen": 3039821824
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.971915747241725e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.3458645730743988,
      "tokens_seen": 3039887360
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.969909729187563e-05,
      "loss": 1.1834,
      "theoretical_loss": 3.345859089053723,
      "tokens_seen": 3039952896
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.9679037111334e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.345853605184374,
      "tokens_seen": 3040018432
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.965897693079238e-05,
      "loss": 1.1839,
      "theoretical_loss": 3.345848121466345,
      "tokens_seen": 3040083968
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.963891675025076e-05,
      "loss": 1.2249,
      "theoretical_loss": 3.345842637899628,
      "tokens_seen": 3040149504
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.961885656970913e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.345837154484215,
      "tokens_seen": 3040215040
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.959879638916751e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.3458316712201,
      "tokens_seen": 3040280576
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.957873620862588e-05,
      "loss": 1.2312,
      "theoretical_loss": 3.345826188107274,
      "tokens_seen": 3040346112
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.955867602808425e-05,
      "loss": 1.1766,
      "theoretical_loss": 3.3458207051457305,
      "tokens_seen": 3040411648
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.953861584754262e-05,
      "loss": 1.207,
      "theoretical_loss": 3.345815222335462,
      "tokens_seen": 3040477184
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.9518555667001e-05,
      "loss": 1.2766,
      "theoretical_loss": 3.3458097396764606,
      "tokens_seen": 3040542720
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.949849548645937e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.3458042571687194,
      "tokens_seen": 3040608256
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.947843530591775e-05,
      "loss": 1.2565,
      "theoretical_loss": 3.3457987748122306,
      "tokens_seen": 3040673792
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.945837512537613e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.345793292606987,
      "tokens_seen": 3040739328
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.94383149448345e-05,
      "loss": 1.3532,
      "theoretical_loss": 3.3457878105529812,
      "tokens_seen": 3040804864
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.941825476429288e-05,
      "loss": 1.2469,
      "theoretical_loss": 3.3457823286502055,
      "tokens_seen": 3040870400
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.939819458375125e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.345776846898653,
      "tokens_seen": 3040935936
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.937813440320964e-05,
      "loss": 1.1849,
      "theoretical_loss": 3.345771365298315,
      "tokens_seen": 3041001472
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.9358074222668e-05,
      "loss": 1.2494,
      "theoretical_loss": 3.345765883849186,
      "tokens_seen": 3041067008
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.933801404212638e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.345760402551257,
      "tokens_seen": 3041132544
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.931795386158476e-05,
      "loss": 1.3831,
      "theoretical_loss": 3.345754921404521,
      "tokens_seen": 3041198080
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.929789368104313e-05,
      "loss": 1.2123,
      "theoretical_loss": 3.345749440408971,
      "tokens_seen": 3041263616
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.927783350050151e-05,
      "loss": 1.3638,
      "theoretical_loss": 3.345743959564599,
      "tokens_seen": 3041329152
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.925777331995988e-05,
      "loss": 1.2335,
      "theoretical_loss": 3.345738478871398,
      "tokens_seen": 3041394688
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.923771313941826e-05,
      "loss": 1.2094,
      "theoretical_loss": 3.34573299832936,
      "tokens_seen": 3041460224
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.921765295887664e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.3457275179384784,
      "tokens_seen": 3041525760
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.919759277833501e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.345722037698745,
      "tokens_seen": 3041591296
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.917753259779339e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.345716557610153,
      "tokens_seen": 3041656832
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.915747241725175e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.3457110776726946,
      "tokens_seen": 3041722368
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.913741223671012e-05,
      "loss": 1.3873,
      "theoretical_loss": 3.3457055978863623,
      "tokens_seen": 3041787904
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.91173520561685e-05,
      "loss": 1.3208,
      "theoretical_loss": 3.345700118251149,
      "tokens_seen": 3041853440
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.909729187562688e-05,
      "loss": 1.17,
      "theoretical_loss": 3.345694638767047,
      "tokens_seen": 3041918976
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.907723169508525e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.345689159434049,
      "tokens_seen": 3041984512
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.905717151454363e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.345683680252148,
      "tokens_seen": 3042050048
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.9037111334002e-05,
      "loss": 1.1907,
      "theoretical_loss": 3.345678201221336,
      "tokens_seen": 3042115584
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.901705115346038e-05,
      "loss": 1.4159,
      "theoretical_loss": 3.345672722341605,
      "tokens_seen": 3042181120
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.899699097291877e-05,
      "loss": 1.4238,
      "theoretical_loss": 3.345667243612949,
      "tokens_seen": 3042246656
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.897693079237714e-05,
      "loss": 1.2007,
      "theoretical_loss": 3.3456617650353597,
      "tokens_seen": 3042312192
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.895687061183551e-05,
      "loss": 1.291,
      "theoretical_loss": 3.34565628660883,
      "tokens_seen": 3042377728
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.893681043129388e-05,
      "loss": 1.2181,
      "theoretical_loss": 3.3456508083333523,
      "tokens_seen": 3042443264
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.891675025075226e-05,
      "loss": 1.3695,
      "theoretical_loss": 3.345645330208919,
      "tokens_seen": 3042508800
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.889669007021063e-05,
      "loss": 1.295,
      "theoretical_loss": 3.3456398522355233,
      "tokens_seen": 3042574336
    },
    {
      "epoch": 0.84,
      "objective/train/docs_used": 1105520,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.30586838722229,
      "objective/train/theoretical_loss": 3.345634374413157,
      "objective/train/tokens_used": 1413098976,
      "theoretical_loss": 3.345634374413157,
      "tokens_seen": 3042639872
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.887662988966901e-05,
      "loss": 1.3113,
      "theoretical_loss": 3.345634374413157,
      "tokens_seen": 3042639872
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.885656970912739e-05,
      "loss": 1.0604,
      "theoretical_loss": 3.3456288967418133,
      "tokens_seen": 3042705408
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.883650952858576e-05,
      "loss": 1.2855,
      "theoretical_loss": 3.3456234192214844,
      "tokens_seen": 3042770944
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.881644934804414e-05,
      "loss": 1.4043,
      "theoretical_loss": 3.345617941852163,
      "tokens_seen": 3042836480
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.879638916750251e-05,
      "loss": 1.3062,
      "theoretical_loss": 3.345612464633842,
      "tokens_seen": 3042902016
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.877632898696088e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.345606987566514,
      "tokens_seen": 3042967552
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.875626880641925e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.345601510650171,
      "tokens_seen": 3043033088
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.873620862587763e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.3455960338848056,
      "tokens_seen": 3043098624
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.8716148445336e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.3455905572704108,
      "tokens_seen": 3043164160
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.869608826479438e-05,
      "loss": 1.1424,
      "theoretical_loss": 3.345585080806979,
      "tokens_seen": 3043229696
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.867602808425275e-05,
      "loss": 1.1993,
      "theoretical_loss": 3.345579604494503,
      "tokens_seen": 3043295232
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.865596790371113e-05,
      "loss": 1.2826,
      "theoretical_loss": 3.3455741283329754,
      "tokens_seen": 3043360768
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.863590772316952e-05,
      "loss": 1.2021,
      "theoretical_loss": 3.345568652322388,
      "tokens_seen": 3043426304
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.86158475426279e-05,
      "loss": 1.4122,
      "theoretical_loss": 3.345563176462735,
      "tokens_seen": 3043491840
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.859578736208627e-05,
      "loss": 1.2292,
      "theoretical_loss": 3.345557700754007,
      "tokens_seen": 3043557376
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.857572718154463e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.345552225196198,
      "tokens_seen": 3043622912
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.855566700100301e-05,
      "loss": 1.2662,
      "theoretical_loss": 3.3455467497893006,
      "tokens_seen": 3043688448
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.853560682046139e-05,
      "loss": 1.2496,
      "theoretical_loss": 3.3455412745333066,
      "tokens_seen": 3043753984
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.851554663991976e-05,
      "loss": 1.2065,
      "theoretical_loss": 3.345535799428209,
      "tokens_seen": 3043819520
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.849548645937814e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.3455303244740007,
      "tokens_seen": 3043885056
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.847542627883651e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.3455248496706735,
      "tokens_seen": 3043950592
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.845536609829489e-05,
      "loss": 1.245,
      "theoretical_loss": 3.3455193750182204,
      "tokens_seen": 3044016128
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.843530591775326e-05,
      "loss": 1.3172,
      "theoretical_loss": 3.345513900516634,
      "tokens_seen": 3044081664
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.841524573721164e-05,
      "loss": 1.2299,
      "theoretical_loss": 3.3455084261659076,
      "tokens_seen": 3044147200
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.839518555667002e-05,
      "loss": 1.2205,
      "theoretical_loss": 3.345502951966033,
      "tokens_seen": 3044212736
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.837512537612838e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.3454974779170024,
      "tokens_seen": 3044278272
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.835506519558675e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.345492004018809,
      "tokens_seen": 3044343808
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.833500501504513e-05,
      "loss": 1.3151,
      "theoretical_loss": 3.3454865302714456,
      "tokens_seen": 3044409344
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.83149448345035e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.3454810566749043,
      "tokens_seen": 3044474880
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.829488465396188e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.3454755832291783,
      "tokens_seen": 3044540416
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.827482447342026e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.3454701099342596,
      "tokens_seen": 3044605952
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.825476429287865e-05,
      "loss": 1.1928,
      "theoretical_loss": 3.345464636790141,
      "tokens_seen": 3044671488
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.823470411233702e-05,
      "loss": 1.2542,
      "theoretical_loss": 3.3454591637968147,
      "tokens_seen": 3044737024
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.82146439317954e-05,
      "loss": 1.2294,
      "theoretical_loss": 3.3454536909542743,
      "tokens_seen": 3044802560
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.819458375125377e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.3454482182625114,
      "tokens_seen": 3044868096
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.817452357071214e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.3454427457215195,
      "tokens_seen": 3044933632
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.815446339017051e-05,
      "loss": 1.296,
      "theoretical_loss": 3.3454372733312905,
      "tokens_seen": 3044999168
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.813440320962889e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.345431801091817,
      "tokens_seen": 3045064704
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.811434302908726e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.345426329003092,
      "tokens_seen": 3045130240
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.809428284854564e-05,
      "loss": 1.2766,
      "theoretical_loss": 3.3454208570651076,
      "tokens_seen": 3045195776
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.807422266800401e-05,
      "loss": 1.2456,
      "theoretical_loss": 3.345415385277857,
      "tokens_seen": 3045261312
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.805416248746239e-05,
      "loss": 1.1379,
      "theoretical_loss": 3.345409913641333,
      "tokens_seen": 3045326848
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.803410230692077e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.345404442155527,
      "tokens_seen": 3045392384
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.801404212637914e-05,
      "loss": 1.1134,
      "theoretical_loss": 3.345398970820433,
      "tokens_seen": 3045457920
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.79939819458375e-05,
      "loss": 1.1874,
      "theoretical_loss": 3.3453934996360424,
      "tokens_seen": 3045523456
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.797392176529588e-05,
      "loss": 1.3489,
      "theoretical_loss": 3.3453880286023487,
      "tokens_seen": 3045588992
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.795386158475426e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.345382557719344,
      "tokens_seen": 3045654528
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.793380140421263e-05,
      "loss": 1.2418,
      "theoretical_loss": 3.345377086987021,
      "tokens_seen": 3045720064
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.791374122367101e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.345371616405372,
      "tokens_seen": 3045785600
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.789368104312938e-05,
      "loss": 1.3273,
      "theoretical_loss": 3.345366145974391,
      "tokens_seen": 3045851136
    },
    {
      "epoch": 0.85,
      "objective/train/docs_used": 1106703,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1013998985290527,
      "objective/train/theoretical_loss": 3.345360675694069,
      "objective/train/tokens_used": 1416375776,
      "theoretical_loss": 3.345360675694069,
      "tokens_seen": 3045916672
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.787362086258777e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.345360675694069,
      "tokens_seen": 3045916672
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.785356068204615e-05,
      "loss": 1.2663,
      "theoretical_loss": 3.345355205564399,
      "tokens_seen": 3045982208
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.783350050150452e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.3453497355853745,
      "tokens_seen": 3046047744
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.78134403209629e-05,
      "loss": 1.272,
      "theoretical_loss": 3.345344265756987,
      "tokens_seen": 3046113280
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.779338014042126e-05,
      "loss": 1.2552,
      "theoretical_loss": 3.345338796079229,
      "tokens_seen": 3046178816
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.777331995987964e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.3453333265520944,
      "tokens_seen": 3046244352
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.775325977933801e-05,
      "loss": 1.1895,
      "theoretical_loss": 3.345327857175575,
      "tokens_seen": 3046309888
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.773319959879639e-05,
      "loss": 1.281,
      "theoretical_loss": 3.345322387949663,
      "tokens_seen": 3046375424
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.771313941825477e-05,
      "loss": 1.2038,
      "theoretical_loss": 3.345316918874352,
      "tokens_seen": 3046440960
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.769307923771314e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.3453114499496337,
      "tokens_seen": 3046506496
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.767301905717152e-05,
      "loss": 1.264,
      "theoretical_loss": 3.3453059811755015,
      "tokens_seen": 3046572032
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.765295887662989e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.345300512551947,
      "tokens_seen": 3046637568
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.763289869608827e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.3452950440789637,
      "tokens_seen": 3046703104
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.761283851554664e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.3452895757565444,
      "tokens_seen": 3046768640
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.759277833500501e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.345284107584681,
      "tokens_seen": 3046834176
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.757271815446338e-05,
      "loss": 1.3483,
      "theoretical_loss": 3.3452786395633662,
      "tokens_seen": 3046899712
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.755265797392176e-05,
      "loss": 1.2527,
      "theoretical_loss": 3.3452731716925928,
      "tokens_seen": 3046965248
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.753259779338013e-05,
      "loss": 1.2011,
      "theoretical_loss": 3.345267703972354,
      "tokens_seen": 3047030784
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.751253761283852e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.345262236402641,
      "tokens_seen": 3047096320
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.74924774322969e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.3452567689834476,
      "tokens_seen": 3047161856
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.747241725175528e-05,
      "loss": 1.2859,
      "theoretical_loss": 3.3452513017147663,
      "tokens_seen": 3047227392
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.745235707121365e-05,
      "loss": 1.3824,
      "theoretical_loss": 3.3452458345965894,
      "tokens_seen": 3047292928
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.743229689067203e-05,
      "loss": 1.2301,
      "theoretical_loss": 3.3452403676289095,
      "tokens_seen": 3047358464
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.74122367101304e-05,
      "loss": 1.2148,
      "theoretical_loss": 3.3452349008117195,
      "tokens_seen": 3047424000
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.739217652958876e-05,
      "loss": 1.302,
      "theoretical_loss": 3.3452294341450117,
      "tokens_seen": 3047489536
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.737211634904714e-05,
      "loss": 1.2552,
      "theoretical_loss": 3.3452239676287787,
      "tokens_seen": 3047555072
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.735205616850552e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.345218501263014,
      "tokens_seen": 3047620608
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.733199598796389e-05,
      "loss": 1.3161,
      "theoretical_loss": 3.345213035047709,
      "tokens_seen": 3047686144
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.731193580742227e-05,
      "loss": 1.262,
      "theoretical_loss": 3.3452075689828566,
      "tokens_seen": 3047751680
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.729187562688064e-05,
      "loss": 1.3165,
      "theoretical_loss": 3.3452021030684502,
      "tokens_seen": 3047817216
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.727181544633902e-05,
      "loss": 1.3408,
      "theoretical_loss": 3.3451966373044817,
      "tokens_seen": 3047882752
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.72517552657974e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.345191171690944,
      "tokens_seen": 3047948288
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.723169508525577e-05,
      "loss": 1.3317,
      "theoretical_loss": 3.345185706227829,
      "tokens_seen": 3048013824
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.721163490471413e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.3451802409151306,
      "tokens_seen": 3048079360
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.719157472417251e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.3451747757528407,
      "tokens_seen": 3048144896
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.717151454363088e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.3451693107409524,
      "tokens_seen": 3048210432
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.715145436308926e-05,
      "loss": 1.2745,
      "theoretical_loss": 3.3451638458794575,
      "tokens_seen": 3048275968
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.713139418254765e-05,
      "loss": 1.2469,
      "theoretical_loss": 3.345158381168349,
      "tokens_seen": 3048341504
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.711133400200603e-05,
      "loss": 1.2074,
      "theoretical_loss": 3.3451529166076197,
      "tokens_seen": 3048407040
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.70912738214644e-05,
      "loss": 1.3387,
      "theoretical_loss": 3.3451474521972626,
      "tokens_seen": 3048472576
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.707121364092278e-05,
      "loss": 1.1404,
      "theoretical_loss": 3.3451419879372692,
      "tokens_seen": 3048538112
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.705115346038115e-05,
      "loss": 1.1787,
      "theoretical_loss": 3.3451365238276334,
      "tokens_seen": 3048603648
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.703109327983953e-05,
      "loss": 1.3549,
      "theoretical_loss": 3.345131059868347,
      "tokens_seen": 3048669184
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.701103309929789e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.3451255960594026,
      "tokens_seen": 3048734720
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.699097291875627e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.345120132400793,
      "tokens_seen": 3048800256
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.697091273821464e-05,
      "loss": 1.3509,
      "theoretical_loss": 3.3451146688925117,
      "tokens_seen": 3048865792
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.695085255767302e-05,
      "loss": 1.2119,
      "theoretical_loss": 3.3451092055345497,
      "tokens_seen": 3048931328
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.69307923771314e-05,
      "loss": 1.3092,
      "theoretical_loss": 3.345103742326901,
      "tokens_seen": 3048996864
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.691073219658977e-05,
      "loss": 1.1888,
      "theoretical_loss": 3.345098279269558,
      "tokens_seen": 3049062400
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.689067201604815e-05,
      "loss": 1.239,
      "theoretical_loss": 3.345092816362512,
      "tokens_seen": 3049127936
    },
    {
      "epoch": 0.85,
      "objective/train/docs_used": 1107994,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2179710865020752,
      "objective/train/theoretical_loss": 3.3450873536057575,
      "objective/train/tokens_used": 1419652576,
      "theoretical_loss": 3.3450873536057575,
      "tokens_seen": 3049193472
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.687061183550652e-05,
      "loss": 1.1453,
      "theoretical_loss": 3.3450873536057575,
      "tokens_seen": 3049193472
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.68505516549649e-05,
      "loss": 1.2262,
      "theoretical_loss": 3.3450818909992863,
      "tokens_seen": 3049259008
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.683049147442327e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.345076428543091,
      "tokens_seen": 3049324544
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.681043129388164e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.3450709662371643,
      "tokens_seen": 3049390080
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.679037111334001e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.3450655040814987,
      "tokens_seen": 3049455616
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.677031093279839e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.3450600420760868,
      "tokens_seen": 3049521152
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.675025075225678e-05,
      "loss": 1.2185,
      "theoretical_loss": 3.3450545802209217,
      "tokens_seen": 3049586688
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.673019057171515e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.3450491185159956,
      "tokens_seen": 3049652224
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.671013039117353e-05,
      "loss": 1.2031,
      "theoretical_loss": 3.345043656961302,
      "tokens_seen": 3049717760
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.66900702106319e-05,
      "loss": 1.2488,
      "theoretical_loss": 3.345038195556832,
      "tokens_seen": 3049783296
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.667001003009028e-05,
      "loss": 1.261,
      "theoretical_loss": 3.3450327343025794,
      "tokens_seen": 3049848832
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.664994984954866e-05,
      "loss": 1.2556,
      "theoretical_loss": 3.345027273198536,
      "tokens_seen": 3049914368
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.662988966900703e-05,
      "loss": 1.1875,
      "theoretical_loss": 3.3450218122446955,
      "tokens_seen": 3049979904
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.66098294884654e-05,
      "loss": 1.2153,
      "theoretical_loss": 3.34501635144105,
      "tokens_seen": 3050045440
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.658976930792377e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.345010890787592,
      "tokens_seen": 3050110976
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.656970912738215e-05,
      "loss": 1.1388,
      "theoretical_loss": 3.3450054302843144,
      "tokens_seen": 3050176512
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.654964894684052e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.3449999699312096,
      "tokens_seen": 3050242048
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.65295887662989e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.34499450972827,
      "tokens_seen": 3050307584
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.650952858575727e-05,
      "loss": 1.328,
      "theoretical_loss": 3.3449890496754895,
      "tokens_seen": 3050373120
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.648946840521565e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.344983589772859,
      "tokens_seen": 3050438656
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.646940822467402e-05,
      "loss": 1.2059,
      "theoretical_loss": 3.3449781300203725,
      "tokens_seen": 3050504192
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.64493480441324e-05,
      "loss": 1.3123,
      "theoretical_loss": 3.344972670418022,
      "tokens_seen": 3050569728
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.642928786359076e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.3449672109658,
      "tokens_seen": 3050635264
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.640922768304914e-05,
      "loss": 1.258,
      "theoretical_loss": 3.3449617516637,
      "tokens_seen": 3050700800
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.638916750250751e-05,
      "loss": 1.269,
      "theoretical_loss": 3.3449562925117133,
      "tokens_seen": 3050766336
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.63691073219659e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.344950833509834,
      "tokens_seen": 3050831872
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.634904714142428e-05,
      "loss": 1.243,
      "theoretical_loss": 3.344945374658054,
      "tokens_seen": 3050897408
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.632898696088265e-05,
      "loss": 1.1236,
      "theoretical_loss": 3.344939915956366,
      "tokens_seen": 3050962944
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.630892678034103e-05,
      "loss": 1.2043,
      "theoretical_loss": 3.3449344574047624,
      "tokens_seen": 3051028480
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.62888665997994e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.3449289990032365,
      "tokens_seen": 3051094016
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.626880641925778e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.3449235407517803,
      "tokens_seen": 3051159552
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.624874623871616e-05,
      "loss": 1.2791,
      "theoretical_loss": 3.344918082650387,
      "tokens_seen": 3051225088
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.622868605817452e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.344912624699049,
      "tokens_seen": 3051290624
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.62086258776329e-05,
      "loss": 1.1661,
      "theoretical_loss": 3.3449071668977584,
      "tokens_seen": 3051356160
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.618856569709127e-05,
      "loss": 1.288,
      "theoretical_loss": 3.3449017092465088,
      "tokens_seen": 3051421696
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.616850551654965e-05,
      "loss": 1.1701,
      "theoretical_loss": 3.3448962517452925,
      "tokens_seen": 3051487232
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.614844533600802e-05,
      "loss": 1.1225,
      "theoretical_loss": 3.344890794394102,
      "tokens_seen": 3051552768
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.61283851554664e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.3448853371929297,
      "tokens_seen": 3051618304
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.610832497492477e-05,
      "loss": 1.2091,
      "theoretical_loss": 3.344879880141769,
      "tokens_seen": 3051683840
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.608826479438315e-05,
      "loss": 1.2292,
      "theoretical_loss": 3.3448744232406122,
      "tokens_seen": 3051749376
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.606820461384153e-05,
      "loss": 1.2552,
      "theoretical_loss": 3.3448689664894515,
      "tokens_seen": 3051814912
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.60481444332999e-05,
      "loss": 1.3826,
      "theoretical_loss": 3.34486350988828,
      "tokens_seen": 3051880448
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.602808425275826e-05,
      "loss": 1.1521,
      "theoretical_loss": 3.3448580534370906,
      "tokens_seen": 3051945984
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.600802407221665e-05,
      "loss": 1.3393,
      "theoretical_loss": 3.344852597135876,
      "tokens_seen": 3052011520
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.598796389167503e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.344847140984628,
      "tokens_seen": 3052077056
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.59679037111334e-05,
      "loss": 1.2322,
      "theoretical_loss": 3.34484168498334,
      "tokens_seen": 3052142592
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.594784353059178e-05,
      "loss": 1.3441,
      "theoretical_loss": 3.3448362291320044,
      "tokens_seen": 3052208128
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.592778335005016e-05,
      "loss": 1.2018,
      "theoretical_loss": 3.344830773430614,
      "tokens_seen": 3052273664
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.590772316950853e-05,
      "loss": 1.2171,
      "theoretical_loss": 3.3448253178791614,
      "tokens_seen": 3052339200
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.588766298896691e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.3448198624776393,
      "tokens_seen": 3052404736
    },
    {
      "epoch": 0.85,
      "objective/train/docs_used": 1109048,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2406476736068726,
      "objective/train/theoretical_loss": 3.34481440722604,
      "objective/train/tokens_used": 1422929376,
      "theoretical_loss": 3.34481440722604,
      "tokens_seen": 3052470272
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.586760280842528e-05,
      "loss": 1.2638,
      "theoretical_loss": 3.34481440722604,
      "tokens_seen": 3052470272
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.584754262788365e-05,
      "loss": 1.1846,
      "theoretical_loss": 3.3448089521243567,
      "tokens_seen": 3052535808
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.582748244734202e-05,
      "loss": 1.3334,
      "theoretical_loss": 3.3448034971725815,
      "tokens_seen": 3052601344
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.58074222668004e-05,
      "loss": 1.3485,
      "theoretical_loss": 3.3447980423707078,
      "tokens_seen": 3052666880
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.578736208625877e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.3447925877187275,
      "tokens_seen": 3052732416
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.576730190571715e-05,
      "loss": 1.2404,
      "theoretical_loss": 3.3447871332166335,
      "tokens_seen": 3052797952
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.574724172517553e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.344781678864419,
      "tokens_seen": 3052863488
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.57271815446339e-05,
      "loss": 1.135,
      "theoretical_loss": 3.344776224662076,
      "tokens_seen": 3052929024
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.570712136409228e-05,
      "loss": 1.1667,
      "theoretical_loss": 3.3447707706095975,
      "tokens_seen": 3052994560
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.568706118355065e-05,
      "loss": 1.2028,
      "theoretical_loss": 3.344765316706976,
      "tokens_seen": 3053060096
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.566700100300903e-05,
      "loss": 1.2377,
      "theoretical_loss": 3.344759862954204,
      "tokens_seen": 3053125632
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.564694082246739e-05,
      "loss": 1.3642,
      "theoretical_loss": 3.3447544093512747,
      "tokens_seen": 3053191168
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.562688064192578e-05,
      "loss": 1.212,
      "theoretical_loss": 3.34474895589818,
      "tokens_seen": 3053256704
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.560682046138416e-05,
      "loss": 1.1834,
      "theoretical_loss": 3.344743502594914,
      "tokens_seen": 3053322240
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.558676028084253e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.3447380494414674,
      "tokens_seen": 3053387776
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.556670010030091e-05,
      "loss": 1.2072,
      "theoretical_loss": 3.3447325964378347,
      "tokens_seen": 3053453312
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.554663991975928e-05,
      "loss": 1.3418,
      "theoretical_loss": 3.344727143584007,
      "tokens_seen": 3053518848
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.552657973921766e-05,
      "loss": 1.3296,
      "theoretical_loss": 3.344721690879978,
      "tokens_seen": 3053584384
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.550651955867604e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.3447162383257405,
      "tokens_seen": 3053649920
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.548645937813441e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.3447107859212863,
      "tokens_seen": 3053715456
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.546639919759279e-05,
      "loss": 1.1581,
      "theoretical_loss": 3.3447053336666084,
      "tokens_seen": 3053780992
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.544633901705115e-05,
      "loss": 1.2152,
      "theoretical_loss": 3.3446998815616995,
      "tokens_seen": 3053846528
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.542627883650952e-05,
      "loss": 1.3208,
      "theoretical_loss": 3.3446944296065526,
      "tokens_seen": 3053912064
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.54062186559679e-05,
      "loss": 1.2924,
      "theoretical_loss": 3.34468897780116,
      "tokens_seen": 3053977600
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.538615847542628e-05,
      "loss": 1.2169,
      "theoretical_loss": 3.344683526145515,
      "tokens_seen": 3054043136
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.536609829488465e-05,
      "loss": 1.1749,
      "theoretical_loss": 3.344678074639609,
      "tokens_seen": 3054108672
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.534603811434303e-05,
      "loss": 1.2251,
      "theoretical_loss": 3.344672623283436,
      "tokens_seen": 3054174208
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.53259779338014e-05,
      "loss": 1.1302,
      "theoretical_loss": 3.344667172076988,
      "tokens_seen": 3054239744
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.530591775325978e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.344661721020258,
      "tokens_seen": 3054305280
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.528585757271816e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.344656270113238,
      "tokens_seen": 3054370816
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.526579739217652e-05,
      "loss": 1.2062,
      "theoretical_loss": 3.3446508193559215,
      "tokens_seen": 3054436352
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.524573721163491e-05,
      "loss": 1.1535,
      "theoretical_loss": 3.3446453687483007,
      "tokens_seen": 3054501888
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.522567703109328e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.3446399182903686,
      "tokens_seen": 3054567424
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.520561685055166e-05,
      "loss": 1.2164,
      "theoretical_loss": 3.344634467982117,
      "tokens_seen": 3054632960
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.518555667001003e-05,
      "loss": 1.2359,
      "theoretical_loss": 3.3446290178235403,
      "tokens_seen": 3054698496
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.516549648946841e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.3446235678146294,
      "tokens_seen": 3054764032
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.514543630892679e-05,
      "loss": 1.1268,
      "theoretical_loss": 3.344618117955378,
      "tokens_seen": 3054829568
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.512537612838516e-05,
      "loss": 1.182,
      "theoretical_loss": 3.3446126682457784,
      "tokens_seen": 3054895104
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.510531594784354e-05,
      "loss": 1.1957,
      "theoretical_loss": 3.3446072186858236,
      "tokens_seen": 3054960640
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.508525576730191e-05,
      "loss": 1.1926,
      "theoretical_loss": 3.344601769275506,
      "tokens_seen": 3055026176
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.506519558676028e-05,
      "loss": 1.2495,
      "theoretical_loss": 3.3445963200148183,
      "tokens_seen": 3055091712
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.504513540621865e-05,
      "loss": 1.1646,
      "theoretical_loss": 3.344590870903753,
      "tokens_seen": 3055157248
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.502507522567703e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.3445854219423032,
      "tokens_seen": 3055222784
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.50050150451354e-05,
      "loss": 1.1632,
      "theoretical_loss": 3.3445799731304615,
      "tokens_seen": 3055288320
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.498495486459378e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.34457452446822,
      "tokens_seen": 3055353856
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.496489468405215e-05,
      "loss": 1.3392,
      "theoretical_loss": 3.3445690759555724,
      "tokens_seen": 3055419392
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.494483450351053e-05,
      "loss": 1.2089,
      "theoretical_loss": 3.344563627592511,
      "tokens_seen": 3055484928
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.49247743229689e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.3445581793790278,
      "tokens_seen": 3055550464
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.490471414242728e-05,
      "loss": 1.2053,
      "theoretical_loss": 3.344552731315116,
      "tokens_seen": 3055616000
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.488465396188567e-05,
      "loss": 1.317,
      "theoretical_loss": 3.3445472834007686,
      "tokens_seen": 3055681536
    },
    {
      "epoch": 0.85,
      "objective/train/docs_used": 1110166,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.205101490020752,
      "objective/train/theoretical_loss": 3.3445418356359777,
      "objective/train/tokens_used": 1426206176,
      "theoretical_loss": 3.3445418356359777,
      "tokens_seen": 3055747072
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.486459378134403e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.3445418356359777,
      "tokens_seen": 3055747072
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.484453360080241e-05,
      "loss": 1.0473,
      "theoretical_loss": 3.3445363880207366,
      "tokens_seen": 3055812608
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.482447342026079e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.344530940555037,
      "tokens_seen": 3055878144
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.480441323971916e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.3445254932388733,
      "tokens_seen": 3055943680
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.478435305917754e-05,
      "loss": 1.1828,
      "theoretical_loss": 3.3445200460722364,
      "tokens_seen": 3056009216
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.476429287863591e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.34451459905512,
      "tokens_seen": 3056074752
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.474423269809429e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.3445091521875163,
      "tokens_seen": 3056140288
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.472417251755266e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.3445037054694184,
      "tokens_seen": 3056205824
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.470411233701104e-05,
      "loss": 1.2063,
      "theoretical_loss": 3.3444982589008188,
      "tokens_seen": 3056271360
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.468405215646942e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.34449281248171,
      "tokens_seen": 3056336896
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.466399197592778e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.344487366212085,
      "tokens_seen": 3056402432
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.464393179538615e-05,
      "loss": 1.1827,
      "theoretical_loss": 3.3444819200919365,
      "tokens_seen": 3056467968
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.462387161484453e-05,
      "loss": 1.2308,
      "theoretical_loss": 3.3444764741212567,
      "tokens_seen": 3056533504
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.46038114343029e-05,
      "loss": 1.139,
      "theoretical_loss": 3.3444710283000387,
      "tokens_seen": 3056599040
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.458375125376128e-05,
      "loss": 1.2342,
      "theoretical_loss": 3.3444655826282754,
      "tokens_seen": 3056664576
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.456369107321966e-05,
      "loss": 1.2063,
      "theoretical_loss": 3.3444601371059592,
      "tokens_seen": 3056730112
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.454363089267803e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.344454691733083,
      "tokens_seen": 3056795648
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.452357071213641e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.3444492465096394,
      "tokens_seen": 3056861184
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.45035105315948e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.3444438014356206,
      "tokens_seen": 3056926720
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.448345035105317e-05,
      "loss": 1.199,
      "theoretical_loss": 3.3444383565110196,
      "tokens_seen": 3056992256
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.446339017051154e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.3444329117358294,
      "tokens_seen": 3057057792
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.444332998996991e-05,
      "loss": 1.277,
      "theoretical_loss": 3.3444274671100427,
      "tokens_seen": 3057123328
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.442326980942829e-05,
      "loss": 1.1964,
      "theoretical_loss": 3.344422022633652,
      "tokens_seen": 3057188864
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.440320962888666e-05,
      "loss": 1.2089,
      "theoretical_loss": 3.34441657830665,
      "tokens_seen": 3057254400
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.438314944834504e-05,
      "loss": 1.1274,
      "theoretical_loss": 3.3444111341290297,
      "tokens_seen": 3057319936
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.436308926780341e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.344405690100783,
      "tokens_seen": 3057385472
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.434302908726179e-05,
      "loss": 1.1875,
      "theoretical_loss": 3.344400246221903,
      "tokens_seen": 3057451008
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.432296890672017e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.344394802492383,
      "tokens_seen": 3057516544
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.430290872617854e-05,
      "loss": 1.3029,
      "theoretical_loss": 3.344389358912215,
      "tokens_seen": 3057582080
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.42828485456369e-05,
      "loss": 1.2308,
      "theoretical_loss": 3.344383915481392,
      "tokens_seen": 3057647616
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.426278836509528e-05,
      "loss": 1.0972,
      "theoretical_loss": 3.3443784721999066,
      "tokens_seen": 3057713152
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.424272818455366e-05,
      "loss": 1.1398,
      "theoretical_loss": 3.3443730290677514,
      "tokens_seen": 3057778688
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.422266800401203e-05,
      "loss": 1.2086,
      "theoretical_loss": 3.3443675860849194,
      "tokens_seen": 3057844224
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.420260782347041e-05,
      "loss": 1.2069,
      "theoretical_loss": 3.344362143251403,
      "tokens_seen": 3057909760
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.418254764292878e-05,
      "loss": 1.2721,
      "theoretical_loss": 3.344356700567195,
      "tokens_seen": 3057975296
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.416248746238716e-05,
      "loss": 1.1739,
      "theoretical_loss": 3.3443512580322885,
      "tokens_seen": 3058040832
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.414242728184553e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.344345815646675,
      "tokens_seen": 3058106368
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.412236710130392e-05,
      "loss": 1.1791,
      "theoretical_loss": 3.3443403734103487,
      "tokens_seen": 3058171904
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.41023069207623e-05,
      "loss": 1.1926,
      "theoretical_loss": 3.344334931323302,
      "tokens_seen": 3058237440
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.408224674022066e-05,
      "loss": 1.1356,
      "theoretical_loss": 3.3443294893855264,
      "tokens_seen": 3058302976
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.406218655967904e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.344324047597016,
      "tokens_seen": 3058368512
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.404212637913741e-05,
      "loss": 1.2529,
      "theoretical_loss": 3.344318605957763,
      "tokens_seen": 3058434048
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.402206619859579e-05,
      "loss": 1.3529,
      "theoretical_loss": 3.34431316446776,
      "tokens_seen": 3058499584
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.400200601805417e-05,
      "loss": 1.346,
      "theoretical_loss": 3.3443077231269993,
      "tokens_seen": 3058565120
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.398194583751254e-05,
      "loss": 1.3231,
      "theoretical_loss": 3.3443022819354744,
      "tokens_seen": 3058630656
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.396188565697092e-05,
      "loss": 1.0801,
      "theoretical_loss": 3.344296840893178,
      "tokens_seen": 3058696192
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.394182547642929e-05,
      "loss": 1.119,
      "theoretical_loss": 3.3442914000001025,
      "tokens_seen": 3058761728
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.392176529588767e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.34428595925624,
      "tokens_seen": 3058827264
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.390170511534604e-05,
      "loss": 1.2394,
      "theoretical_loss": 3.3442805186615843,
      "tokens_seen": 3058892800
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.38816449348044e-05,
      "loss": 1.1967,
      "theoretical_loss": 3.344275078216128,
      "tokens_seen": 3058958336
    },
    {
      "epoch": 0.85,
      "objective/train/docs_used": 1111330,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.197962999343872,
      "objective/train/theoretical_loss": 3.344269637919863,
      "objective/train/tokens_used": 1429482976,
      "theoretical_loss": 3.344269637919863,
      "tokens_seen": 3059023872
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.386158475426278e-05,
      "loss": 1.1617,
      "theoretical_loss": 3.344269637919863,
      "tokens_seen": 3059023872
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.384152457372116e-05,
      "loss": 1.2778,
      "theoretical_loss": 3.3442641977727825,
      "tokens_seen": 3059089408
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.382146439317953e-05,
      "loss": 1.277,
      "theoretical_loss": 3.3442587577748792,
      "tokens_seen": 3059154944
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.380140421263791e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.344253317926146,
      "tokens_seen": 3059220480
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.378134403209629e-05,
      "loss": 1.1773,
      "theoretical_loss": 3.344247878226575,
      "tokens_seen": 3059286016
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.376128385155466e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.34424243867616,
      "tokens_seen": 3059351552
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.374122367101305e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.3442369992748926,
      "tokens_seen": 3059417088
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.372116349047143e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.3442315600227657,
      "tokens_seen": 3059482624
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.37011033099298e-05,
      "loss": 1.1235,
      "theoretical_loss": 3.344226120919773,
      "tokens_seen": 3059548160
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.368104312938816e-05,
      "loss": 1.1502,
      "theoretical_loss": 3.344220681965906,
      "tokens_seen": 3059613696
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.366098294884654e-05,
      "loss": 1.2667,
      "theoretical_loss": 3.344215243161158,
      "tokens_seen": 3059679232
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.364092276830492e-05,
      "loss": 1.2644,
      "theoretical_loss": 3.3442098045055215,
      "tokens_seen": 3059744768
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.362086258776329e-05,
      "loss": 1.1519,
      "theoretical_loss": 3.3442043659989893,
      "tokens_seen": 3059810304
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.360080240722167e-05,
      "loss": 1.2404,
      "theoretical_loss": 3.3441989276415542,
      "tokens_seen": 3059875840
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.358074222668004e-05,
      "loss": 1.1967,
      "theoretical_loss": 3.3441934894332093,
      "tokens_seen": 3059941376
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.356068204613842e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.3441880513739464,
      "tokens_seen": 3060006912
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.35406218655968e-05,
      "loss": 1.1793,
      "theoretical_loss": 3.344182613463759,
      "tokens_seen": 3060072448
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.352056168505517e-05,
      "loss": 1.3346,
      "theoretical_loss": 3.3441771757026397,
      "tokens_seen": 3060137984
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.350050150451353e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.344171738090581,
      "tokens_seen": 3060203520
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.348044132397191e-05,
      "loss": 1.1588,
      "theoretical_loss": 3.3441663006275753,
      "tokens_seen": 3060269056
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.346038114343028e-05,
      "loss": 1.2061,
      "theoretical_loss": 3.344160863313616,
      "tokens_seen": 3060334592
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.344032096288866e-05,
      "loss": 1.2305,
      "theoretical_loss": 3.3441554261486957,
      "tokens_seen": 3060400128
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.342026078234704e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.3441499891328066,
      "tokens_seen": 3060465664
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.340020060180541e-05,
      "loss": 1.2271,
      "theoretical_loss": 3.344144552265942,
      "tokens_seen": 3060531200
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.33801404212638e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.3441391155480944,
      "tokens_seen": 3060596736
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.336008024072218e-05,
      "loss": 1.1939,
      "theoretical_loss": 3.344133678979256,
      "tokens_seen": 3060662272
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.334002006018055e-05,
      "loss": 1.2042,
      "theoretical_loss": 3.344128242559421,
      "tokens_seen": 3060727808
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.331995987963893e-05,
      "loss": 1.3652,
      "theoretical_loss": 3.344122806288581,
      "tokens_seen": 3060793344
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.329989969909729e-05,
      "loss": 1.2253,
      "theoretical_loss": 3.344117370166728,
      "tokens_seen": 3060858880
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.327983951855567e-05,
      "loss": 1.1215,
      "theoretical_loss": 3.3441119341938568,
      "tokens_seen": 3060924416
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.325977933801404e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.344106498369958,
      "tokens_seen": 3060989952
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.323971915747242e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.344101062695026,
      "tokens_seen": 3061055488
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.32196589769308e-05,
      "loss": 1.1661,
      "theoretical_loss": 3.3440956271690525,
      "tokens_seen": 3061121024
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.319959879638917e-05,
      "loss": 1.277,
      "theoretical_loss": 3.3440901917920307,
      "tokens_seen": 3061186560
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.317953861584755e-05,
      "loss": 1.1809,
      "theoretical_loss": 3.344084756563953,
      "tokens_seen": 3061252096
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.315947843530592e-05,
      "loss": 1.1162,
      "theoretical_loss": 3.3440793214848124,
      "tokens_seen": 3061317632
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.31394182547643e-05,
      "loss": 1.115,
      "theoretical_loss": 3.3440738865546016,
      "tokens_seen": 3061383168
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.311935807422267e-05,
      "loss": 1.3299,
      "theoretical_loss": 3.344068451773313,
      "tokens_seen": 3061448704
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.309929789368104e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.34406301714094,
      "tokens_seen": 3061514240
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.307923771313941e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.3440575826574745,
      "tokens_seen": 3061579776
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.305917753259779e-05,
      "loss": 1.2766,
      "theoretical_loss": 3.34405214832291,
      "tokens_seen": 3061645312
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.303911735205616e-05,
      "loss": 1.1243,
      "theoretical_loss": 3.3440467141372388,
      "tokens_seen": 3061710848
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.301905717151454e-05,
      "loss": 1.3593,
      "theoretical_loss": 3.3440412801004538,
      "tokens_seen": 3061776384
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.299899699097293e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.3440358462125475,
      "tokens_seen": 3061841920
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.29789368104313e-05,
      "loss": 1.1931,
      "theoretical_loss": 3.344030412473513,
      "tokens_seen": 3061907456
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.295887662988968e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.3440249788833425,
      "tokens_seen": 3061972992
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.293881644934806e-05,
      "loss": 1.3444,
      "theoretical_loss": 3.3440195454420296,
      "tokens_seen": 3062038528
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.291875626880642e-05,
      "loss": 1.2291,
      "theoretical_loss": 3.344014112149566,
      "tokens_seen": 3062104064
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.28986960882648e-05,
      "loss": 1.264,
      "theoretical_loss": 3.3440086790059453,
      "tokens_seen": 3062169600
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.287863590772317e-05,
      "loss": 1.2129,
      "theoretical_loss": 3.3440032460111597,
      "tokens_seen": 3062235136
    },
    {
      "epoch": 0.86,
      "objective/train/docs_used": 1112415,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.172864556312561,
      "objective/train/theoretical_loss": 3.343997813165202,
      "objective/train/tokens_used": 1432759776,
      "theoretical_loss": 3.343997813165202,
      "tokens_seen": 3062300672
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.285857572718154e-05,
      "loss": 1.1978,
      "theoretical_loss": 3.343997813165202,
      "tokens_seen": 3062300672
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.283851554663992e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.3439923804680656,
      "tokens_seen": 3062366208
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.28184553660983e-05,
      "loss": 1.335,
      "theoretical_loss": 3.343986947919742,
      "tokens_seen": 3062431744
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.279839518555667e-05,
      "loss": 1.2176,
      "theoretical_loss": 3.343981515520225,
      "tokens_seen": 3062497280
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.277833500501505e-05,
      "loss": 1.343,
      "theoretical_loss": 3.343976083269507,
      "tokens_seen": 3062562816
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.275827482447342e-05,
      "loss": 1.356,
      "theoretical_loss": 3.343970651167581,
      "tokens_seen": 3062628352
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.27382146439318e-05,
      "loss": 1.3856,
      "theoretical_loss": 3.343965219214439,
      "tokens_seen": 3062693888
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.271815446339016e-05,
      "loss": 1.246,
      "theoretical_loss": 3.343959787410074,
      "tokens_seen": 3062759424
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.269809428284854e-05,
      "loss": 1.2012,
      "theoretical_loss": 3.3439543557544793,
      "tokens_seen": 3062824960
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.267803410230691e-05,
      "loss": 1.2288,
      "theoretical_loss": 3.3439489242476474,
      "tokens_seen": 3062890496
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.265797392176529e-05,
      "loss": 1.282,
      "theoretical_loss": 3.343943492889571,
      "tokens_seen": 3062956032
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.263791374122367e-05,
      "loss": 1.2314,
      "theoretical_loss": 3.3439380616802423,
      "tokens_seen": 3063021568
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.261785356068205e-05,
      "loss": 1.1952,
      "theoretical_loss": 3.343932630619655,
      "tokens_seen": 3063087104
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.259779338014043e-05,
      "loss": 1.2614,
      "theoretical_loss": 3.343927199707801,
      "tokens_seen": 3063152640
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.25777331995988e-05,
      "loss": 1.2516,
      "theoretical_loss": 3.343921768944673,
      "tokens_seen": 3063218176
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.255767301905718e-05,
      "loss": 1.3118,
      "theoretical_loss": 3.343916338330265,
      "tokens_seen": 3063283712
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.253761283851556e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3439109078645686,
      "tokens_seen": 3063349248
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.251755265797392e-05,
      "loss": 1.4123,
      "theoretical_loss": 3.3439054775475765,
      "tokens_seen": 3063414784
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.24974924774323e-05,
      "loss": 1.1974,
      "theoretical_loss": 3.3439000473792824,
      "tokens_seen": 3063480320
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.247743229689067e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.3438946173596777,
      "tokens_seen": 3063545856
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.245737211634905e-05,
      "loss": 1.2119,
      "theoretical_loss": 3.3438891874887564,
      "tokens_seen": 3063611392
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.243731193580742e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.3438837577665104,
      "tokens_seen": 3063676928
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.24172517552658e-05,
      "loss": 1.2663,
      "theoretical_loss": 3.3438783281929334,
      "tokens_seen": 3063742464
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.239719157472417e-05,
      "loss": 1.2644,
      "theoretical_loss": 3.3438728987680166,
      "tokens_seen": 3063808000
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.237713139418255e-05,
      "loss": 1.2199,
      "theoretical_loss": 3.3438674694917543,
      "tokens_seen": 3063873536
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.235707121364093e-05,
      "loss": 1.2371,
      "theoretical_loss": 3.343862040364139,
      "tokens_seen": 3063939072
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.23370110330993e-05,
      "loss": 1.2474,
      "theoretical_loss": 3.343856611385162,
      "tokens_seen": 3064004608
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.231695085255766e-05,
      "loss": 1.3425,
      "theoretical_loss": 3.343851182554818,
      "tokens_seen": 3064070144
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.229689067201604e-05,
      "loss": 1.309,
      "theoretical_loss": 3.3438457538730986,
      "tokens_seen": 3064135680
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.227683049147442e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.343840325339997,
      "tokens_seen": 3064201216
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.22567703109328e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.3438348969555056,
      "tokens_seen": 3064266752
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.223671013039118e-05,
      "loss": 1.1515,
      "theoretical_loss": 3.3438294687196173,
      "tokens_seen": 3064332288
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.221664994984956e-05,
      "loss": 1.2391,
      "theoretical_loss": 3.3438240406323247,
      "tokens_seen": 3064397824
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.219658976930793e-05,
      "loss": 1.1839,
      "theoretical_loss": 3.343818612693621,
      "tokens_seen": 3064463360
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.217652958876631e-05,
      "loss": 1.1778,
      "theoretical_loss": 3.3438131849034987,
      "tokens_seen": 3064528896
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.215646940822468e-05,
      "loss": 1.2616,
      "theoretical_loss": 3.343807757261951,
      "tokens_seen": 3064594432
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.213640922768305e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.3438023297689696,
      "tokens_seen": 3064659968
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.211634904714142e-05,
      "loss": 1.1668,
      "theoretical_loss": 3.3437969024245477,
      "tokens_seen": 3064725504
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.20962888665998e-05,
      "loss": 1.252,
      "theoretical_loss": 3.343791475228679,
      "tokens_seen": 3064791040
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.207622868605817e-05,
      "loss": 1.184,
      "theoretical_loss": 3.3437860481813546,
      "tokens_seen": 3064856576
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.205616850551655e-05,
      "loss": 1.2375,
      "theoretical_loss": 3.3437806212825687,
      "tokens_seen": 3064922112
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.203610832497493e-05,
      "loss": 1.2314,
      "theoretical_loss": 3.3437751945323138,
      "tokens_seen": 3064987648
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.20160481444333e-05,
      "loss": 1.3149,
      "theoretical_loss": 3.343769767930582,
      "tokens_seen": 3065053184
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.199598796389168e-05,
      "loss": 1.256,
      "theoretical_loss": 3.343764341477366,
      "tokens_seen": 3065118720
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.197592778335005e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.3437589151726597,
      "tokens_seen": 3065184256
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.195586760280843e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.3437534890164544,
      "tokens_seen": 3065249792
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.193580742226679e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.343748063008744,
      "tokens_seen": 3065315328
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.191574724172517e-05,
      "loss": 1.1851,
      "theoretical_loss": 3.3437426371495214,
      "tokens_seen": 3065380864
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.189568706118354e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.343737211438778,
      "tokens_seen": 3065446400
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.187562688064193e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.343731785876508,
      "tokens_seen": 3065511936
    },
    {
      "epoch": 0.86,
      "objective/train/docs_used": 1113675,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.228407859802246,
      "objective/train/theoretical_loss": 3.343726360462703,
      "objective/train/tokens_used": 1436036576,
      "theoretical_loss": 3.343726360462703,
      "tokens_seen": 3065577472
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.185556670010031e-05,
      "loss": 1.2218,
      "theoretical_loss": 3.343726360462703,
      "tokens_seen": 3065577472
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.183550651955868e-05,
      "loss": 1.1769,
      "theoretical_loss": 3.3437209351973567,
      "tokens_seen": 3065643008
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.181544633901706e-05,
      "loss": 1.3058,
      "theoretical_loss": 3.3437155100804614,
      "tokens_seen": 3065708544
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.179538615847543e-05,
      "loss": 1.3262,
      "theoretical_loss": 3.3437100851120096,
      "tokens_seen": 3065774080
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.177532597793381e-05,
      "loss": 1.1986,
      "theoretical_loss": 3.343704660291995,
      "tokens_seen": 3065839616
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.175526579739219e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.3436992356204094,
      "tokens_seen": 3065905152
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.173520561685055e-05,
      "loss": 1.248,
      "theoretical_loss": 3.3436938110972463,
      "tokens_seen": 3065970688
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.171514543630892e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.343688386722498,
      "tokens_seen": 3066036224
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.16950852557673e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.343682962496157,
      "tokens_seen": 3066101760
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.167502507522568e-05,
      "loss": 1.1593,
      "theoretical_loss": 3.3436775384182162,
      "tokens_seen": 3066167296
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.165496489468405e-05,
      "loss": 1.2305,
      "theoretical_loss": 3.3436721144886694,
      "tokens_seen": 3066232832
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.163490471414243e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.3436666907075083,
      "tokens_seen": 3066298368
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.16148445336008e-05,
      "loss": 1.2365,
      "theoretical_loss": 3.343661267074726,
      "tokens_seen": 3066363904
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.159478435305918e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.343655843590315,
      "tokens_seen": 3066429440
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.157472417251756e-05,
      "loss": 1.1999,
      "theoretical_loss": 3.3436504202542685,
      "tokens_seen": 3066494976
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.155466399197592e-05,
      "loss": 1.1164,
      "theoretical_loss": 3.343644997066579,
      "tokens_seen": 3066560512
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.153460381143429e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.343639574027239,
      "tokens_seen": 3066626048
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.151454363089267e-05,
      "loss": 1.2105,
      "theoretical_loss": 3.343634151136242,
      "tokens_seen": 3066691584
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.149448345035106e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.3436287283935804,
      "tokens_seen": 3066757120
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.147442326980943e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.343623305799247,
      "tokens_seen": 3066822656
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.145436308926781e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.343617883353234,
      "tokens_seen": 3066888192
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.143430290872619e-05,
      "loss": 1.1695,
      "theoretical_loss": 3.3436124610555353,
      "tokens_seen": 3066953728
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.141424272818456e-05,
      "loss": 1.3045,
      "theoretical_loss": 3.3436070389061427,
      "tokens_seen": 3067019264
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.139418254764294e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.3436016169050493,
      "tokens_seen": 3067084800
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.137412236710131e-05,
      "loss": 1.2387,
      "theoretical_loss": 3.343596195052248,
      "tokens_seen": 3067150336
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.135406218655968e-05,
      "loss": 1.2782,
      "theoretical_loss": 3.3435907733477315,
      "tokens_seen": 3067215872
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.133400200601805e-05,
      "loss": 1.2218,
      "theoretical_loss": 3.343585351791493,
      "tokens_seen": 3067281408
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.131394182547643e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.343579930383524,
      "tokens_seen": 3067346944
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.12938816449348e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.343574509123819,
      "tokens_seen": 3067412480
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.127382146439318e-05,
      "loss": 1.221,
      "theoretical_loss": 3.3435690880123694,
      "tokens_seen": 3067478016
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.125376128385155e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.3435636670491684,
      "tokens_seen": 3067543552
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.123370110330993e-05,
      "loss": 1.1655,
      "theoretical_loss": 3.3435582462342093,
      "tokens_seen": 3067609088
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.12136409227683e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.3435528255674845,
      "tokens_seen": 3067674624
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.119358074222668e-05,
      "loss": 1.173,
      "theoretical_loss": 3.343547405048986,
      "tokens_seen": 3067740160
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.117352056168506e-05,
      "loss": 1.1512,
      "theoretical_loss": 3.3435419846787076,
      "tokens_seen": 3067805696
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.115346038114342e-05,
      "loss": 1.1919,
      "theoretical_loss": 3.343536564456642,
      "tokens_seen": 3067871232
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.113340020060181e-05,
      "loss": 1.198,
      "theoretical_loss": 3.3435311443827818,
      "tokens_seen": 3067936768
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.111334002006018e-05,
      "loss": 1.2253,
      "theoretical_loss": 3.3435257244571197,
      "tokens_seen": 3068002304
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.109327983951856e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.3435203046796484,
      "tokens_seen": 3068067840
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.107321965897694e-05,
      "loss": 1.2193,
      "theoretical_loss": 3.343514885050361,
      "tokens_seen": 3068133376
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.105315947843531e-05,
      "loss": 1.2216,
      "theoretical_loss": 3.34350946556925,
      "tokens_seen": 3068198912
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.103309929789369e-05,
      "loss": 1.2122,
      "theoretical_loss": 3.343504046236308,
      "tokens_seen": 3068264448
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.101303911735206e-05,
      "loss": 1.2154,
      "theoretical_loss": 3.343498627051528,
      "tokens_seen": 3068329984
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.099297893681044e-05,
      "loss": 1.3754,
      "theoretical_loss": 3.343493208014903,
      "tokens_seen": 3068395520
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.097291875626882e-05,
      "loss": 1.28,
      "theoretical_loss": 3.343487789126426,
      "tokens_seen": 3068461056
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.095285857572718e-05,
      "loss": 1.3348,
      "theoretical_loss": 3.3434823703860888,
      "tokens_seen": 3068526592
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.093279839518555e-05,
      "loss": 1.1923,
      "theoretical_loss": 3.3434769517938854,
      "tokens_seen": 3068592128
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.091273821464393e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.343471533349808,
      "tokens_seen": 3068657664
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.08926780341023e-05,
      "loss": 1.1897,
      "theoretical_loss": 3.343466115053849,
      "tokens_seen": 3068723200
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.087261785356068e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.3434606969060017,
      "tokens_seen": 3068788736
    },
    {
      "epoch": 0.86,
      "objective/train/docs_used": 1114913,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1459938287734985,
      "objective/train/theoretical_loss": 3.3434552789062586,
      "objective/train/tokens_used": 1439313376,
      "theoretical_loss": 3.3434552789062586,
      "tokens_seen": 3068854272
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.085255767301906e-05,
      "loss": 1.1436,
      "theoretical_loss": 3.3434552789062586,
      "tokens_seen": 3068854272
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.083249749247743e-05,
      "loss": 1.2042,
      "theoretical_loss": 3.3434498610546126,
      "tokens_seen": 3068919808
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.081243731193581e-05,
      "loss": 1.1617,
      "theoretical_loss": 3.343444443351057,
      "tokens_seen": 3068985344
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.079237713139418e-05,
      "loss": 1.2073,
      "theoretical_loss": 3.343439025795584,
      "tokens_seen": 3069050880
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.077231695085255e-05,
      "loss": 1.3145,
      "theoretical_loss": 3.343433608388186,
      "tokens_seen": 3069116416
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.075225677031094e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.343428191128857,
      "tokens_seen": 3069181952
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.073219658976931e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.3434227740175886,
      "tokens_seen": 3069247488
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.071213640922769e-05,
      "loss": 1.1155,
      "theoretical_loss": 3.343417357054374,
      "tokens_seen": 3069313024
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.069207622868606e-05,
      "loss": 1.3589,
      "theoretical_loss": 3.3434119402392066,
      "tokens_seen": 3069378560
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.067201604814444e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.343406523572078,
      "tokens_seen": 3069444096
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.065195586760281e-05,
      "loss": 1.2636,
      "theoretical_loss": 3.3434011070529825,
      "tokens_seen": 3069509632
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.063189568706119e-05,
      "loss": 1.1962,
      "theoretical_loss": 3.3433956906819113,
      "tokens_seen": 3069575168
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.061183550651957e-05,
      "loss": 1.2496,
      "theoretical_loss": 3.3433902744588586,
      "tokens_seen": 3069640704
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.059177532597794e-05,
      "loss": 1.2855,
      "theoretical_loss": 3.343384858383816,
      "tokens_seen": 3069706240
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.05717151454363e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.343379442456777,
      "tokens_seen": 3069771776
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.055165496489468e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.3433740266777345,
      "tokens_seen": 3069837312
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.053159478435306e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.3433686110466807,
      "tokens_seen": 3069902848
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.051153460381143e-05,
      "loss": 1.293,
      "theoretical_loss": 3.343363195563609,
      "tokens_seen": 3069968384
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.049147442326981e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.3433577802285117,
      "tokens_seen": 3070033920
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.047141424272818e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.343352365041382,
      "tokens_seen": 3070099456
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.045135406218656e-05,
      "loss": 1.244,
      "theoretical_loss": 3.3433469500022124,
      "tokens_seen": 3070164992
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.043129388164493e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.343341535110996,
      "tokens_seen": 3070230528
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.041123370110331e-05,
      "loss": 1.2672,
      "theoretical_loss": 3.343336120367725,
      "tokens_seen": 3070296064
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.039117352056169e-05,
      "loss": 1.2072,
      "theoretical_loss": 3.3433307057723933,
      "tokens_seen": 3070361600
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.037111334002006e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.3433252913249927,
      "tokens_seen": 3070427136
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.035105315947844e-05,
      "loss": 1.1981,
      "theoretical_loss": 3.343319877025516,
      "tokens_seen": 3070492672
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.033099297893681e-05,
      "loss": 1.3416,
      "theoretical_loss": 3.3433144628739564,
      "tokens_seen": 3070558208
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.031093279839519e-05,
      "loss": 1.2398,
      "theoretical_loss": 3.343309048870307,
      "tokens_seen": 3070623744
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.029087261785357e-05,
      "loss": 1.27,
      "theoretical_loss": 3.34330363501456,
      "tokens_seen": 3070689280
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.027081243731194e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.343298221306709,
      "tokens_seen": 3070754816
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.025075225677032e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.3432928077467454,
      "tokens_seen": 3070820352
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.023069207622869e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.3432873943346633,
      "tokens_seen": 3070885888
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.021063189568707e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.343281981070455,
      "tokens_seen": 3070951424
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.019057171514544e-05,
      "loss": 1.2769,
      "theoretical_loss": 3.3432765679541134,
      "tokens_seen": 3071016960
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.01705115346038e-05,
      "loss": 1.2028,
      "theoretical_loss": 3.343271154985631,
      "tokens_seen": 3071082496
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.015045135406218e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.343265742165001,
      "tokens_seen": 3071148032
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.013039117352056e-05,
      "loss": 1.2181,
      "theoretical_loss": 3.343260329492216,
      "tokens_seen": 3071213568
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.011033099297893e-05,
      "loss": 1.2297,
      "theoretical_loss": 3.343254916967269,
      "tokens_seen": 3071279104
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.009027081243731e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.343249504590153,
      "tokens_seen": 3071344640
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.007021063189569e-05,
      "loss": 1.2028,
      "theoretical_loss": 3.34324409236086,
      "tokens_seen": 3071410176
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.005015045135406e-05,
      "loss": 1.2566,
      "theoretical_loss": 3.3432386802793834,
      "tokens_seen": 3071475712
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.003009027081244e-05,
      "loss": 1.3447,
      "theoretical_loss": 3.343233268345716,
      "tokens_seen": 3071541248
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.001003009027081e-05,
      "loss": 1.1817,
      "theoretical_loss": 3.3432278565598508,
      "tokens_seen": 3071606784
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.998996990972919e-05,
      "loss": 1.3316,
      "theoretical_loss": 3.34322244492178,
      "tokens_seen": 3071672320
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.996990972918756e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.343217033431497,
      "tokens_seen": 3071737856
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.994984954864594e-05,
      "loss": 1.2085,
      "theoretical_loss": 3.343211622088994,
      "tokens_seen": 3071803392
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.992978936810432e-05,
      "loss": 1.3125,
      "theoretical_loss": 3.343206210894264,
      "tokens_seen": 3071868928
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.990972918756269e-05,
      "loss": 1.3013,
      "theoretical_loss": 3.3432007998473003,
      "tokens_seen": 3071934464
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.988966900702107e-05,
      "loss": 1.319,
      "theoretical_loss": 3.3431953889480956,
      "tokens_seen": 3072000000
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.986960882647944e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.343189978196642,
      "tokens_seen": 3072065536
    },
    {
      "epoch": 0.86,
      "objective/train/docs_used": 1116127,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2608907222747803,
      "objective/train/theoretical_loss": 3.3431845675929335,
      "objective/train/tokens_used": 1442590176,
      "theoretical_loss": 3.3431845675929335,
      "tokens_seen": 3072131072
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.984954864593782e-05,
      "loss": 1.205,
      "theoretical_loss": 3.3431845675929335,
      "tokens_seen": 3072131072
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.98294884653962e-05,
      "loss": 1.2421,
      "theoretical_loss": 3.3431791571369613,
      "tokens_seen": 3072196608
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.980942828485457e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.34317374682872,
      "tokens_seen": 3072262144
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.978936810431293e-05,
      "loss": 1.3149,
      "theoretical_loss": 3.3431683366682012,
      "tokens_seen": 3072327680
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.976930792377131e-05,
      "loss": 1.24,
      "theoretical_loss": 3.343162926655398,
      "tokens_seen": 3072393216
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.974924774322968e-05,
      "loss": 1.2194,
      "theoretical_loss": 3.3431575167903036,
      "tokens_seen": 3072458752
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.972918756268806e-05,
      "loss": 1.2076,
      "theoretical_loss": 3.3431521070729104,
      "tokens_seen": 3072524288
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.970912738214644e-05,
      "loss": 1.2361,
      "theoretical_loss": 3.343146697503211,
      "tokens_seen": 3072589824
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.968906720160481e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.343141288081199,
      "tokens_seen": 3072655360
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.966900702106319e-05,
      "loss": 1.2237,
      "theoretical_loss": 3.3431358788068666,
      "tokens_seen": 3072720896
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.964894684052156e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.3431304696802067,
      "tokens_seen": 3072786432
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.962888665997995e-05,
      "loss": 1.309,
      "theoretical_loss": 3.343125060701212,
      "tokens_seen": 3072851968
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.960882647943833e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.3431196518698756,
      "tokens_seen": 3072917504
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.958876629889669e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.3431142431861907,
      "tokens_seen": 3072983040
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.956870611835507e-05,
      "loss": 1.1674,
      "theoretical_loss": 3.3431088346501494,
      "tokens_seen": 3073048576
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.954864593781344e-05,
      "loss": 1.2051,
      "theoretical_loss": 3.3431034262617443,
      "tokens_seen": 3073114112
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.952858575727182e-05,
      "loss": 1.1876,
      "theoretical_loss": 3.3430980180209695,
      "tokens_seen": 3073179648
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.95085255767302e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.3430926099278166,
      "tokens_seen": 3073245184
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.948846539618857e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.343087201982279,
      "tokens_seen": 3073310720
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.946840521564695e-05,
      "loss": 1.1802,
      "theoretical_loss": 3.343081794184349,
      "tokens_seen": 3073376256
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.944834503510532e-05,
      "loss": 1.1966,
      "theoretical_loss": 3.3430763865340203,
      "tokens_seen": 3073441792
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.94282848545637e-05,
      "loss": 1.3509,
      "theoretical_loss": 3.343070979031285,
      "tokens_seen": 3073507328
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.940822467402207e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.343065571676136,
      "tokens_seen": 3073572864
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.938816449348044e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.3430601644685662,
      "tokens_seen": 3073638400
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.936810431293881e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.343054757408569,
      "tokens_seen": 3073703936
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.934804413239719e-05,
      "loss": 1.1616,
      "theoretical_loss": 3.3430493504961363,
      "tokens_seen": 3073769472
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.932798395185556e-05,
      "loss": 1.1886,
      "theoretical_loss": 3.343043943731262,
      "tokens_seen": 3073835008
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.930792377131394e-05,
      "loss": 1.256,
      "theoretical_loss": 3.3430385371139373,
      "tokens_seen": 3073900544
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.928786359077231e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.3430331306441565,
      "tokens_seen": 3073966080
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.926780341023069e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.343027724321912,
      "tokens_seen": 3074031616
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.924774322968908e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.343022318147196,
      "tokens_seen": 3074097152
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.922768304914746e-05,
      "loss": 1.2256,
      "theoretical_loss": 3.3430169121200026,
      "tokens_seen": 3074162688
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.920762286860582e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.3430115062403236,
      "tokens_seen": 3074228224
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.918756268806419e-05,
      "loss": 1.2345,
      "theoretical_loss": 3.3430061005081524,
      "tokens_seen": 3074293760
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.916750250752257e-05,
      "loss": 1.1778,
      "theoretical_loss": 3.3430006949234814,
      "tokens_seen": 3074359296
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.914744232698094e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.3429952894863035,
      "tokens_seen": 3074424832
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.912738214643932e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.3429898841966117,
      "tokens_seen": 3074490368
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.91073219658977e-05,
      "loss": 1.1998,
      "theoretical_loss": 3.3429844790543988,
      "tokens_seen": 3074555904
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.908726178535607e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.3429790740596577,
      "tokens_seen": 3074621440
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.906720160481445e-05,
      "loss": 1.312,
      "theoretical_loss": 3.342973669212381,
      "tokens_seen": 3074686976
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.904714142427282e-05,
      "loss": 1.1901,
      "theoretical_loss": 3.3429682645125616,
      "tokens_seen": 3074752512
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.90270812437312e-05,
      "loss": 1.2371,
      "theoretical_loss": 3.3429628599601924,
      "tokens_seen": 3074818048
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.900702106318956e-05,
      "loss": 1.0938,
      "theoretical_loss": 3.3429574555552666,
      "tokens_seen": 3074883584
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.898696088264794e-05,
      "loss": 1.2901,
      "theoretical_loss": 3.342952051297776,
      "tokens_seen": 3074949120
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.896690070210631e-05,
      "loss": 1.1924,
      "theoretical_loss": 3.342946647187715,
      "tokens_seen": 3075014656
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.894684052156469e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.342941243225075,
      "tokens_seen": 3075080192
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.892678034102306e-05,
      "loss": 1.2962,
      "theoretical_loss": 3.3429358394098494,
      "tokens_seen": 3075145728
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.890672016048144e-05,
      "loss": 1.3097,
      "theoretical_loss": 3.3429304357420313,
      "tokens_seen": 3075211264
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.888665997993982e-05,
      "loss": 1.236,
      "theoretical_loss": 3.342925032221613,
      "tokens_seen": 3075276800
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.88665997993982e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.3429196288485876,
      "tokens_seen": 3075342336
    },
    {
      "epoch": 0.86,
      "objective/train/docs_used": 1117177,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3147614002227783,
      "objective/train/theoretical_loss": 3.3429142256229483,
      "objective/train/tokens_used": 1445866976,
      "theoretical_loss": 3.3429142256229483,
      "tokens_seen": 3075407872
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.884653961885658e-05,
      "loss": 1.277,
      "theoretical_loss": 3.3429142256229483,
      "tokens_seen": 3075407872
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.882647943831496e-05,
      "loss": 1.2309,
      "theoretical_loss": 3.342908822544687,
      "tokens_seen": 3075473408
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.880641925777332e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.3429034196137977,
      "tokens_seen": 3075538944
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.87863590772317e-05,
      "loss": 1.3641,
      "theoretical_loss": 3.342898016830272,
      "tokens_seen": 3075604480
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.876629889669007e-05,
      "loss": 1.2228,
      "theoretical_loss": 3.342892614194104,
      "tokens_seen": 3075670016
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.874623871614845e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.3428872117052855,
      "tokens_seen": 3075735552
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.872617853560682e-05,
      "loss": 1.1883,
      "theoretical_loss": 3.34288180936381,
      "tokens_seen": 3075801088
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.87061183550652e-05,
      "loss": 1.1291,
      "theoretical_loss": 3.3428764071696704,
      "tokens_seen": 3075866624
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.868605817452357e-05,
      "loss": 1.215,
      "theoretical_loss": 3.3428710051228587,
      "tokens_seen": 3075932160
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.866599799398195e-05,
      "loss": 1.1792,
      "theoretical_loss": 3.342865603223369,
      "tokens_seen": 3075997696
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.864593781344033e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.3428602014711926,
      "tokens_seen": 3076063232
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.862587763289869e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.342854799866324,
      "tokens_seen": 3076128768
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.860581745235706e-05,
      "loss": 1.2506,
      "theoretical_loss": 3.3428493984087546,
      "tokens_seen": 3076194304
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.858575727181544e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.3428439970984782,
      "tokens_seen": 3076259840
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.856569709127382e-05,
      "loss": 1.3367,
      "theoretical_loss": 3.342838595935487,
      "tokens_seen": 3076325376
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.854563691073219e-05,
      "loss": 1.2501,
      "theoretical_loss": 3.3428331949197747,
      "tokens_seen": 3076390912
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.852557673019057e-05,
      "loss": 1.1985,
      "theoretical_loss": 3.3428277940513333,
      "tokens_seen": 3076456448
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.850551654964896e-05,
      "loss": 1.297,
      "theoretical_loss": 3.342822393330156,
      "tokens_seen": 3076521984
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.848545636910733e-05,
      "loss": 1.2036,
      "theoretical_loss": 3.342816992756236,
      "tokens_seen": 3076587520
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.846539618856571e-05,
      "loss": 1.1435,
      "theoretical_loss": 3.342811592329565,
      "tokens_seen": 3076653056
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.844533600802408e-05,
      "loss": 1.1949,
      "theoretical_loss": 3.3428061920501375,
      "tokens_seen": 3076718592
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.842527582748245e-05,
      "loss": 1.2049,
      "theoretical_loss": 3.342800791917945,
      "tokens_seen": 3076784128
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.840521564694082e-05,
      "loss": 1.1968,
      "theoretical_loss": 3.342795391932981,
      "tokens_seen": 3076849664
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.83851554663992e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.3427899920952377,
      "tokens_seen": 3076915200
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.836509528585757e-05,
      "loss": 1.2301,
      "theoretical_loss": 3.342784592404709,
      "tokens_seen": 3076980736
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.834503510531595e-05,
      "loss": 1.2297,
      "theoretical_loss": 3.3427791928613866,
      "tokens_seen": 3077046272
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.832497492477433e-05,
      "loss": 1.2204,
      "theoretical_loss": 3.3427737934652644,
      "tokens_seen": 3077111808
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.83049147442327e-05,
      "loss": 1.2332,
      "theoretical_loss": 3.3427683942163346,
      "tokens_seen": 3077177344
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.828485456369108e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.3427629951145903,
      "tokens_seen": 3077242880
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.826479438314945e-05,
      "loss": 1.2319,
      "theoretical_loss": 3.342757596160024,
      "tokens_seen": 3077308416
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.824473420260783e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.3427521973526293,
      "tokens_seen": 3077373952
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.822467402206619e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.3427467986923984,
      "tokens_seen": 3077439488
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.820461384152457e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.342741400179324,
      "tokens_seen": 3077505024
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.818455366098294e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.3427360018133996,
      "tokens_seen": 3077570560
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.816449348044132e-05,
      "loss": 1.164,
      "theoretical_loss": 3.342730603594618,
      "tokens_seen": 3077636096
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.81444332998997e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.3427252055229713,
      "tokens_seen": 3077701632
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.812437311935808e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.342719807598453,
      "tokens_seen": 3077767168
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.810431293881646e-05,
      "loss": 1.2387,
      "theoretical_loss": 3.342714409821056,
      "tokens_seen": 3077832704
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.808425275827483e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.342709012190773,
      "tokens_seen": 3077898240
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.806419257773321e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.3427036147075966,
      "tokens_seen": 3077963776
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.804413239719159e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.3426982173715203,
      "tokens_seen": 3078029312
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.802407221664995e-05,
      "loss": 1.2258,
      "theoretical_loss": 3.342692820182536,
      "tokens_seen": 3078094848
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.800401203610832e-05,
      "loss": 1.2202,
      "theoretical_loss": 3.3426874231406374,
      "tokens_seen": 3078160384
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.79839518555667e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.342682026245817,
      "tokens_seen": 3078225920
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.796389167502508e-05,
      "loss": 1.2353,
      "theoretical_loss": 3.342676629498068,
      "tokens_seen": 3078291456
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.794383149448345e-05,
      "loss": 1.2144,
      "theoretical_loss": 3.3426712328973824,
      "tokens_seen": 3078356992
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.792377131394183e-05,
      "loss": 1.1927,
      "theoretical_loss": 3.3426658364437545,
      "tokens_seen": 3078422528
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.79037111334002e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.3426604401371756,
      "tokens_seen": 3078488064
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.788365095285858e-05,
      "loss": 1.2723,
      "theoretical_loss": 3.34265504397764,
      "tokens_seen": 3078553600
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.786359077231695e-05,
      "loss": 1.272,
      "theoretical_loss": 3.342649647965139,
      "tokens_seen": 3078619136
    },
    {
      "epoch": 0.87,
      "objective/train/docs_used": 1118389,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3177624940872192,
      "objective/train/theoretical_loss": 3.342644252099667,
      "objective/train/tokens_used": 1449143776,
      "theoretical_loss": 3.342644252099667,
      "tokens_seen": 3078684672
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.784353059177532e-05,
      "loss": 1.2908,
      "theoretical_loss": 3.342644252099667,
      "tokens_seen": 3078684672
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.782347041123369e-05,
      "loss": 1.2306,
      "theoretical_loss": 3.3426388563812157,
      "tokens_seen": 3078750208
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.780341023069207e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.3426334608097785,
      "tokens_seen": 3078815744
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.778335005015044e-05,
      "loss": 1.2074,
      "theoretical_loss": 3.342628065385348,
      "tokens_seen": 3078881280
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.776328986960882e-05,
      "loss": 1.198,
      "theoretical_loss": 3.342622670107918,
      "tokens_seen": 3078946816
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.774322968906721e-05,
      "loss": 1.1445,
      "theoretical_loss": 3.34261727497748,
      "tokens_seen": 3079012352
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.772316950852559e-05,
      "loss": 1.1809,
      "theoretical_loss": 3.342611879994028,
      "tokens_seen": 3079077888
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.770310932798396e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.342606485157554,
      "tokens_seen": 3079143424
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.768304914744234e-05,
      "loss": 1.266,
      "theoretical_loss": 3.3426010904680514,
      "tokens_seen": 3079208960
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.766298896690071e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.342595695925513,
      "tokens_seen": 3079274496
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.764292878635908e-05,
      "loss": 1.2961,
      "theoretical_loss": 3.342590301529931,
      "tokens_seen": 3079340032
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.762286860581745e-05,
      "loss": 1.227,
      "theoretical_loss": 3.3425849072812994,
      "tokens_seen": 3079405568
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.760280842527583e-05,
      "loss": 1.2108,
      "theoretical_loss": 3.3425795131796106,
      "tokens_seen": 3079471104
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.75827482447342e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.342574119224857,
      "tokens_seen": 3079536640
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.756268806419258e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.342568725417032,
      "tokens_seen": 3079602176
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.754262788365095e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.3425633317561285,
      "tokens_seen": 3079667712
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.752256770310933e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.3425579382421393,
      "tokens_seen": 3079733248
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.75025075225677e-05,
      "loss": 1.2418,
      "theoretical_loss": 3.3425525448750566,
      "tokens_seen": 3079798784
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.748244734202608e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.342547151654874,
      "tokens_seen": 3079864320
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.746238716148446e-05,
      "loss": 1.267,
      "theoretical_loss": 3.342541758581585,
      "tokens_seen": 3079929856
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.744232698094282e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.3425363656551808,
      "tokens_seen": 3079995392
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.74222668004012e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.342530972875655,
      "tokens_seen": 3080060928
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.740220661985957e-05,
      "loss": 1.2227,
      "theoretical_loss": 3.3425255802430014,
      "tokens_seen": 3080126464
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.738214643931795e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.342520187757212,
      "tokens_seen": 3080192000
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.736208625877634e-05,
      "loss": 1.3179,
      "theoretical_loss": 3.3425147954182792,
      "tokens_seen": 3080257536
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.734202607823471e-05,
      "loss": 1.1473,
      "theoretical_loss": 3.342509403226197,
      "tokens_seen": 3080323072
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.732196589769309e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.3425040111809574,
      "tokens_seen": 3080388608
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.730190571715146e-05,
      "loss": 1.2664,
      "theoretical_loss": 3.342498619282554,
      "tokens_seen": 3080454144
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.728184553660984e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.342493227530979,
      "tokens_seen": 3080519680
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.726178535606822e-05,
      "loss": 1.2477,
      "theoretical_loss": 3.3424878359262253,
      "tokens_seen": 3080585216
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.724172517552658e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.3424824444682866,
      "tokens_seen": 3080650752
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.722166499498495e-05,
      "loss": 1.1908,
      "theoretical_loss": 3.342477053157155,
      "tokens_seen": 3080716288
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.720160481444333e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.342471661992824,
      "tokens_seen": 3080781824
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.71815446339017e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.3424662709752857,
      "tokens_seen": 3080847360
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.716148445336008e-05,
      "loss": 1.2182,
      "theoretical_loss": 3.3424608801045332,
      "tokens_seen": 3080912896
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.714142427281846e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.34245548938056,
      "tokens_seen": 3080978432
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.712136409227683e-05,
      "loss": 1.26,
      "theoretical_loss": 3.342450098803358,
      "tokens_seen": 3081043968
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.710130391173521e-05,
      "loss": 1.2625,
      "theoretical_loss": 3.3424447083729207,
      "tokens_seen": 3081109504
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.708124373119358e-05,
      "loss": 1.3398,
      "theoretical_loss": 3.3424393180892413,
      "tokens_seen": 3081175040
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.706118355065195e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.342433927952312,
      "tokens_seen": 3081240576
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.704112337011032e-05,
      "loss": 1.265,
      "theoretical_loss": 3.342428537962126,
      "tokens_seen": 3081306112
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.70210631895687e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.3424231481186766,
      "tokens_seen": 3081371648
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.700100300902709e-05,
      "loss": 1.339,
      "theoretical_loss": 3.3424177584219557,
      "tokens_seen": 3081437184
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.698094282848546e-05,
      "loss": 1.2229,
      "theoretical_loss": 3.3424123688719565,
      "tokens_seen": 3081502720
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.696088264794384e-05,
      "loss": 1.288,
      "theoretical_loss": 3.342406979468673,
      "tokens_seen": 3081568256
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.694082246740221e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.3424015902120963,
      "tokens_seen": 3081633792
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.692076228686059e-05,
      "loss": 1.3918,
      "theoretical_loss": 3.3423962011022206,
      "tokens_seen": 3081699328
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.690070210631897e-05,
      "loss": 1.2622,
      "theoretical_loss": 3.342390812139038,
      "tokens_seen": 3081764864
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.688064192577734e-05,
      "loss": 1.1636,
      "theoretical_loss": 3.342385423322542,
      "tokens_seen": 3081830400
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.68605817452357e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.3423800346527255,
      "tokens_seen": 3081895936
    },
    {
      "epoch": 0.87,
      "objective/train/docs_used": 1119488,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2091344594955444,
      "objective/train/theoretical_loss": 3.342374646129581,
      "objective/train/tokens_used": 1452420576,
      "theoretical_loss": 3.342374646129581,
      "tokens_seen": 3081961472
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.684052156469408e-05,
      "loss": 1.1885,
      "theoretical_loss": 3.342374646129581,
      "tokens_seen": 3081961472
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.682046138415246e-05,
      "loss": 1.1314,
      "theoretical_loss": 3.3423692577531012,
      "tokens_seen": 3082027008
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.680040120361083e-05,
      "loss": 1.1791,
      "theoretical_loss": 3.34236386952328,
      "tokens_seen": 3082092544
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.678034102306921e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.3423584814401086,
      "tokens_seen": 3082158080
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.676028084252758e-05,
      "loss": 1.1877,
      "theoretical_loss": 3.3423530935035815,
      "tokens_seen": 3082223616
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.674022066198596e-05,
      "loss": 1.1862,
      "theoretical_loss": 3.342347705713691,
      "tokens_seen": 3082289152
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.672016048144433e-05,
      "loss": 1.2283,
      "theoretical_loss": 3.34234231807043,
      "tokens_seen": 3082354688
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.670010030090271e-05,
      "loss": 1.2037,
      "theoretical_loss": 3.342336930573791,
      "tokens_seen": 3082420224
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.668004012036109e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.3423315432237675,
      "tokens_seen": 3082485760
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.665997993981945e-05,
      "loss": 1.3585,
      "theoretical_loss": 3.342326156020352,
      "tokens_seen": 3082551296
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.663991975927782e-05,
      "loss": 1.2095,
      "theoretical_loss": 3.342320768963538,
      "tokens_seen": 3082616832
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.661985957873621e-05,
      "loss": 1.2005,
      "theoretical_loss": 3.342315382053317,
      "tokens_seen": 3082682368
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.659979939819459e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.3423099952896838,
      "tokens_seen": 3082747904
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.657973921765296e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.34230460867263,
      "tokens_seen": 3082813440
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.655967903711134e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.342299222202149,
      "tokens_seen": 3082878976
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.653961885656972e-05,
      "loss": 1.1524,
      "theoretical_loss": 3.342293835878233,
      "tokens_seen": 3082944512
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.651955867602809e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.3422884497008756,
      "tokens_seen": 3083010048
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.649949849548647e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.3422830636700698,
      "tokens_seen": 3083075584
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.647943831494484e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.342277677785808,
      "tokens_seen": 3083141120
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.64593781344032e-05,
      "loss": 1.2377,
      "theoretical_loss": 3.3422722920480834,
      "tokens_seen": 3083206656
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.643931795386158e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.3422669064568886,
      "tokens_seen": 3083272192
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.641925777331996e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.342261521012217,
      "tokens_seen": 3083337728
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.639919759277833e-05,
      "loss": 1.2293,
      "theoretical_loss": 3.3422561357140608,
      "tokens_seen": 3083403264
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.637913741223671e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.3422507505624135,
      "tokens_seen": 3083468800
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.635907723169509e-05,
      "loss": 1.2392,
      "theoretical_loss": 3.342245365557268,
      "tokens_seen": 3083534336
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.633901705115346e-05,
      "loss": 1.2362,
      "theoretical_loss": 3.3422399806986167,
      "tokens_seen": 3083599872
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.631895687061184e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.342234595986453,
      "tokens_seen": 3083665408
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.629889669007021e-05,
      "loss": 1.262,
      "theoretical_loss": 3.3422292114207695,
      "tokens_seen": 3083730944
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.627883650952857e-05,
      "loss": 1.1807,
      "theoretical_loss": 3.3422238270015594,
      "tokens_seen": 3083796480
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.625877632898695e-05,
      "loss": 1.1751,
      "theoretical_loss": 3.3422184427288153,
      "tokens_seen": 3083862016
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.623871614844534e-05,
      "loss": 1.2099,
      "theoretical_loss": 3.34221305860253,
      "tokens_seen": 3083927552
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.621865596790372e-05,
      "loss": 1.3728,
      "theoretical_loss": 3.3422076746226965,
      "tokens_seen": 3083993088
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.619859578736209e-05,
      "loss": 1.1834,
      "theoretical_loss": 3.3422022907893085,
      "tokens_seen": 3084058624
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.617853560682047e-05,
      "loss": 1.1474,
      "theoretical_loss": 3.3421969071023576,
      "tokens_seen": 3084124160
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.615847542627884e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.3421915235618376,
      "tokens_seen": 3084189696
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.613841524573722e-05,
      "loss": 1.2128,
      "theoretical_loss": 3.3421861401677413,
      "tokens_seen": 3084255232
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.61183550651956e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.3421807569200612,
      "tokens_seen": 3084320768
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.609829488465397e-05,
      "loss": 1.283,
      "theoretical_loss": 3.3421753738187903,
      "tokens_seen": 3084386304
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.607823470411233e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.342169990863922,
      "tokens_seen": 3084451840
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.605817452357071e-05,
      "loss": 1.1953,
      "theoretical_loss": 3.342164608055449,
      "tokens_seen": 3084517376
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.603811434302908e-05,
      "loss": 1.1716,
      "theoretical_loss": 3.3421592253933636,
      "tokens_seen": 3084582912
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.601805416248746e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.3421538428776594,
      "tokens_seen": 3084648448
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.599799398194584e-05,
      "loss": 1.2203,
      "theoretical_loss": 3.342148460508329,
      "tokens_seen": 3084713984
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.597793380140421e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.342143078285366,
      "tokens_seen": 3084779520
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.595787362086259e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.342137696208762,
      "tokens_seen": 3084845056
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.593781344032096e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.3421323142785107,
      "tokens_seen": 3084910592
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.591775325977934e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.342126932494605,
      "tokens_seen": 3084976128
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.589769307923771e-05,
      "loss": 1.2645,
      "theoretical_loss": 3.342121550857038,
      "tokens_seen": 3085041664
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.587763289869609e-05,
      "loss": 1.278,
      "theoretical_loss": 3.342116169365802,
      "tokens_seen": 3085107200
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.585757271815447e-05,
      "loss": 1.1933,
      "theoretical_loss": 3.3421107880208902,
      "tokens_seen": 3085172736
    },
    {
      "epoch": 0.87,
      "objective/train/docs_used": 1120785,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3070566654205322,
      "objective/train/theoretical_loss": 3.342105406822296,
      "objective/train/tokens_used": 1455697376,
      "theoretical_loss": 3.342105406822296,
      "tokens_seen": 3085238272
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.583751253761284e-05,
      "loss": 1.3057,
      "theoretical_loss": 3.342105406822296,
      "tokens_seen": 3085238272
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.581745235707122e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.3421000257700118,
      "tokens_seen": 3085303808
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.57973921765296e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.3420946448640305,
      "tokens_seen": 3085369344
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.577733199598797e-05,
      "loss": 1.1958,
      "theoretical_loss": 3.342089264104345,
      "tokens_seen": 3085434880
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.575727181544635e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.3420838834909485,
      "tokens_seen": 3085500416
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.573721163490472e-05,
      "loss": 1.4305,
      "theoretical_loss": 3.342078503023834,
      "tokens_seen": 3085565952
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.57171514543631e-05,
      "loss": 1.1885,
      "theoretical_loss": 3.342073122702994,
      "tokens_seen": 3085631488
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.569709127382147e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.3420677425284215,
      "tokens_seen": 3085697024
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.567703109327984e-05,
      "loss": 1.2065,
      "theoretical_loss": 3.3420623625001094,
      "tokens_seen": 3085762560
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.565697091273821e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.342056982618051,
      "tokens_seen": 3085828096
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.563691073219659e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.342051602882239,
      "tokens_seen": 3085893632
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.561685055165496e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.342046223292666,
      "tokens_seen": 3085959168
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.559679037111334e-05,
      "loss": 1.1932,
      "theoretical_loss": 3.3420408438493254,
      "tokens_seen": 3086024704
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.557673019057171e-05,
      "loss": 1.176,
      "theoretical_loss": 3.3420354645522097,
      "tokens_seen": 3086090240
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.555667001003009e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.342030085401312,
      "tokens_seen": 3086155776
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.553660982948847e-05,
      "loss": 1.218,
      "theoretical_loss": 3.3420247063966255,
      "tokens_seen": 3086221312
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.551654964894684e-05,
      "loss": 1.2161,
      "theoretical_loss": 3.3420193275381425,
      "tokens_seen": 3086286848
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.549648946840522e-05,
      "loss": 1.2674,
      "theoretical_loss": 3.3420139488258567,
      "tokens_seen": 3086352384
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.547642928786359e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.34200857025976,
      "tokens_seen": 3086417920
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.545636910732197e-05,
      "loss": 1.2085,
      "theoretical_loss": 3.3420031918398467,
      "tokens_seen": 3086483456
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.543630892678034e-05,
      "loss": 1.1755,
      "theoretical_loss": 3.3419978135661084,
      "tokens_seen": 3086548992
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.541624874623872e-05,
      "loss": 1.2001,
      "theoretical_loss": 3.341992435438539,
      "tokens_seen": 3086614528
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.53961885656971e-05,
      "loss": 1.1976,
      "theoretical_loss": 3.3419870574571306,
      "tokens_seen": 3086680064
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.537612838515547e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.3419816796218766,
      "tokens_seen": 3086745600
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.535606820461385e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.34197630193277,
      "tokens_seen": 3086811136
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.533600802407222e-05,
      "loss": 1.1794,
      "theoretical_loss": 3.3419709243898037,
      "tokens_seen": 3086876672
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.53159478435306e-05,
      "loss": 1.2965,
      "theoretical_loss": 3.34196554699297,
      "tokens_seen": 3086942208
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.529588766298896e-05,
      "loss": 1.1408,
      "theoretical_loss": 3.341960169742263,
      "tokens_seen": 3087007744
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.527582748244734e-05,
      "loss": 1.193,
      "theoretical_loss": 3.341954792637675,
      "tokens_seen": 3087073280
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.525576730190571e-05,
      "loss": 1.2064,
      "theoretical_loss": 3.3419494156791982,
      "tokens_seen": 3087138816
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.523570712136409e-05,
      "loss": 1.0993,
      "theoretical_loss": 3.3419440388668264,
      "tokens_seen": 3087204352
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.521564694082246e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.3419386622005525,
      "tokens_seen": 3087269888
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.519558676028084e-05,
      "loss": 1.2574,
      "theoretical_loss": 3.341933285680369,
      "tokens_seen": 3087335424
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.517552657973922e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.3419279093062695,
      "tokens_seen": 3087400960
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.515546639919759e-05,
      "loss": 1.2585,
      "theoretical_loss": 3.3419225330782467,
      "tokens_seen": 3087466496
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.513540621865597e-05,
      "loss": 1.2379,
      "theoretical_loss": 3.341917156996293,
      "tokens_seen": 3087532032
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.511534603811436e-05,
      "loss": 1.1762,
      "theoretical_loss": 3.3419117810604018,
      "tokens_seen": 3087597568
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.509528585757272e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.3419064052705654,
      "tokens_seen": 3087663104
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.50752256770311e-05,
      "loss": 1.2362,
      "theoretical_loss": 3.341901029626778,
      "tokens_seen": 3087728640
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.505516549648947e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.3418956541290314,
      "tokens_seen": 3087794176
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.503510531594785e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.341890278777319,
      "tokens_seen": 3087859712
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.501504513540622e-05,
      "loss": 1.1844,
      "theoretical_loss": 3.341884903571634,
      "tokens_seen": 3087925248
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.49949849548646e-05,
      "loss": 1.1943,
      "theoretical_loss": 3.3418795285119685,
      "tokens_seen": 3087990784
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.497492477432297e-05,
      "loss": 1.22,
      "theoretical_loss": 3.341874153598316,
      "tokens_seen": 3088056320
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.495486459378135e-05,
      "loss": 1.2247,
      "theoretical_loss": 3.341868778830669,
      "tokens_seen": 3088121856
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.493480441323973e-05,
      "loss": 1.1761,
      "theoretical_loss": 3.3418634042090214,
      "tokens_seen": 3088187392
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.491474423269809e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.3418580297333653,
      "tokens_seen": 3088252928
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.489468405215646e-05,
      "loss": 1.2859,
      "theoretical_loss": 3.3418526554036942,
      "tokens_seen": 3088318464
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.487462387161484e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.34184728122,
      "tokens_seen": 3088384000
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.485456369107322e-05,
      "loss": 1.1519,
      "theoretical_loss": 3.3418419071822774,
      "tokens_seen": 3088449536
    },
    {
      "epoch": 0.87,
      "objective/train/docs_used": 1122042,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2953546047210693,
      "objective/train/theoretical_loss": 3.3418365332905173,
      "objective/train/tokens_used": 1458974176,
      "theoretical_loss": 3.3418365332905173,
      "tokens_seen": 3088515072
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.483450351053159e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.3418365332905173,
      "tokens_seen": 3088515072
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.481444332998997e-05,
      "loss": 1.1937,
      "theoretical_loss": 3.3418311595447143,
      "tokens_seen": 3088580608
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.479438314944834e-05,
      "loss": 1.2097,
      "theoretical_loss": 3.34182578594486,
      "tokens_seen": 3088646144
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.477432296890672e-05,
      "loss": 1.3098,
      "theoretical_loss": 3.3418204124909483,
      "tokens_seen": 3088711680
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.47542627883651e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.341815039182972,
      "tokens_seen": 3088777216
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.473420260782348e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.3418096660209238,
      "tokens_seen": 3088842752
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.471414242728185e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.3418042930047966,
      "tokens_seen": 3088908288
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.469408224674022e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.3417989201345835,
      "tokens_seen": 3088973824
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.46740220661986e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.341793547410277,
      "tokens_seen": 3089039360
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.465396188565697e-05,
      "loss": 1.2007,
      "theoretical_loss": 3.341788174831871,
      "tokens_seen": 3089104896
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.463390170511535e-05,
      "loss": 1.1841,
      "theoretical_loss": 3.341782802399358,
      "tokens_seen": 3089170432
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.461384152457372e-05,
      "loss": 1.2498,
      "theoretical_loss": 3.3417774301127303,
      "tokens_seen": 3089235968
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.45937813440321e-05,
      "loss": 1.2002,
      "theoretical_loss": 3.3417720579719816,
      "tokens_seen": 3089301504
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.457372116349048e-05,
      "loss": 1.2387,
      "theoretical_loss": 3.3417666859771047,
      "tokens_seen": 3089367040
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.455366098294885e-05,
      "loss": 1.2,
      "theoretical_loss": 3.3417613141280924,
      "tokens_seen": 3089432576
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.453360080240723e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.3417559424249377,
      "tokens_seen": 3089498112
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.451354062186559e-05,
      "loss": 1.2499,
      "theoretical_loss": 3.3417505708676334,
      "tokens_seen": 3089563648
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.449348044132397e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.341745199456173,
      "tokens_seen": 3089629184
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.447342026078234e-05,
      "loss": 1.2121,
      "theoretical_loss": 3.3417398281905486,
      "tokens_seen": 3089694720
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.445336008024072e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.341734457070754,
      "tokens_seen": 3089760256
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.44332998996991e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.341729086096781,
      "tokens_seen": 3089825792
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.441323971915747e-05,
      "loss": 1.3125,
      "theoretical_loss": 3.3417237152686243,
      "tokens_seen": 3089891328
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.439317953861585e-05,
      "loss": 1.3464,
      "theoretical_loss": 3.3417183445862753,
      "tokens_seen": 3089956864
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.437311935807423e-05,
      "loss": 1.185,
      "theoretical_loss": 3.3417129740497273,
      "tokens_seen": 3090022400
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.435305917753261e-05,
      "loss": 1.2053,
      "theoretical_loss": 3.341707603658974,
      "tokens_seen": 3090087936
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.433299899699099e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.3417022334140074,
      "tokens_seen": 3090153472
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.431293881644935e-05,
      "loss": 1.2495,
      "theoretical_loss": 3.341696863314821,
      "tokens_seen": 3090219008
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.429287863590772e-05,
      "loss": 1.1847,
      "theoretical_loss": 3.3416914933614073,
      "tokens_seen": 3090284544
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.42728184553661e-05,
      "loss": 1.1721,
      "theoretical_loss": 3.34168612355376,
      "tokens_seen": 3090350080
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.425275827482448e-05,
      "loss": 1.131,
      "theoretical_loss": 3.341680753891871,
      "tokens_seen": 3090415616
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.423269809428285e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.3416753843757343,
      "tokens_seen": 3090481152
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.421263791374123e-05,
      "loss": 1.1549,
      "theoretical_loss": 3.3416700150053424,
      "tokens_seen": 3090546688
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.41925777331996e-05,
      "loss": 1.216,
      "theoretical_loss": 3.341664645780688,
      "tokens_seen": 3090612224
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.417251755265798e-05,
      "loss": 1.214,
      "theoretical_loss": 3.3416592767017645,
      "tokens_seen": 3090677760
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.415245737211635e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3416539077685643,
      "tokens_seen": 3090743296
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.413239719157472e-05,
      "loss": 1.2497,
      "theoretical_loss": 3.3416485389810813,
      "tokens_seen": 3090808832
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.411233701103309e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.3416431703393075,
      "tokens_seen": 3090874368
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.409227683049147e-05,
      "loss": 1.1487,
      "theoretical_loss": 3.341637801843236,
      "tokens_seen": 3090939904
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.407221664994984e-05,
      "loss": 1.1276,
      "theoretical_loss": 3.3416324334928604,
      "tokens_seen": 3091005440
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.405215646940822e-05,
      "loss": 1.1784,
      "theoretical_loss": 3.341627065288173,
      "tokens_seen": 3091070976
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.40320962888666e-05,
      "loss": 1.2342,
      "theoretical_loss": 3.3416216972291672,
      "tokens_seen": 3091136512
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.401203610832497e-05,
      "loss": 1.2473,
      "theoretical_loss": 3.3416163293158356,
      "tokens_seen": 3091202048
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.399197592778336e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.3416109615481715,
      "tokens_seen": 3091267584
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.397191574724174e-05,
      "loss": 1.2388,
      "theoretical_loss": 3.3416055939261673,
      "tokens_seen": 3091333120
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.395185556670011e-05,
      "loss": 1.1963,
      "theoretical_loss": 3.341600226449817,
      "tokens_seen": 3091398656
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.393179538615847e-05,
      "loss": 1.1633,
      "theoretical_loss": 3.341594859119112,
      "tokens_seen": 3091464192
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.391173520561685e-05,
      "loss": 1.2009,
      "theoretical_loss": 3.3415894919340468,
      "tokens_seen": 3091529728
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.389167502507523e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.3415841248946134,
      "tokens_seen": 3091595264
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.38716148445336e-05,
      "loss": 1.2147,
      "theoretical_loss": 3.3415787580008054,
      "tokens_seen": 3091660800
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.385155466399198e-05,
      "loss": 1.245,
      "theoretical_loss": 3.341573391252615,
      "tokens_seen": 3091726336
    },
    {
      "epoch": 0.87,
      "objective/train/docs_used": 1123259,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2203683853149414,
      "objective/train/theoretical_loss": 3.341568024650036,
      "objective/train/tokens_used": 1462250976,
      "theoretical_loss": 3.341568024650036,
      "tokens_seen": 3091791872
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.383149448345035e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.341568024650036,
      "tokens_seen": 3091791872
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.381143430290873e-05,
      "loss": 1.1939,
      "theoretical_loss": 3.3415626581930606,
      "tokens_seen": 3091857408
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.37913741223671e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.3415572918816823,
      "tokens_seen": 3091922944
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.377131394182548e-05,
      "loss": 1.4183,
      "theoretical_loss": 3.341551925715894,
      "tokens_seen": 3091988480
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.375125376128386e-05,
      "loss": 1.264,
      "theoretical_loss": 3.3415465596956886,
      "tokens_seen": 3092054016
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.373119358074222e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.3415411938210586,
      "tokens_seen": 3092119552
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.37111334002006e-05,
      "loss": 1.3256,
      "theoretical_loss": 3.341535828091998,
      "tokens_seen": 3092185088
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.369107321965897e-05,
      "loss": 1.203,
      "theoretical_loss": 3.3415304625084987,
      "tokens_seen": 3092250624
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.367101303911735e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.341525097070554,
      "tokens_seen": 3092316160
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.365095285857572e-05,
      "loss": 1.3098,
      "theoretical_loss": 3.3415197317781575,
      "tokens_seen": 3092381696
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.36308926780341e-05,
      "loss": 1.2132,
      "theoretical_loss": 3.3415143666313014,
      "tokens_seen": 3092447232
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.361083249749249e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.3415090016299787,
      "tokens_seen": 3092512768
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.359077231695086e-05,
      "loss": 1.2093,
      "theoretical_loss": 3.341503636774183,
      "tokens_seen": 3092578304
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.357071213640924e-05,
      "loss": 1.2278,
      "theoretical_loss": 3.3414982720639066,
      "tokens_seen": 3092643840
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.355065195586761e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.3414929074991426,
      "tokens_seen": 3092709376
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.353059177532598e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.341487543079884,
      "tokens_seen": 3092774912
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.351053159478435e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.341482178806124,
      "tokens_seen": 3092840448
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.349047141424273e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.341476814677856,
      "tokens_seen": 3092905984
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.34704112337011e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.341471450695072,
      "tokens_seen": 3092971520
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.345035105315948e-05,
      "loss": 1.1896,
      "theoretical_loss": 3.341466086857765,
      "tokens_seen": 3093037056
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.343029087261786e-05,
      "loss": 1.2855,
      "theoretical_loss": 3.3414607231659286,
      "tokens_seen": 3093102592
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.341023069207623e-05,
      "loss": 1.2324,
      "theoretical_loss": 3.3414553596195558,
      "tokens_seen": 3093168128
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.339017051153461e-05,
      "loss": 1.1845,
      "theoretical_loss": 3.341449996218639,
      "tokens_seen": 3093233664
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.337011033099298e-05,
      "loss": 1.2195,
      "theoretical_loss": 3.3414446329631717,
      "tokens_seen": 3093299200
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.335005015045135e-05,
      "loss": 1.258,
      "theoretical_loss": 3.3414392698531463,
      "tokens_seen": 3093364736
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.332998996990972e-05,
      "loss": 1.4057,
      "theoretical_loss": 3.3414339068885566,
      "tokens_seen": 3093430272
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.33099297893681e-05,
      "loss": 1.18,
      "theoretical_loss": 3.3414285440693945,
      "tokens_seen": 3093495808
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.328986960882647e-05,
      "loss": 1.3387,
      "theoretical_loss": 3.341423181395654,
      "tokens_seen": 3093561344
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.326980942828485e-05,
      "loss": 1.3045,
      "theoretical_loss": 3.3414178188673276,
      "tokens_seen": 3093626880
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.324974924774324e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.3414124564844085,
      "tokens_seen": 3093692416
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.322968906720161e-05,
      "loss": 1.2431,
      "theoretical_loss": 3.341407094246889,
      "tokens_seen": 3093757952
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.320962888665999e-05,
      "loss": 1.3941,
      "theoretical_loss": 3.341401732154763,
      "tokens_seen": 3093823488
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.318956870611837e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.341396370208023,
      "tokens_seen": 3093889024
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.316950852557674e-05,
      "loss": 1.2336,
      "theoretical_loss": 3.341391008406662,
      "tokens_seen": 3093954560
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.31494483450351e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.3413856467506724,
      "tokens_seen": 3094020096
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.312938816449348e-05,
      "loss": 1.361,
      "theoretical_loss": 3.3413802852400485,
      "tokens_seen": 3094085632
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.310932798395186e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.3413749238747825,
      "tokens_seen": 3094151168
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.308926780341023e-05,
      "loss": 1.2301,
      "theoretical_loss": 3.3413695626548674,
      "tokens_seen": 3094216704
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.306920762286861e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.341364201580296,
      "tokens_seen": 3094282240
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.304914744232698e-05,
      "loss": 1.1271,
      "theoretical_loss": 3.341358840651062,
      "tokens_seen": 3094347776
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.302908726178536e-05,
      "loss": 1.2249,
      "theoretical_loss": 3.3413534798671574,
      "tokens_seen": 3094413312
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.300902708124373e-05,
      "loss": 1.1055,
      "theoretical_loss": 3.341348119228576,
      "tokens_seen": 3094478848
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.298896690070211e-05,
      "loss": 1.2265,
      "theoretical_loss": 3.34134275873531,
      "tokens_seen": 3094544384
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.296890672016049e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.3413373983873536,
      "tokens_seen": 3094609920
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.294884653961885e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.3413320381846985,
      "tokens_seen": 3094675456
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.292878635907722e-05,
      "loss": 1.1069,
      "theoretical_loss": 3.3413266781273383,
      "tokens_seen": 3094740992
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.29087261785356e-05,
      "loss": 1.3151,
      "theoretical_loss": 3.341321318215266,
      "tokens_seen": 3094806528
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.288866599799398e-05,
      "loss": 1.173,
      "theoretical_loss": 3.341315958448474,
      "tokens_seen": 3094872064
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.286860581745236e-05,
      "loss": 1.2616,
      "theoretical_loss": 3.3413105988269565,
      "tokens_seen": 3094937600
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.284854563691074e-05,
      "loss": 1.1863,
      "theoretical_loss": 3.3413052393507057,
      "tokens_seen": 3095003136
    },
    {
      "epoch": 0.88,
      "objective/train/docs_used": 1124452,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.351484775543213,
      "objective/train/theoretical_loss": 3.3412998800197142,
      "objective/train/tokens_used": 1465527776,
      "theoretical_loss": 3.3412998800197142,
      "tokens_seen": 3095068672
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.282848545636912e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.3412998800197142,
      "tokens_seen": 3095068672
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.280842527582749e-05,
      "loss": 1.2323,
      "theoretical_loss": 3.3412945208339755,
      "tokens_seen": 3095134208
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.278836509528587e-05,
      "loss": 1.3331,
      "theoretical_loss": 3.341289161793483,
      "tokens_seen": 3095199744
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.276830491474424e-05,
      "loss": 1.1998,
      "theoretical_loss": 3.3412838028982286,
      "tokens_seen": 3095265280
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.27482447342026e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.341278444148206,
      "tokens_seen": 3095330816
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.272818455366098e-05,
      "loss": 1.3901,
      "theoretical_loss": 3.341273085543408,
      "tokens_seen": 3095396352
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.270812437311936e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.341267727083828,
      "tokens_seen": 3095461888
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.268806419257773e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.341262368769459,
      "tokens_seen": 3095527424
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.266800401203611e-05,
      "loss": 1.1279,
      "theoretical_loss": 3.341257010600293,
      "tokens_seen": 3095592960
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.264794383149448e-05,
      "loss": 1.1551,
      "theoretical_loss": 3.341251652576324,
      "tokens_seen": 3095658496
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.262788365095286e-05,
      "loss": 1.3007,
      "theoretical_loss": 3.3412462946975445,
      "tokens_seen": 3095724032
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.260782347041124e-05,
      "loss": 1.2125,
      "theoretical_loss": 3.3412409369639473,
      "tokens_seen": 3095789568
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.258776328986961e-05,
      "loss": 1.1926,
      "theoretical_loss": 3.341235579375526,
      "tokens_seen": 3095855104
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.256770310932797e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.3412302219322734,
      "tokens_seen": 3095920640
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.254764292878635e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.3412248646341824,
      "tokens_seen": 3095986176
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.252758274824473e-05,
      "loss": 1.336,
      "theoretical_loss": 3.341219507481246,
      "tokens_seen": 3096051712
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.25075225677031e-05,
      "loss": 1.3127,
      "theoretical_loss": 3.341214150473457,
      "tokens_seen": 3096117248
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.248746238716149e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.341208793610809,
      "tokens_seen": 3096182784
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.246740220661985e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.341203436893294,
      "tokens_seen": 3096248320
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.244734202607823e-05,
      "loss": 1.1846,
      "theoretical_loss": 3.3411980803209063,
      "tokens_seen": 3096313856
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.24272818455366e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.341192723893638,
      "tokens_seen": 3096379392
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.240722166499498e-05,
      "loss": 1.2387,
      "theoretical_loss": 3.341187367611482,
      "tokens_seen": 3096444928
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.238716148445337e-05,
      "loss": 1.1735,
      "theoretical_loss": 3.3411820114744315,
      "tokens_seen": 3096510464
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.236710130391173e-05,
      "loss": 1.168,
      "theoretical_loss": 3.3411766554824798,
      "tokens_seen": 3096576000
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.234704112337011e-05,
      "loss": 1.3184,
      "theoretical_loss": 3.3411712996356195,
      "tokens_seen": 3096641536
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.232698094282848e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.341165943933844,
      "tokens_seen": 3096707072
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.230692076228686e-05,
      "loss": 1.2032,
      "theoretical_loss": 3.341160588377146,
      "tokens_seen": 3096772608
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.228686058174524e-05,
      "loss": 1.2316,
      "theoretical_loss": 3.3411552329655185,
      "tokens_seen": 3096838144
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.226680040120361e-05,
      "loss": 1.1589,
      "theoretical_loss": 3.341149877698955,
      "tokens_seen": 3096903680
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.224674022066199e-05,
      "loss": 1.156,
      "theoretical_loss": 3.3411445225774474,
      "tokens_seen": 3096969216
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.222668004012036e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.34113916760099,
      "tokens_seen": 3097034752
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.220661985957874e-05,
      "loss": 1.3819,
      "theoretical_loss": 3.3411338127695744,
      "tokens_seen": 3097100288
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.218655967903711e-05,
      "loss": 1.1877,
      "theoretical_loss": 3.341128458083195,
      "tokens_seen": 3097165824
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.216649949849549e-05,
      "loss": 1.2291,
      "theoretical_loss": 3.341123103541844,
      "tokens_seen": 3097231360
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.214643931795387e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.3411177491455146,
      "tokens_seen": 3097296896
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.212637913741224e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.3411123948942,
      "tokens_seen": 3097362432
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.210631895687062e-05,
      "loss": 1.2829,
      "theoretical_loss": 3.3411070407878927,
      "tokens_seen": 3097427968
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.2086258776329e-05,
      "loss": 1.1434,
      "theoretical_loss": 3.341101686826586,
      "tokens_seen": 3097493504
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.206619859578736e-05,
      "loss": 1.1952,
      "theoretical_loss": 3.341096333010273,
      "tokens_seen": 3097559040
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.204613841524573e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.3410909793389463,
      "tokens_seen": 3097624576
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.202607823470412e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.3410856258125996,
      "tokens_seen": 3097690112
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.20060180541625e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.3410802724312254,
      "tokens_seen": 3097755648
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.198595787362086e-05,
      "loss": 1.288,
      "theoretical_loss": 3.341074919194817,
      "tokens_seen": 3097821184
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.196589769307923e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.341069566103367,
      "tokens_seen": 3097886720
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.194583751253761e-05,
      "loss": 1.2051,
      "theoretical_loss": 3.341064213156869,
      "tokens_seen": 3097952256
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.192577733199599e-05,
      "loss": 1.3285,
      "theoretical_loss": 3.3410588603553153,
      "tokens_seen": 3098017792
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.190571715145436e-05,
      "loss": 1.2204,
      "theoretical_loss": 3.3410535076986996,
      "tokens_seen": 3098083328
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.188565697091274e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.341048155187014,
      "tokens_seen": 3098148864
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.186559679037111e-05,
      "loss": 1.304,
      "theoretical_loss": 3.3410428028202523,
      "tokens_seen": 3098214400
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.184553660982949e-05,
      "loss": 1.1779,
      "theoretical_loss": 3.3410374505984075,
      "tokens_seen": 3098279936
    },
    {
      "epoch": 0.88,
      "objective/train/docs_used": 1125757,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2427351474761963,
      "objective/train/theoretical_loss": 3.3410320985214725,
      "objective/train/tokens_used": 1468804576,
      "theoretical_loss": 3.3410320985214725,
      "tokens_seen": 3098345472
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.182547642928787e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.3410320985214725,
      "tokens_seen": 3098345472
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.180541624874624e-05,
      "loss": 1.3113,
      "theoretical_loss": 3.34102674658944,
      "tokens_seen": 3098411008
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.178535606820462e-05,
      "loss": 1.1997,
      "theoretical_loss": 3.3410213948023033,
      "tokens_seen": 3098476544
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.176529588766299e-05,
      "loss": 1.1998,
      "theoretical_loss": 3.3410160431600553,
      "tokens_seen": 3098542080
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.174523570712137e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.3410106916626887,
      "tokens_seen": 3098607616
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.172517552657974e-05,
      "loss": 1.2111,
      "theoretical_loss": 3.3410053403101974,
      "tokens_seen": 3098673152
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.170511534603812e-05,
      "loss": 1.318,
      "theoretical_loss": 3.3409999891025737,
      "tokens_seen": 3098738688
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.168505516549648e-05,
      "loss": 1.3577,
      "theoretical_loss": 3.3409946380398106,
      "tokens_seen": 3098804224
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.166499498495486e-05,
      "loss": 1.221,
      "theoretical_loss": 3.340989287121902,
      "tokens_seen": 3098869760
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.164493480441325e-05,
      "loss": 1.1947,
      "theoretical_loss": 3.3409839363488394,
      "tokens_seen": 3098935296
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.162487462387162e-05,
      "loss": 1.2318,
      "theoretical_loss": 3.340978585720617,
      "tokens_seen": 3099000832
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.160481444333e-05,
      "loss": 1.2255,
      "theoretical_loss": 3.3409732352372274,
      "tokens_seen": 3099066368
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.158475426278836e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.3409678848986637,
      "tokens_seen": 3099131904
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.156469408224674e-05,
      "loss": 1.4005,
      "theoretical_loss": 3.340962534704919,
      "tokens_seen": 3099197440
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.154463390170511e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.340957184655986,
      "tokens_seen": 3099262976
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.152457372116349e-05,
      "loss": 1.2999,
      "theoretical_loss": 3.3409518347518583,
      "tokens_seen": 3099328512
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.150451354062186e-05,
      "loss": 1.1839,
      "theoretical_loss": 3.340946484992528,
      "tokens_seen": 3099394048
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.148445336008024e-05,
      "loss": 1.2593,
      "theoretical_loss": 3.340941135377989,
      "tokens_seen": 3099459584
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.146439317953862e-05,
      "loss": 1.3403,
      "theoretical_loss": 3.340935785908234,
      "tokens_seen": 3099525120
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.144433299899699e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.340930436583256,
      "tokens_seen": 3099590656
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.142427281845537e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.340925087403048,
      "tokens_seen": 3099656192
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.140421263791374e-05,
      "loss": 1.1589,
      "theoretical_loss": 3.340919738367603,
      "tokens_seen": 3099721728
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.138415245737212e-05,
      "loss": 1.3281,
      "theoretical_loss": 3.3409143894769144,
      "tokens_seen": 3099787264
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.13640922768305e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.3409090407309745,
      "tokens_seen": 3099852800
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.134403209628887e-05,
      "loss": 1.2322,
      "theoretical_loss": 3.3409036921297766,
      "tokens_seen": 3099918336
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.132397191574725e-05,
      "loss": 1.2501,
      "theoretical_loss": 3.3408983436733144,
      "tokens_seen": 3099983872
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.130391173520561e-05,
      "loss": 1.3634,
      "theoretical_loss": 3.34089299536158,
      "tokens_seen": 3100049408
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.128385155466398e-05,
      "loss": 1.2189,
      "theoretical_loss": 3.3408876471945668,
      "tokens_seen": 3100114944
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.126379137412237e-05,
      "loss": 1.2496,
      "theoretical_loss": 3.340882299172268,
      "tokens_seen": 3100180480
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.124373119358075e-05,
      "loss": 1.2441,
      "theoretical_loss": 3.340876951294676,
      "tokens_seen": 3100246016
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.122367101303913e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.3408716035617845,
      "tokens_seen": 3100311552
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.120361083249749e-05,
      "loss": 1.1541,
      "theoretical_loss": 3.3408662559735864,
      "tokens_seen": 3100377088
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.118355065195586e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.3408609085300744,
      "tokens_seen": 3100442624
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.116349047141424e-05,
      "loss": 1.1904,
      "theoretical_loss": 3.3408555612312423,
      "tokens_seen": 3100508160
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.114343029087262e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.340850214077082,
      "tokens_seen": 3100573696
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.112337011033099e-05,
      "loss": 1.3495,
      "theoretical_loss": 3.340844867067587,
      "tokens_seen": 3100639232
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.110330992978937e-05,
      "loss": 1.2118,
      "theoretical_loss": 3.340839520202751,
      "tokens_seen": 3100704768
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.108324974924774e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.340834173482566,
      "tokens_seen": 3100770304
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.106318956870612e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.3408288269070257,
      "tokens_seen": 3100835840
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.10431293881645e-05,
      "loss": 1.2219,
      "theoretical_loss": 3.340823480476123,
      "tokens_seen": 3100901376
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.102306920762287e-05,
      "loss": 1.1932,
      "theoretical_loss": 3.340818134189851,
      "tokens_seen": 3100966912
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.1003009027081246e-05,
      "loss": 1.2275,
      "theoretical_loss": 3.340812788048202,
      "tokens_seen": 3101032448
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.098294884653962e-05,
      "loss": 1.2312,
      "theoretical_loss": 3.3408074420511698,
      "tokens_seen": 3101097984
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.096288866599799e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.3408020961987472,
      "tokens_seen": 3101163520
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0942828485456366e-05,
      "loss": 1.192,
      "theoretical_loss": 3.3407967504909277,
      "tokens_seen": 3101229056
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.092276830491474e-05,
      "loss": 1.1062,
      "theoretical_loss": 3.3407914049277037,
      "tokens_seen": 3101294592
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0902708124373125e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.3407860595090684,
      "tokens_seen": 3101360128
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.08826479438315e-05,
      "loss": 1.1511,
      "theoretical_loss": 3.3407807142350148,
      "tokens_seen": 3101425664
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.086258776328987e-05,
      "loss": 1.2025,
      "theoretical_loss": 3.3407753691055357,
      "tokens_seen": 3101491200
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0842527582748245e-05,
      "loss": 1.1946,
      "theoretical_loss": 3.340770024120625,
      "tokens_seen": 3101556736
    },
    {
      "epoch": 0.88,
      "objective/train/docs_used": 1126928,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1931419372558594,
      "objective/train/theoretical_loss": 3.3407646792802748,
      "objective/train/tokens_used": 1472081376,
      "theoretical_loss": 3.3407646792802748,
      "tokens_seen": 3101622272
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.082246740220662e-05,
      "loss": 1.2137,
      "theoretical_loss": 3.3407646792802748,
      "tokens_seen": 3101622272
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0802407221665e-05,
      "loss": 1.1793,
      "theoretical_loss": 3.3407593345844786,
      "tokens_seen": 3101687808
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.078234704112337e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.3407539900332295,
      "tokens_seen": 3101753344
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.076228686058174e-05,
      "loss": 1.1753,
      "theoretical_loss": 3.3407486456265203,
      "tokens_seen": 3101818880
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.074222668004012e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.3407433013643444,
      "tokens_seen": 3101884416
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.072216649949849e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.340737957246694,
      "tokens_seen": 3101949952
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0702106318956876e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.3407326132735635,
      "tokens_seen": 3102015488
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.068204613841525e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.3407272694449444,
      "tokens_seen": 3102081024
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.066198595787362e-05,
      "loss": 1.1462,
      "theoretical_loss": 3.340721925760831,
      "tokens_seen": 3102146560
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0641925777331996e-05,
      "loss": 1.1678,
      "theoretical_loss": 3.3407165822212153,
      "tokens_seen": 3102212096
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.062186559679037e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.340711238826091,
      "tokens_seen": 3102277632
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.060180541624875e-05,
      "loss": 1.278,
      "theoretical_loss": 3.340705895575451,
      "tokens_seen": 3102343168
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0581745235707124e-05,
      "loss": 1.295,
      "theoretical_loss": 3.3407005524692885,
      "tokens_seen": 3102408704
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.056168505516549e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.3406952095075964,
      "tokens_seen": 3102474240
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.054162487462387e-05,
      "loss": 1.2778,
      "theoretical_loss": 3.3406898666903677,
      "tokens_seen": 3102539776
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.052156469408225e-05,
      "loss": 1.318,
      "theoretical_loss": 3.3406845240175955,
      "tokens_seen": 3102605312
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.050150451354063e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.3406791814892727,
      "tokens_seen": 3102670848
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0481444332999e-05,
      "loss": 1.1638,
      "theoretical_loss": 3.3406738391053925,
      "tokens_seen": 3102736384
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.046138415245737e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.340668496865948,
      "tokens_seen": 3102801920
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.044132397191575e-05,
      "loss": 1.1666,
      "theoretical_loss": 3.340663154770932,
      "tokens_seen": 3102867456
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0421263791374123e-05,
      "loss": 1.1958,
      "theoretical_loss": 3.3406578128203375,
      "tokens_seen": 3102932992
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.04012036108325e-05,
      "loss": 1.196,
      "theoretical_loss": 3.340652471014158,
      "tokens_seen": 3102998528
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0381143430290875e-05,
      "loss": 1.1192,
      "theoretical_loss": 3.3406471293523863,
      "tokens_seen": 3103064064
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0361083249749244e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.340641787835015,
      "tokens_seen": 3103129600
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0341023069207627e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.340636446462038,
      "tokens_seen": 3103195136
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0320962888666e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.340631105233448,
      "tokens_seen": 3103260672
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.030090270812438e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.340625764149238,
      "tokens_seen": 3103326208
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0280842527582754e-05,
      "loss": 1.233,
      "theoretical_loss": 3.340620423209401,
      "tokens_seen": 3103391744
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.026078234704112e-05,
      "loss": 1.2161,
      "theoretical_loss": 3.3406150824139296,
      "tokens_seen": 3103457280
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.02407221664995e-05,
      "loss": 1.2005,
      "theoretical_loss": 3.3406097417628176,
      "tokens_seen": 3103522816
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0220661985957875e-05,
      "loss": 1.2195,
      "theoretical_loss": 3.3406044012560576,
      "tokens_seen": 3103588352
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.020060180541625e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.340599060893643,
      "tokens_seen": 3103653888
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.018054162487462e-05,
      "loss": 1.2282,
      "theoretical_loss": 3.3405937206755665,
      "tokens_seen": 3103719424
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0160481444332995e-05,
      "loss": 1.1573,
      "theoretical_loss": 3.3405883806018215,
      "tokens_seen": 3103784960
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.014042126379138e-05,
      "loss": 1.178,
      "theoretical_loss": 3.340583040672401,
      "tokens_seen": 3103850496
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0120361083249754e-05,
      "loss": 1.3732,
      "theoretical_loss": 3.3405777008872977,
      "tokens_seen": 3103916032
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.010030090270813e-05,
      "loss": 1.3299,
      "theoretical_loss": 3.340572361246505,
      "tokens_seen": 3103981568
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.00802407221665e-05,
      "loss": 1.2345,
      "theoretical_loss": 3.3405670217500156,
      "tokens_seen": 3104047104
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0060180541624874e-05,
      "loss": 1.1837,
      "theoretical_loss": 3.340561682397823,
      "tokens_seen": 3104112640
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.004012036108325e-05,
      "loss": 1.2317,
      "theoretical_loss": 3.34055634318992,
      "tokens_seen": 3104178176
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0020060180541626e-05,
      "loss": 1.2507,
      "theoretical_loss": 3.3405510041263,
      "tokens_seen": 3104243712
    },
    {
      "epoch": 0.88,
      "learning_rate": 6e-05,
      "loss": 1.2443,
      "theoretical_loss": 3.340545665206955,
      "tokens_seen": 3104309248
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.997993981945837e-05,
      "loss": 1.2084,
      "theoretical_loss": 3.340540326431879,
      "tokens_seen": 3104374784
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.995987963891675e-05,
      "loss": 1.2653,
      "theoretical_loss": 3.3405349878010653,
      "tokens_seen": 3104440320
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.993981945837513e-05,
      "loss": 1.2299,
      "theoretical_loss": 3.3405296493145062,
      "tokens_seen": 3104505856
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9919759277833505e-05,
      "loss": 1.2184,
      "theoretical_loss": 3.340524310972195,
      "tokens_seen": 3104571392
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.989969909729188e-05,
      "loss": 1.1504,
      "theoretical_loss": 3.340518972774125,
      "tokens_seen": 3104636928
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.987963891675025e-05,
      "loss": 1.2239,
      "theoretical_loss": 3.340513634720289,
      "tokens_seen": 3104702464
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9859578736208625e-05,
      "loss": 1.1887,
      "theoretical_loss": 3.3405082968106803,
      "tokens_seen": 3104768000
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9839518555667e-05,
      "loss": 1.4095,
      "theoretical_loss": 3.3405029590452915,
      "tokens_seen": 3104833536
    },
    {
      "epoch": 0.88,
      "objective/train/docs_used": 1128234,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2622343301773071,
      "objective/train/theoretical_loss": 3.340497621424116,
      "objective/train/tokens_used": 1475358176,
      "theoretical_loss": 3.340497621424116,
      "tokens_seen": 3104899072
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.981945837512538e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.340497621424116,
      "tokens_seen": 3104899072
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.979939819458375e-05,
      "loss": 1.2092,
      "theoretical_loss": 3.340492283947147,
      "tokens_seen": 3104964608
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.977933801404212e-05,
      "loss": 1.1484,
      "theoretical_loss": 3.3404869466143774,
      "tokens_seen": 3105030144
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9759277833500504e-05,
      "loss": 1.2252,
      "theoretical_loss": 3.3404816094258,
      "tokens_seen": 3105095680
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.973921765295888e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.3404762723814083,
      "tokens_seen": 3105161216
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9719157472417256e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.3404709354811954,
      "tokens_seen": 3105226752
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.969909729187563e-05,
      "loss": 1.3244,
      "theoretical_loss": 3.340465598725154,
      "tokens_seen": 3105292288
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9679037111334e-05,
      "loss": 1.3497,
      "theoretical_loss": 3.340460262113277,
      "tokens_seen": 3105357824
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9658976930792376e-05,
      "loss": 1.2545,
      "theoretical_loss": 3.340454925645558,
      "tokens_seen": 3105423360
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.963891675025075e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.3404495893219894,
      "tokens_seen": 3105488896
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.961885656970913e-05,
      "loss": 1.2094,
      "theoretical_loss": 3.3404442531425653,
      "tokens_seen": 3105554432
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9598796389167504e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.3404389171072775,
      "tokens_seen": 3105619968
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.957873620862588e-05,
      "loss": 1.1886,
      "theoretical_loss": 3.34043358121612,
      "tokens_seen": 3105685504
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9558676028084255e-05,
      "loss": 1.3585,
      "theoretical_loss": 3.3404282454690857,
      "tokens_seen": 3105751040
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.953861584754263e-05,
      "loss": 1.3469,
      "theoretical_loss": 3.3404229098661675,
      "tokens_seen": 3105816576
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.951855566700101e-05,
      "loss": 1.2002,
      "theoretical_loss": 3.3404175744073585,
      "tokens_seen": 3105882112
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.949849548645938e-05,
      "loss": 1.1512,
      "theoretical_loss": 3.3404122390926516,
      "tokens_seen": 3105947648
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.947843530591775e-05,
      "loss": 1.3575,
      "theoretical_loss": 3.3404069039220405,
      "tokens_seen": 3106013184
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.945837512537613e-05,
      "loss": 1.1912,
      "theoretical_loss": 3.3404015688955173,
      "tokens_seen": 3106078720
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9438314944834503e-05,
      "loss": 1.2294,
      "theoretical_loss": 3.3403962340130757,
      "tokens_seen": 3106144256
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.941825476429288e-05,
      "loss": 1.254,
      "theoretical_loss": 3.340390899274709,
      "tokens_seen": 3106209792
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9398194583751255e-05,
      "loss": 1.2046,
      "theoretical_loss": 3.3403855646804095,
      "tokens_seen": 3106275328
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.937813440320963e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.340380230230171,
      "tokens_seen": 3106340864
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9358074222668007e-05,
      "loss": 1.1637,
      "theoretical_loss": 3.340374895923986,
      "tokens_seen": 3106406400
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.933801404212638e-05,
      "loss": 1.264,
      "theoretical_loss": 3.340369561761848,
      "tokens_seen": 3106471936
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.931795386158476e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.3403642277437497,
      "tokens_seen": 3106537472
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.929789368104313e-05,
      "loss": 1.2091,
      "theoretical_loss": 3.3403588938696847,
      "tokens_seen": 3106603008
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.92778335005015e-05,
      "loss": 1.3145,
      "theoretical_loss": 3.3403535601396452,
      "tokens_seen": 3106668544
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.925777331995988e-05,
      "loss": 1.1576,
      "theoretical_loss": 3.3403482265536253,
      "tokens_seen": 3106734080
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9237713139418255e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.3403428931116177,
      "tokens_seen": 3106799616
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.921765295887663e-05,
      "loss": 1.289,
      "theoretical_loss": 3.340337559813615,
      "tokens_seen": 3106865152
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9197592778335006e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.3403322266596107,
      "tokens_seen": 3106930688
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.917753259779338e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.340326893649598,
      "tokens_seen": 3106996224
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.915747241725176e-05,
      "loss": 1.1894,
      "theoretical_loss": 3.3403215607835697,
      "tokens_seen": 3107061760
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9137412236710134e-05,
      "loss": 1.1648,
      "theoretical_loss": 3.3403162280615186,
      "tokens_seen": 3107127296
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.911735205616851e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.3403108954834386,
      "tokens_seen": 3107192832
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.909729187562688e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.340305563049322,
      "tokens_seen": 3107258368
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9077231695085254e-05,
      "loss": 1.3377,
      "theoretical_loss": 3.3403002307591625,
      "tokens_seen": 3107323904
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.905717151454363e-05,
      "loss": 1.1734,
      "theoretical_loss": 3.340294898612953,
      "tokens_seen": 3107389440
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9037111334002006e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.340289566610686,
      "tokens_seen": 3107454976
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.901705115346039e-05,
      "loss": 1.1824,
      "theoretical_loss": 3.3402842347523554,
      "tokens_seen": 3107520512
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.899699097291876e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.3402789030379534,
      "tokens_seen": 3107586048
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.897693079237713e-05,
      "loss": 1.2971,
      "theoretical_loss": 3.340273571467474,
      "tokens_seen": 3107651584
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.895687061183551e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.34026824004091,
      "tokens_seen": 3107717120
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8936810431293885e-05,
      "loss": 1.2149,
      "theoretical_loss": 3.3402629087582545,
      "tokens_seen": 3107782656
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.891675025075226e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.3402575776195,
      "tokens_seen": 3107848192
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.889669007021063e-05,
      "loss": 1.35,
      "theoretical_loss": 3.34025224662464,
      "tokens_seen": 3107913728
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8876629889669005e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.3402469157736676,
      "tokens_seen": 3107979264
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.885656970912738e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.340241585066576,
      "tokens_seen": 3108044800
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8836509528585764e-05,
      "loss": 1.2031,
      "theoretical_loss": 3.340236254503358,
      "tokens_seen": 3108110336
    },
    {
      "epoch": 0.88,
      "objective/train/docs_used": 1129377,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.267437219619751,
      "objective/train/theoretical_loss": 3.340230924084007,
      "objective/train/tokens_used": 1478634976,
      "theoretical_loss": 3.340230924084007,
      "tokens_seen": 3108175872
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.881644934804414e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.340230924084007,
      "tokens_seen": 3108175872
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.879638916750251e-05,
      "loss": 1.2417,
      "theoretical_loss": 3.340225593808516,
      "tokens_seen": 3108241408
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8776328986960884e-05,
      "loss": 1.1716,
      "theoretical_loss": 3.3402202636768776,
      "tokens_seen": 3108306944
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.875626880641926e-05,
      "loss": 1.1749,
      "theoretical_loss": 3.3402149336890856,
      "tokens_seen": 3108372480
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8736208625877636e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.3402096038451328,
      "tokens_seen": 3108438016
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8716148445336005e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.3402042741450124,
      "tokens_seen": 3108503552
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.869608826479438e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.3401989445887175,
      "tokens_seen": 3108569088
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8676028084252756e-05,
      "loss": 1.2731,
      "theoretical_loss": 3.3401936151762404,
      "tokens_seen": 3108634624
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.865596790371113e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3401882859075753,
      "tokens_seen": 3108700160
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8635907723169515e-05,
      "loss": 1.2252,
      "theoretical_loss": 3.340182956782715,
      "tokens_seen": 3108765696
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8615847542627884e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.3401776278016517,
      "tokens_seen": 3108831232
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.859578736208626e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.3401722989643794,
      "tokens_seen": 3108896768
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8575727181544635e-05,
      "loss": 1.285,
      "theoretical_loss": 3.340166970270891,
      "tokens_seen": 3108962304
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.855566700100301e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.34016164172118,
      "tokens_seen": 3109027840
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.853560682046139e-05,
      "loss": 1.1552,
      "theoretical_loss": 3.340156313315239,
      "tokens_seen": 3109093376
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8515546639919756e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.340150985053061,
      "tokens_seen": 3109158912
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.849548645937813e-05,
      "loss": 1.2307,
      "theoretical_loss": 3.340145656934639,
      "tokens_seen": 3109224448
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.847542627883651e-05,
      "loss": 1.3699,
      "theoretical_loss": 3.3401403289599663,
      "tokens_seen": 3109289984
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.845536609829489e-05,
      "loss": 1.3362,
      "theoretical_loss": 3.3401350011290365,
      "tokens_seen": 3109355520
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8435305917753266e-05,
      "loss": 1.2085,
      "theoretical_loss": 3.340129673441842,
      "tokens_seen": 3109421056
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8415245737211635e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.340124345898376,
      "tokens_seen": 3109486592
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.839518555667001e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.3401190184986316,
      "tokens_seen": 3109552128
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8375125376128387e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.340113691242602,
      "tokens_seen": 3109617664
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.835506519558676e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.3401083641302804,
      "tokens_seen": 3109683200
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.833500501504514e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.34010303716166,
      "tokens_seen": 3109748736
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.831494483450351e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.3400977103367335,
      "tokens_seen": 3109814272
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.829488465396188e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.3400923836554943,
      "tokens_seen": 3109879808
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8274824473420266e-05,
      "loss": 1.2201,
      "theoretical_loss": 3.340087057117935,
      "tokens_seen": 3109945344
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.825476429287864e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.3400817307240493,
      "tokens_seen": 3110010880
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.823470411233702e-05,
      "loss": 1.2344,
      "theoretical_loss": 3.3400764044738303,
      "tokens_seen": 3110076416
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8214643931795386e-05,
      "loss": 1.1968,
      "theoretical_loss": 3.3400710783672705,
      "tokens_seen": 3110141952
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.819458375125376e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.3400657524043638,
      "tokens_seen": 3110207488
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.817452357071214e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.3400604265851026,
      "tokens_seen": 3110273024
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.8154463390170514e-05,
      "loss": 1.2416,
      "theoretical_loss": 3.34005510090948,
      "tokens_seen": 3110338560
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.813440320962889e-05,
      "loss": 1.1812,
      "theoretical_loss": 3.34004977537749,
      "tokens_seen": 3110404096
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.811434302908726e-05,
      "loss": 1.3911,
      "theoretical_loss": 3.3400444499891244,
      "tokens_seen": 3110469632
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.8094282848545634e-05,
      "loss": 1.3125,
      "theoretical_loss": 3.3400391247443775,
      "tokens_seen": 3110535168
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.807422266800402e-05,
      "loss": 1.1755,
      "theoretical_loss": 3.3400337996432414,
      "tokens_seen": 3110600704
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.805416248746239e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.3400284746857096,
      "tokens_seen": 3110666240
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.803410230692077e-05,
      "loss": 1.3322,
      "theoretical_loss": 3.340023149871776,
      "tokens_seen": 3110731776
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.801404212637914e-05,
      "loss": 1.198,
      "theoretical_loss": 3.340017825201432,
      "tokens_seen": 3110797312
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.799398194583751e-05,
      "loss": 1.2571,
      "theoretical_loss": 3.340012500674672,
      "tokens_seen": 3110862848
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.797392176529589e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.3400071762914894,
      "tokens_seen": 3110928384
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7953861584754265e-05,
      "loss": 1.3194,
      "theoretical_loss": 3.340001852051876,
      "tokens_seen": 3110993920
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7933801404212634e-05,
      "loss": 1.2506,
      "theoretical_loss": 3.339996527955826,
      "tokens_seen": 3111059456
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.791374122367101e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.339991204003332,
      "tokens_seen": 3111124992
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.789368104312939e-05,
      "loss": 1.2344,
      "theoretical_loss": 3.339985880194387,
      "tokens_seen": 3111190528
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.787362086258777e-05,
      "loss": 1.2174,
      "theoretical_loss": 3.3399805565289844,
      "tokens_seen": 3111256064
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7853560682046144e-05,
      "loss": 1.1783,
      "theoretical_loss": 3.339975233007117,
      "tokens_seen": 3111321600
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.783350050150451e-05,
      "loss": 1.3798,
      "theoretical_loss": 3.3399699096287785,
      "tokens_seen": 3111387136
    },
    {
      "epoch": 0.89,
      "objective/train/docs_used": 1130437,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.337208867073059,
      "objective/train/theoretical_loss": 3.339964586393961,
      "objective/train/tokens_used": 1481911776,
      "theoretical_loss": 3.339964586393961,
      "tokens_seen": 3111452672
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.781344032096289e-05,
      "loss": 1.2394,
      "theoretical_loss": 3.339964586393961,
      "tokens_seen": 3111452672
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7793380140421264e-05,
      "loss": 1.2615,
      "theoretical_loss": 3.339959263302659,
      "tokens_seen": 3111518208
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.777331995987964e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.3399539403548646,
      "tokens_seen": 3111583744
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7753259779338016e-05,
      "loss": 1.2084,
      "theoretical_loss": 3.339948617550571,
      "tokens_seen": 3111649280
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7733199598796385e-05,
      "loss": 1.1771,
      "theoretical_loss": 3.3399432948897716,
      "tokens_seen": 3111714816
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.771313941825477e-05,
      "loss": 1.1843,
      "theoretical_loss": 3.339937972372459,
      "tokens_seen": 3111780352
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.769307923771314e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.339932649998627,
      "tokens_seen": 3111845888
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.767301905717152e-05,
      "loss": 1.3057,
      "theoretical_loss": 3.3399273277682684,
      "tokens_seen": 3111911424
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7652958876629895e-05,
      "loss": 1.1854,
      "theoretical_loss": 3.339922005681376,
      "tokens_seen": 3111976960
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7632898696088264e-05,
      "loss": 1.195,
      "theoretical_loss": 3.3399166837379433,
      "tokens_seen": 3112042496
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.761283851554664e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.339911361937964,
      "tokens_seen": 3112108032
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7592778335005015e-05,
      "loss": 1.217,
      "theoretical_loss": 3.33990604028143,
      "tokens_seen": 3112173568
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.757271815446339e-05,
      "loss": 1.1899,
      "theoretical_loss": 3.3399007187683347,
      "tokens_seen": 3112239104
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.755265797392177e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.3398953973986716,
      "tokens_seen": 3112304640
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7532597793380136e-05,
      "loss": 1.3839,
      "theoretical_loss": 3.3398900761724333,
      "tokens_seen": 3112370176
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.751253761283852e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.339884755089614,
      "tokens_seen": 3112435712
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7492477432296894e-05,
      "loss": 1.3874,
      "theoretical_loss": 3.339879434150206,
      "tokens_seen": 3112501248
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.747241725175527e-05,
      "loss": 1.2426,
      "theoretical_loss": 3.339874113354202,
      "tokens_seen": 3112566784
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7452357071213646e-05,
      "loss": 1.0943,
      "theoretical_loss": 3.339868792701596,
      "tokens_seen": 3112632320
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7432296890672015e-05,
      "loss": 1.25,
      "theoretical_loss": 3.3398634721923806,
      "tokens_seen": 3112697856
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.741223671013039e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.339858151826549,
      "tokens_seen": 3112763392
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7392176529588767e-05,
      "loss": 1.1358,
      "theoretical_loss": 3.339852831604095,
      "tokens_seen": 3112828928
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.737211634904714e-05,
      "loss": 1.3453,
      "theoretical_loss": 3.3398475115250106,
      "tokens_seen": 3112894464
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.735205616850552e-05,
      "loss": 1.315,
      "theoretical_loss": 3.3398421915892893,
      "tokens_seen": 3112960000
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7331995987963894e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.3398368717969245,
      "tokens_seen": 3113025536
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.731193580742227e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.3398315521479094,
      "tokens_seen": 3113091072
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7291875626880646e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.3398262326422365,
      "tokens_seen": 3113156608
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.727181544633902e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.3398209132798993,
      "tokens_seen": 3113222144
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.725175526579739e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.3398155940608913,
      "tokens_seen": 3113287680
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7231695085255766e-05,
      "loss": 1.3129,
      "theoretical_loss": 3.339810274985205,
      "tokens_seen": 3113353216
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.721163490471414e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.339804956052834,
      "tokens_seen": 3113418752
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.719157472417252e-05,
      "loss": 1.2159,
      "theoretical_loss": 3.339799637263771,
      "tokens_seen": 3113484288
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7171514543630893e-05,
      "loss": 1.1763,
      "theoretical_loss": 3.339794318618009,
      "tokens_seen": 3113549824
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.715145436308926e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.3397890001155415,
      "tokens_seen": 3113615360
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7131394182547645e-05,
      "loss": 1.2427,
      "theoretical_loss": 3.3397836817563618,
      "tokens_seen": 3113680896
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.711133400200602e-05,
      "loss": 1.3064,
      "theoretical_loss": 3.3397783635404625,
      "tokens_seen": 3113746432
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.70912738214644e-05,
      "loss": 1.271,
      "theoretical_loss": 3.339773045467837,
      "tokens_seen": 3113811968
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.707121364092277e-05,
      "loss": 1.1153,
      "theoretical_loss": 3.339767727538479,
      "tokens_seen": 3113877504
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.705115346038114e-05,
      "loss": 1.3252,
      "theoretical_loss": 3.339762409752381,
      "tokens_seen": 3113943040
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.703109327983952e-05,
      "loss": 1.2571,
      "theoretical_loss": 3.3397570921095356,
      "tokens_seen": 3114008576
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.701103309929789e-05,
      "loss": 1.4015,
      "theoretical_loss": 3.3397517746099363,
      "tokens_seen": 3114074112
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.699097291875627e-05,
      "loss": 1.2625,
      "theoretical_loss": 3.339746457253577,
      "tokens_seen": 3114139648
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6970912738214645e-05,
      "loss": 1.2543,
      "theoretical_loss": 3.33974114004045,
      "tokens_seen": 3114205184
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.695085255767302e-05,
      "loss": 1.3036,
      "theoretical_loss": 3.3397358229705487,
      "tokens_seen": 3114270720
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6930792377131396e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.3397305060438667,
      "tokens_seen": 3114336256
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.691073219658977e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.339725189260396,
      "tokens_seen": 3114401792
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.689067201604815e-05,
      "loss": 1.1943,
      "theoretical_loss": 3.3397198726201305,
      "tokens_seen": 3114467328
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6870611835506524e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.3397145561230634,
      "tokens_seen": 3114532864
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.685055165496489e-05,
      "loss": 1.207,
      "theoretical_loss": 3.3397092397691877,
      "tokens_seen": 3114598400
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.683049147442327e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.339703923558496,
      "tokens_seen": 3114663936
    },
    {
      "epoch": 0.89,
      "objective/train/docs_used": 1131665,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2482120990753174,
      "objective/train/theoretical_loss": 3.3396986074909822,
      "objective/train/tokens_used": 1485188576,
      "theoretical_loss": 3.3396986074909822,
      "tokens_seen": 3114729472
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6810431293881644e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.3396986074909822,
      "tokens_seen": 3114729472
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.679037111334002e-05,
      "loss": 1.1763,
      "theoretical_loss": 3.339693291566639,
      "tokens_seen": 3114795008
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.67703109327984e-05,
      "loss": 1.2199,
      "theoretical_loss": 3.33968797578546,
      "tokens_seen": 3114860544
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.675025075225677e-05,
      "loss": 1.2279,
      "theoretical_loss": 3.3396826601474374,
      "tokens_seen": 3114926080
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.673019057171515e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.339677344652565,
      "tokens_seen": 3114991616
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.671013039117352e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.3396720293008366,
      "tokens_seen": 3115057152
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.66900702106319e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.339666714092244,
      "tokens_seen": 3115122688
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6670010030090275e-05,
      "loss": 1.2063,
      "theoretical_loss": 3.339661399026781,
      "tokens_seen": 3115188224
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6649949849548644e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.3396560841044405,
      "tokens_seen": 3115253760
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.662988966900702e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.3396507693252158,
      "tokens_seen": 3115319296
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6609829488465395e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.3396454546891006,
      "tokens_seen": 3115384832
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.658976930792377e-05,
      "loss": 1.2183,
      "theoretical_loss": 3.339640140196087,
      "tokens_seen": 3115450368
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6569709127382154e-05,
      "loss": 1.3358,
      "theoretical_loss": 3.3396348258461686,
      "tokens_seen": 3115515904
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.654964894684052e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.3396295116393384,
      "tokens_seen": 3115581440
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.65295887662989e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.33962419757559,
      "tokens_seen": 3115646976
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6509528585757274e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.3396188836549157,
      "tokens_seen": 3115712512
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.648946840521565e-05,
      "loss": 1.2258,
      "theoretical_loss": 3.3396135698773097,
      "tokens_seen": 3115778048
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.646940822467402e-05,
      "loss": 1.2355,
      "theoretical_loss": 3.339608256242764,
      "tokens_seen": 3115843584
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6449348044132395e-05,
      "loss": 1.3547,
      "theoretical_loss": 3.339602942751273,
      "tokens_seen": 3115909120
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.642928786359077e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.339597629402829,
      "tokens_seen": 3115974656
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6409227683049147e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.339592316197425,
      "tokens_seen": 3116040192
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.638916750250753e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.3395870031350547,
      "tokens_seen": 3116105728
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.63691073219659e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.3395816902157107,
      "tokens_seen": 3116171264
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6349047141424274e-05,
      "loss": 1.2642,
      "theoretical_loss": 3.3395763774393865,
      "tokens_seen": 3116236800
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.632898696088265e-05,
      "loss": 1.2901,
      "theoretical_loss": 3.3395710648060755,
      "tokens_seen": 3116302336
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6308926780341025e-05,
      "loss": 1.2017,
      "theoretical_loss": 3.33956575231577,
      "tokens_seen": 3116367872
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.62888665997994e-05,
      "loss": 1.3384,
      "theoretical_loss": 3.339560439968464,
      "tokens_seen": 3116433408
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.626880641925777e-05,
      "loss": 1.2243,
      "theoretical_loss": 3.3395551277641506,
      "tokens_seen": 3116498944
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6248746238716146e-05,
      "loss": 1.3302,
      "theoretical_loss": 3.3395498157028225,
      "tokens_seen": 3116564480
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.622868605817452e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.339544503784473,
      "tokens_seen": 3116630016
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6208625877632904e-05,
      "loss": 1.3954,
      "theoretical_loss": 3.3395391920090947,
      "tokens_seen": 3116695552
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.618856569709128e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.3395338803766816,
      "tokens_seen": 3116761088
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.616850551654965e-05,
      "loss": 1.4061,
      "theoretical_loss": 3.3395285688872267,
      "tokens_seen": 3116826624
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6148445336008025e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3395232575407228,
      "tokens_seen": 3116892160
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.61283851554664e-05,
      "loss": 1.1868,
      "theoretical_loss": 3.339517946337163,
      "tokens_seen": 3116957696
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.610832497492478e-05,
      "loss": 1.3542,
      "theoretical_loss": 3.339512635276541,
      "tokens_seen": 3117023232
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.608826479438315e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.33950732435885,
      "tokens_seen": 3117088768
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.606820461384152e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.339502013584082,
      "tokens_seen": 3117154304
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.60481444332999e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.339496702952231,
      "tokens_seen": 3117219840
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.602808425275827e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.3394913924632905,
      "tokens_seen": 3117285376
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6008024072216656e-05,
      "loss": 1.2231,
      "theoretical_loss": 3.3394860821172534,
      "tokens_seen": 3117350912
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.598796389167503e-05,
      "loss": 1.2309,
      "theoretical_loss": 3.339480771914112,
      "tokens_seen": 3117416448
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.59679037111334e-05,
      "loss": 1.3296,
      "theoretical_loss": 3.3394754618538607,
      "tokens_seen": 3117481984
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5947843530591776e-05,
      "loss": 1.3286,
      "theoretical_loss": 3.3394701519364918,
      "tokens_seen": 3117547520
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.592778335005015e-05,
      "loss": 1.3183,
      "theoretical_loss": 3.3394648421619983,
      "tokens_seen": 3117613056
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.590772316950853e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.3394595325303746,
      "tokens_seen": 3117678592
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5887662988966904e-05,
      "loss": 1.3293,
      "theoretical_loss": 3.3394542230416127,
      "tokens_seen": 3117744128
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.586760280842527e-05,
      "loss": 1.3526,
      "theoretical_loss": 3.3394489136957057,
      "tokens_seen": 3117809664
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.584754262788365e-05,
      "loss": 1.2459,
      "theoretical_loss": 3.3394436044926477,
      "tokens_seen": 3117875200
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.582748244734203e-05,
      "loss": 1.3964,
      "theoretical_loss": 3.339438295432431,
      "tokens_seen": 3117940736
    },
    {
      "epoch": 0.89,
      "objective/train/docs_used": 1132901,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4164438247680664,
      "objective/train/theoretical_loss": 3.3394329865150487,
      "objective/train/tokens_used": 1488465376,
      "theoretical_loss": 3.3394329865150487,
      "tokens_seen": 3118006272
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.580742226680041e-05,
      "loss": 1.3957,
      "theoretical_loss": 3.3394329865150487,
      "tokens_seen": 3118006272
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5787362086258776e-05,
      "loss": 1.3443,
      "theoretical_loss": 3.3394276777404945,
      "tokens_seen": 3118071808
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.576730190571715e-05,
      "loss": 1.3855,
      "theoretical_loss": 3.3394223691087617,
      "tokens_seen": 3118137344
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.574724172517553e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.339417060619843,
      "tokens_seen": 3118202880
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.57271815446339e-05,
      "loss": 1.2613,
      "theoretical_loss": 3.3394117522737314,
      "tokens_seen": 3118268416
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.570712136409228e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.3394064440704208,
      "tokens_seen": 3118333952
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.568706118355065e-05,
      "loss": 1.3286,
      "theoretical_loss": 3.3394011360099034,
      "tokens_seen": 3118399488
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5667001003009024e-05,
      "loss": 1.2758,
      "theoretical_loss": 3.339395828092173,
      "tokens_seen": 3118465024
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5646940822467406e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.339390520317223,
      "tokens_seen": 3118530560
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.562688064192578e-05,
      "loss": 1.2574,
      "theoretical_loss": 3.3393852126850456,
      "tokens_seen": 3118596096
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.560682046138416e-05,
      "loss": 1.3733,
      "theoretical_loss": 3.3393799051956345,
      "tokens_seen": 3118661632
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.558676028084253e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.339374597848983,
      "tokens_seen": 3118727168
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.55667001003009e-05,
      "loss": 1.345,
      "theoretical_loss": 3.3393692906450845,
      "tokens_seen": 3118792704
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.554663991975928e-05,
      "loss": 1.4749,
      "theoretical_loss": 3.3393639835839313,
      "tokens_seen": 3118858240
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5526579739217654e-05,
      "loss": 1.2123,
      "theoretical_loss": 3.3393586766655172,
      "tokens_seen": 3118923776
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.550651955867603e-05,
      "loss": 1.3549,
      "theoretical_loss": 3.339353369889835,
      "tokens_seen": 3118989312
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.54864593781344e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.3393480632568786,
      "tokens_seen": 3119054848
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5466399197592775e-05,
      "loss": 1.2307,
      "theoretical_loss": 3.3393427567666403,
      "tokens_seen": 3119120384
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.544633901705116e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.339337450419114,
      "tokens_seen": 3119185920
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.542627883650953e-05,
      "loss": 1.3645,
      "theoretical_loss": 3.339332144214292,
      "tokens_seen": 3119251456
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.540621865596791e-05,
      "loss": 1.2856,
      "theoretical_loss": 3.339326838152168,
      "tokens_seen": 3119316992
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.538615847542628e-05,
      "loss": 1.2573,
      "theoretical_loss": 3.339321532232735,
      "tokens_seen": 3119382528
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5366098294884654e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.3393162264559866,
      "tokens_seen": 3119448064
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.534603811434303e-05,
      "loss": 1.1851,
      "theoretical_loss": 3.3393109208219154,
      "tokens_seen": 3119513600
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5325977933801405e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.339305615330515,
      "tokens_seen": 3119579136
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.530591775325978e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.339300309981778,
      "tokens_seen": 3119644672
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.528585757271815e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.3392950047756984,
      "tokens_seen": 3119710208
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.526579739217653e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.3392896997122685,
      "tokens_seen": 3119775744
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.524573721163491e-05,
      "loss": 1.2185,
      "theoretical_loss": 3.3392843947914823,
      "tokens_seen": 3119841280
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5225677031093284e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.3392790900133322,
      "tokens_seen": 3119906816
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.520561685055166e-05,
      "loss": 1.3362,
      "theoretical_loss": 3.3392737853778116,
      "tokens_seen": 3119972352
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.518555667001003e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.339268480884914,
      "tokens_seen": 3120037888
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5165496489468405e-05,
      "loss": 1.3987,
      "theoretical_loss": 3.339263176534632,
      "tokens_seen": 3120103424
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.514543630892678e-05,
      "loss": 1.337,
      "theoretical_loss": 3.3392578723269595,
      "tokens_seen": 3120168960
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.512537612838516e-05,
      "loss": 1.3798,
      "theoretical_loss": 3.339252568261889,
      "tokens_seen": 3120234496
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5105315947843526e-05,
      "loss": 1.2024,
      "theoretical_loss": 3.339247264339414,
      "tokens_seen": 3120300032
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.508525576730191e-05,
      "loss": 1.3523,
      "theoretical_loss": 3.339241960559528,
      "tokens_seen": 3120365568
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5065195586760284e-05,
      "loss": 1.1827,
      "theoretical_loss": 3.339236656922224,
      "tokens_seen": 3120431104
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.504513540621866e-05,
      "loss": 1.3958,
      "theoretical_loss": 3.339231353427494,
      "tokens_seen": 3120496640
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5025075225677036e-05,
      "loss": 1.3365,
      "theoretical_loss": 3.339226050075333,
      "tokens_seen": 3120562176
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5005015045135405e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.3392207468657324,
      "tokens_seen": 3120627712
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.498495486459378e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.339215443798687,
      "tokens_seen": 3120693248
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4964894684052156e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.339210140874189,
      "tokens_seen": 3120758784
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.494483450351053e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.339204838092232,
      "tokens_seen": 3120824320
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.492477432296891e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.339199535452809,
      "tokens_seen": 3120889856
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.490471414242728e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.339194232955913,
      "tokens_seen": 3120955392
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.488465396188566e-05,
      "loss": 1.23,
      "theoretical_loss": 3.3391889306015377,
      "tokens_seen": 3121020928
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4864593781344035e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.3391836283896756,
      "tokens_seen": 3121086464
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.484453360080241e-05,
      "loss": 1.3373,
      "theoretical_loss": 3.3391783263203205,
      "tokens_seen": 3121152000
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.482447342026079e-05,
      "loss": 1.3693,
      "theoretical_loss": 3.3391730243934648,
      "tokens_seen": 3121217536
    },
    {
      "epoch": 0.89,
      "objective/train/docs_used": 1133992,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2980782985687256,
      "objective/train/theoretical_loss": 3.339167722609103,
      "objective/train/tokens_used": 1491742176,
      "theoretical_loss": 3.339167722609103,
      "tokens_seen": 3121283072
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4804413239719156e-05,
      "loss": 1.2572,
      "theoretical_loss": 3.339167722609103,
      "tokens_seen": 3121283072
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.478435305917753e-05,
      "loss": 1.3695,
      "theoretical_loss": 3.339162420967227,
      "tokens_seen": 3121348608
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.476429287863591e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.33915711946783,
      "tokens_seen": 3121414144
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.474423269809428e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.3391518181109063,
      "tokens_seen": 3121479680
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.472417251755266e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.339146516896448,
      "tokens_seen": 3121545216
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4704112337011035e-05,
      "loss": 1.4929,
      "theoretical_loss": 3.3391412158244487,
      "tokens_seen": 3121610752
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.468405215646941e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.339135914894902,
      "tokens_seen": 3121676288
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4663991975927786e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.3391306141078,
      "tokens_seen": 3121741824
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.464393179538616e-05,
      "loss": 1.2193,
      "theoretical_loss": 3.339125313463137,
      "tokens_seen": 3121807360
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.462387161484454e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.3391200129609055,
      "tokens_seen": 3121872896
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.460381143430291e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.3391147126010985,
      "tokens_seen": 3121938432
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.458375125376128e-05,
      "loss": 1.3236,
      "theoretical_loss": 3.33910941238371,
      "tokens_seen": 3122003968
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.456369107321966e-05,
      "loss": 1.2004,
      "theoretical_loss": 3.339104112308733,
      "tokens_seen": 3122069504
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4543630892678034e-05,
      "loss": 1.2254,
      "theoretical_loss": 3.33909881237616,
      "tokens_seen": 3122135040
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.452357071213641e-05,
      "loss": 1.4918,
      "theoretical_loss": 3.3390935125859844,
      "tokens_seen": 3122200576
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4503510531594786e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.3390882129382002,
      "tokens_seen": 3122266112
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.448345035105316e-05,
      "loss": 1.2232,
      "theoretical_loss": 3.3390829134328,
      "tokens_seen": 3122331648
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.446339017051154e-05,
      "loss": 1.3578,
      "theoretical_loss": 3.3390776140697764,
      "tokens_seen": 3122397184
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.444332998996991e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.3390723148491235,
      "tokens_seen": 3122462720
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.442326980942829e-05,
      "loss": 1.3346,
      "theoretical_loss": 3.3390670157708344,
      "tokens_seen": 3122528256
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.440320962888666e-05,
      "loss": 1.3279,
      "theoretical_loss": 3.3390617168349017,
      "tokens_seen": 3122593792
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4383149448345034e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.3390564180413187,
      "tokens_seen": 3122659328
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.436308926780341e-05,
      "loss": 1.2622,
      "theoretical_loss": 3.339051119390079,
      "tokens_seen": 3122724864
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4343029087261785e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.339045820881176,
      "tokens_seen": 3122790400
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.432296890672016e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.339040522514602,
      "tokens_seen": 3122855936
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.430290872617854e-05,
      "loss": 1.3958,
      "theoretical_loss": 3.339035224290351,
      "tokens_seen": 3122921472
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.428284854563691e-05,
      "loss": 1.2865,
      "theoretical_loss": 3.3390299262084158,
      "tokens_seen": 3122987008
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.426278836509529e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.3390246282687897,
      "tokens_seen": 3123052544
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4242728184553664e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.3390193304714657,
      "tokens_seen": 3123118080
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4222668004012033e-05,
      "loss": 1.3067,
      "theoretical_loss": 3.3390140328164373,
      "tokens_seen": 3123183616
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.420260782347041e-05,
      "loss": 1.254,
      "theoretical_loss": 3.3390087353036972,
      "tokens_seen": 3123249152
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4182547642928785e-05,
      "loss": 1.2016,
      "theoretical_loss": 3.3390034379332394,
      "tokens_seen": 3123314688
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.416248746238716e-05,
      "loss": 1.3223,
      "theoretical_loss": 3.338998140705056,
      "tokens_seen": 3123380224
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4142427281845543e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.338992843619142,
      "tokens_seen": 3123445760
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.412236710130391e-05,
      "loss": 1.2281,
      "theoretical_loss": 3.3389875466754884,
      "tokens_seen": 3123511296
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.410230692076229e-05,
      "loss": 1.24,
      "theoretical_loss": 3.3389822498740895,
      "tokens_seen": 3123576832
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4082246740220664e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.3389769532149387,
      "tokens_seen": 3123642368
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.406218655967904e-05,
      "loss": 1.4761,
      "theoretical_loss": 3.3389716566980283,
      "tokens_seen": 3123707904
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4042126379137416e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3389663603233526,
      "tokens_seen": 3123773440
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4022066198595785e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.3389610640909044,
      "tokens_seen": 3123838976
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.400200601805416e-05,
      "loss": 1.2429,
      "theoretical_loss": 3.3389557680006767,
      "tokens_seen": 3123904512
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3981945837512536e-05,
      "loss": 1.3699,
      "theoretical_loss": 3.338950472052663,
      "tokens_seen": 3123970048
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.396188565697091e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.3389451762468556,
      "tokens_seen": 3124035584
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3941825476429295e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.338939880583249,
      "tokens_seen": 3124101120
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3921765295887664e-05,
      "loss": 1.3569,
      "theoretical_loss": 3.3389345850618355,
      "tokens_seen": 3124166656
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.390170511534604e-05,
      "loss": 1.1049,
      "theoretical_loss": 3.3389292896826084,
      "tokens_seen": 3124232192
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3881644934804415e-05,
      "loss": 1.214,
      "theoretical_loss": 3.3389239944455613,
      "tokens_seen": 3124297728
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.386158475426279e-05,
      "loss": 1.2059,
      "theoretical_loss": 3.338918699350687,
      "tokens_seen": 3124363264
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.384152457372117e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.3389134043979793,
      "tokens_seen": 3124428800
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3821464393179536e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.3389081095874307,
      "tokens_seen": 3124494336
    },
    {
      "epoch": 0.89,
      "objective/train/docs_used": 1135223,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2332755327224731,
      "objective/train/theoretical_loss": 3.338902814919035,
      "objective/train/tokens_used": 1495018976,
      "theoretical_loss": 3.338902814919035,
      "tokens_seen": 3124559872
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.380140421263791e-05,
      "loss": 1.2549,
      "theoretical_loss": 3.338902814919035,
      "tokens_seen": 3124559872
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.378134403209629e-05,
      "loss": 1.228,
      "theoretical_loss": 3.3388975203927846,
      "tokens_seen": 3124625408
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.376128385155467e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.3388922260086735,
      "tokens_seen": 3124690944
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3741223671013046e-05,
      "loss": 1.1697,
      "theoretical_loss": 3.3388869317666945,
      "tokens_seen": 3124756480
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3721163490471415e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.338881637666841,
      "tokens_seen": 3124822016
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.370110330992979e-05,
      "loss": 1.2027,
      "theoretical_loss": 3.338876343709106,
      "tokens_seen": 3124887552
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3681043129388166e-05,
      "loss": 1.3303,
      "theoretical_loss": 3.3388710498934824,
      "tokens_seen": 3124953088
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.366098294884654e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.338865756219964,
      "tokens_seen": 3125018624
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.364092276830491e-05,
      "loss": 1.2024,
      "theoretical_loss": 3.3388604626885448,
      "tokens_seen": 3125084160
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.362086258776329e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.3388551692992157,
      "tokens_seen": 3125149696
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.360080240722166e-05,
      "loss": 1.3154,
      "theoretical_loss": 3.338849876051972,
      "tokens_seen": 3125215232
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3580742226680045e-05,
      "loss": 1.3403,
      "theoretical_loss": 3.338844582946806,
      "tokens_seen": 3125280768
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.356068204613842e-05,
      "loss": 1.3437,
      "theoretical_loss": 3.338839289983711,
      "tokens_seen": 3125346304
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.354062186559679e-05,
      "loss": 1.251,
      "theoretical_loss": 3.3388339971626797,
      "tokens_seen": 3125411840
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3520561685055166e-05,
      "loss": 1.2187,
      "theoretical_loss": 3.3388287044837064,
      "tokens_seen": 3125477376
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.350050150451354e-05,
      "loss": 1.3163,
      "theoretical_loss": 3.338823411946784,
      "tokens_seen": 3125542912
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.348044132397192e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.338818119551905,
      "tokens_seen": 3125608448
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.346038114343029e-05,
      "loss": 1.2459,
      "theoretical_loss": 3.338812827299063,
      "tokens_seen": 3125673984
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.344032096288866e-05,
      "loss": 1.2299,
      "theoretical_loss": 3.338807535188252,
      "tokens_seen": 3125739520
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.342026078234704e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.3388022432194635,
      "tokens_seen": 3125805056
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3400200601805414e-05,
      "loss": 1.243,
      "theoretical_loss": 3.3387969513926925,
      "tokens_seen": 3125870592
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3380140421263796e-05,
      "loss": 1.2342,
      "theoretical_loss": 3.338791659707931,
      "tokens_seen": 3125936128
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.336008024072217e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.338786368165173,
      "tokens_seen": 3126001664
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.334002006018054e-05,
      "loss": 1.2041,
      "theoretical_loss": 3.338781076764411,
      "tokens_seen": 3126067200
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.331995987963892e-05,
      "loss": 1.2135,
      "theoretical_loss": 3.338775785505639,
      "tokens_seen": 3126132736
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.329989969909729e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.338770494388849,
      "tokens_seen": 3126198272
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.327983951855567e-05,
      "loss": 1.3333,
      "theoretical_loss": 3.338765203414036,
      "tokens_seen": 3126263808
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3259779338014044e-05,
      "loss": 1.2329,
      "theoretical_loss": 3.3387599125811915,
      "tokens_seen": 3126329344
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3239719157472413e-05,
      "loss": 1.3566,
      "theoretical_loss": 3.3387546218903097,
      "tokens_seen": 3126394880
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.321965897693079e-05,
      "loss": 1.2451,
      "theoretical_loss": 3.338749331341383,
      "tokens_seen": 3126460416
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.319959879638917e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.338744040934406,
      "tokens_seen": 3126525952
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.317953861584755e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.3387387506693704,
      "tokens_seen": 3126591488
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.315947843530592e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.3387334605462704,
      "tokens_seen": 3126657024
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.313941825476429e-05,
      "loss": 1.3931,
      "theoretical_loss": 3.3387281705650986,
      "tokens_seen": 3126722560
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.311935807422267e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.3387228807258493,
      "tokens_seen": 3126788096
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.3099297893681044e-05,
      "loss": 1.4309,
      "theoretical_loss": 3.338717591028514,
      "tokens_seen": 3126853632
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.307923771313942e-05,
      "loss": 1.2244,
      "theoretical_loss": 3.3387123014730875,
      "tokens_seen": 3126919168
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.3059177532597796e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.3387070120595617,
      "tokens_seen": 3126984704
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.3039117352056165e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.338701722787931,
      "tokens_seen": 3127050240
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.301905717151455e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.338696433658188,
      "tokens_seen": 3127115776
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.299899699097292e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.3386911446703262,
      "tokens_seen": 3127181312
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.29789368104313e-05,
      "loss": 1.3906,
      "theoretical_loss": 3.3386858558243384,
      "tokens_seen": 3127246848
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2958876629889675e-05,
      "loss": 1.2248,
      "theoretical_loss": 3.3386805671202184,
      "tokens_seen": 3127312384
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2938816449348044e-05,
      "loss": 1.3297,
      "theoretical_loss": 3.3386752785579588,
      "tokens_seen": 3127377920
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.291875626880642e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.338669990137553,
      "tokens_seen": 3127443456
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2898696088264795e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.3386647018589946,
      "tokens_seen": 3127508992
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.287863590772317e-05,
      "loss": 1.2924,
      "theoretical_loss": 3.3386594137222767,
      "tokens_seen": 3127574528
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.285857572718154e-05,
      "loss": 1.2733,
      "theoretical_loss": 3.338654125727392,
      "tokens_seen": 3127640064
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2838515546639916e-05,
      "loss": 1.3411,
      "theoretical_loss": 3.338648837874334,
      "tokens_seen": 3127705600
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.28184553660983e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.3386435501630967,
      "tokens_seen": 3127771136
    },
    {
      "epoch": 0.9,
      "objective/train/docs_used": 1136308,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3864843845367432,
      "objective/train/theoretical_loss": 3.338638262593672,
      "objective/train/tokens_used": 1498295776,
      "theoretical_loss": 3.338638262593672,
      "tokens_seen": 3127836672
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2798395185556674e-05,
      "loss": 1.4017,
      "theoretical_loss": 3.338638262593672,
      "tokens_seen": 3127836672
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.277833500501505e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.338632975166054,
      "tokens_seen": 3127902208
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.275827482447342e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.338627687880236,
      "tokens_seen": 3127967744
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2738214643931795e-05,
      "loss": 1.1933,
      "theoretical_loss": 3.3386224007362104,
      "tokens_seen": 3128033280
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.271815446339017e-05,
      "loss": 1.2099,
      "theoretical_loss": 3.3386171137339713,
      "tokens_seen": 3128098816
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2698094282848546e-05,
      "loss": 1.2349,
      "theoretical_loss": 3.3386118268735117,
      "tokens_seen": 3128164352
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.267803410230692e-05,
      "loss": 1.2314,
      "theoretical_loss": 3.3386065401548244,
      "tokens_seen": 3128229888
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.265797392176529e-05,
      "loss": 1.1569,
      "theoretical_loss": 3.3386012535779033,
      "tokens_seen": 3128295424
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2637913741223674e-05,
      "loss": 1.3129,
      "theoretical_loss": 3.338595967142741,
      "tokens_seen": 3128360960
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.261785356068205e-05,
      "loss": 1.3413,
      "theoretical_loss": 3.338590680849331,
      "tokens_seen": 3128426496
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2597793380140425e-05,
      "loss": 1.3588,
      "theoretical_loss": 3.3385853946976667,
      "tokens_seen": 3128492032
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.25777331995988e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.3385801086877414,
      "tokens_seen": 3128557568
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.255767301905717e-05,
      "loss": 1.2002,
      "theoretical_loss": 3.3385748228195475,
      "tokens_seen": 3128623104
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2537612838515546e-05,
      "loss": 1.2857,
      "theoretical_loss": 3.338569537093079,
      "tokens_seen": 3128688640
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.251755265797392e-05,
      "loss": 1.1989,
      "theoretical_loss": 3.338564251508329,
      "tokens_seen": 3128754176
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.24974924774323e-05,
      "loss": 1.2785,
      "theoretical_loss": 3.338558966065291,
      "tokens_seen": 3128819712
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.247743229689067e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.3385536807639578,
      "tokens_seen": 3128885248
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.245737211634905e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.3385483956043225,
      "tokens_seen": 3128950784
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2437311935807425e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.3385431105863788,
      "tokens_seen": 3129016320
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.24172517552658e-05,
      "loss": 1.3517,
      "theoretical_loss": 3.3385378257101195,
      "tokens_seen": 3129081856
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2397191574724176e-05,
      "loss": 1.258,
      "theoretical_loss": 3.3385325409755384,
      "tokens_seen": 3129147392
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.237713139418255e-05,
      "loss": 1.1842,
      "theoretical_loss": 3.338527256382628,
      "tokens_seen": 3129212928
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.235707121364092e-05,
      "loss": 1.2903,
      "theoretical_loss": 3.338521971931382,
      "tokens_seen": 3129278464
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.23370110330993e-05,
      "loss": 1.1866,
      "theoretical_loss": 3.338516687621794,
      "tokens_seen": 3129344000
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.231695085255767e-05,
      "loss": 1.2388,
      "theoretical_loss": 3.3385114034538566,
      "tokens_seen": 3129409536
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.229689067201605e-05,
      "loss": 1.2309,
      "theoretical_loss": 3.338506119427563,
      "tokens_seen": 3129475072
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2276830491474424e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.3385008355429067,
      "tokens_seen": 3129540608
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.22567703109328e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.3384955517998813,
      "tokens_seen": 3129606144
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2236710130391176e-05,
      "loss": 1.2108,
      "theoretical_loss": 3.338490268198479,
      "tokens_seen": 3129671680
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.221664994984955e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.3384849847386944,
      "tokens_seen": 3129737216
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.219658976930793e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.3384797014205194,
      "tokens_seen": 3129802752
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2176529588766297e-05,
      "loss": 1.33,
      "theoretical_loss": 3.338474418243948,
      "tokens_seen": 3129868288
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.215646940822467e-05,
      "loss": 1.4026,
      "theoretical_loss": 3.3384691352089737,
      "tokens_seen": 3129933824
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.213640922768305e-05,
      "loss": 1.369,
      "theoretical_loss": 3.338463852315589,
      "tokens_seen": 3129999360
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2116349047141424e-05,
      "loss": 1.3477,
      "theoretical_loss": 3.3384585695637874,
      "tokens_seen": 3130064896
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.20962888665998e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.3384532869535626,
      "tokens_seen": 3130130432
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2076228686058176e-05,
      "loss": 1.1733,
      "theoretical_loss": 3.338448004484907,
      "tokens_seen": 3130195968
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.205616850551655e-05,
      "loss": 1.3776,
      "theoretical_loss": 3.338442722157815,
      "tokens_seen": 3130261504
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.203610832497493e-05,
      "loss": 1.2528,
      "theoretical_loss": 3.3384374399722785,
      "tokens_seen": 3130327040
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.20160481444333e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.338432157928292,
      "tokens_seen": 3130392576
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.199598796389168e-05,
      "loss": 1.4126,
      "theoretical_loss": 3.3384268760258475,
      "tokens_seen": 3130458112
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.197592778335005e-05,
      "loss": 1.2297,
      "theoretical_loss": 3.338421594264939,
      "tokens_seen": 3130523648
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1955867602808424e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.3384163126455597,
      "tokens_seen": 3130589184
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.19358074222668e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.338411031167703,
      "tokens_seen": 3130654720
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1915747241725175e-05,
      "loss": 1.1913,
      "theoretical_loss": 3.3384057498313617,
      "tokens_seen": 3130720256
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.189568706118355e-05,
      "loss": 1.3617,
      "theoretical_loss": 3.338400468636529,
      "tokens_seen": 3130785792
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.187562688064193e-05,
      "loss": 1.2078,
      "theoretical_loss": 3.338395187583199,
      "tokens_seen": 3130851328
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.18555667001003e-05,
      "loss": 1.2324,
      "theoretical_loss": 3.338389906671364,
      "tokens_seen": 3130916864
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.183550651955868e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.3383846259010177,
      "tokens_seen": 3130982400
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1815446339017054e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.3383793452721533,
      "tokens_seen": 3131047936
    },
    {
      "epoch": 0.9,
      "objective/train/docs_used": 1137606,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1512203216552734,
      "objective/train/theoretical_loss": 3.338374064784764,
      "objective/train/tokens_used": 1501572576,
      "theoretical_loss": 3.338374064784764,
      "tokens_seen": 3131113472
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.179538615847543e-05,
      "loss": 1.1797,
      "theoretical_loss": 3.338374064784764,
      "tokens_seen": 3131113472
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.17753259779338e-05,
      "loss": 1.2284,
      "theoretical_loss": 3.3383687844388428,
      "tokens_seen": 3131179008
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1755265797392175e-05,
      "loss": 1.2421,
      "theoretical_loss": 3.3383635042343833,
      "tokens_seen": 3131244544
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.173520561685055e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.3383582241713787,
      "tokens_seen": 3131310080
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1715145436308926e-05,
      "loss": 1.288,
      "theoretical_loss": 3.3383529442498228,
      "tokens_seen": 3131375616
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.169508525576731e-05,
      "loss": 1.2117,
      "theoretical_loss": 3.3383476644697074,
      "tokens_seen": 3131441152
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.167502507522568e-05,
      "loss": 1.1477,
      "theoretical_loss": 3.338342384831027,
      "tokens_seen": 3131506688
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1654964894684054e-05,
      "loss": 1.3232,
      "theoretical_loss": 3.3383371053337743,
      "tokens_seen": 3131572224
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.163490471414243e-05,
      "loss": 1.268,
      "theoretical_loss": 3.3383318259779426,
      "tokens_seen": 3131637760
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1614844533600805e-05,
      "loss": 1.2622,
      "theoretical_loss": 3.3383265467635255,
      "tokens_seen": 3131703296
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.159478435305918e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.3383212676905156,
      "tokens_seen": 3131768832
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.157472417251755e-05,
      "loss": 1.1975,
      "theoretical_loss": 3.3383159887589073,
      "tokens_seen": 3131834368
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1554663991975926e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.3383107099686926,
      "tokens_seen": 3131899904
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.15346038114343e-05,
      "loss": 1.2132,
      "theoretical_loss": 3.3383054313198652,
      "tokens_seen": 3131965440
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1514543630892684e-05,
      "loss": 1.1878,
      "theoretical_loss": 3.3383001528124185,
      "tokens_seen": 3132030976
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.149448345035106e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.338294874446346,
      "tokens_seen": 3132096512
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.147442326980943e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.3382895962216406,
      "tokens_seen": 3132162048
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1454363089267805e-05,
      "loss": 1.2192,
      "theoretical_loss": 3.338284318138295,
      "tokens_seen": 3132227584
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.143430290872618e-05,
      "loss": 1.2537,
      "theoretical_loss": 3.3382790401963036,
      "tokens_seen": 3132293120
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1414242728184556e-05,
      "loss": 1.2361,
      "theoretical_loss": 3.3382737623956587,
      "tokens_seen": 3132358656
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1394182547642925e-05,
      "loss": 1.2377,
      "theoretical_loss": 3.3382684847363544,
      "tokens_seen": 3132424192
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.13741223671013e-05,
      "loss": 1.2989,
      "theoretical_loss": 3.3382632072183833,
      "tokens_seen": 3132489728
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.135406218655968e-05,
      "loss": 1.2481,
      "theoretical_loss": 3.3382579298417387,
      "tokens_seen": 3132555264
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.133400200601805e-05,
      "loss": 1.141,
      "theoretical_loss": 3.3382526526064145,
      "tokens_seen": 3132620800
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1313941825476435e-05,
      "loss": 1.3588,
      "theoretical_loss": 3.338247375512403,
      "tokens_seen": 3132686336
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1293881644934804e-05,
      "loss": 1.1879,
      "theoretical_loss": 3.3382420985596983,
      "tokens_seen": 3132751872
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.127382146439318e-05,
      "loss": 1.4233,
      "theoretical_loss": 3.338236821748293,
      "tokens_seen": 3132817408
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1253761283851556e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.338231545078181,
      "tokens_seen": 3132882944
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.123370110330993e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.338226268549355,
      "tokens_seen": 3132948480
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.121364092276831e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.338220992161809,
      "tokens_seen": 3133014016
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1193580742226677e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.3382157159155352,
      "tokens_seen": 3133079552
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.117352056168505e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.3382104398105277,
      "tokens_seen": 3133145088
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.115346038114343e-05,
      "loss": 1.2118,
      "theoretical_loss": 3.338205163846779,
      "tokens_seen": 3133210624
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.113340020060181e-05,
      "loss": 1.4645,
      "theoretical_loss": 3.3381998880242834,
      "tokens_seen": 3133276160
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1113340020060187e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.3381946123430337,
      "tokens_seen": 3133341696
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1093279839518556e-05,
      "loss": 1.332,
      "theoretical_loss": 3.3381893368030227,
      "tokens_seen": 3133407232
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.107321965897693e-05,
      "loss": 1.2453,
      "theoretical_loss": 3.338184061404244,
      "tokens_seen": 3133472768
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.105315947843531e-05,
      "loss": 1.1524,
      "theoretical_loss": 3.338178786146691,
      "tokens_seen": 3133538304
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.103309929789368e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.338173511030357,
      "tokens_seen": 3133603840
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.101303911735206e-05,
      "loss": 1.3605,
      "theoretical_loss": 3.3381682360552354,
      "tokens_seen": 3133669376
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.099297893681043e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.3381629612213186,
      "tokens_seen": 3133734912
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0972918756268804e-05,
      "loss": 1.4332,
      "theoretical_loss": 3.338157686528601,
      "tokens_seen": 3133800448
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0952858575727186e-05,
      "loss": 1.3297,
      "theoretical_loss": 3.338152411977075,
      "tokens_seen": 3133865984
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.093279839518556e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.338147137566734,
      "tokens_seen": 3133931520
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.091273821464394e-05,
      "loss": 1.3529,
      "theoretical_loss": 3.3381418632975723,
      "tokens_seen": 3133997056
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.089267803410231e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.3381365891695816,
      "tokens_seen": 3134062592
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.087261785356068e-05,
      "loss": 1.2445,
      "theoretical_loss": 3.338131315182756,
      "tokens_seen": 3134128128
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.085255767301906e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.3381260413370892,
      "tokens_seen": 3134193664
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0832497492477434e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.3381207676325735,
      "tokens_seen": 3134259200
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.081243731193581e-05,
      "loss": 1.3216,
      "theoretical_loss": 3.3381154940692026,
      "tokens_seen": 3134324736
    },
    {
      "epoch": 0.9,
      "objective/train/docs_used": 1138754,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4435052871704102,
      "objective/train/theoretical_loss": 3.3381102206469704,
      "objective/train/tokens_used": 1504849376,
      "theoretical_loss": 3.3381102206469704,
      "tokens_seen": 3134390272
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.079237713139418e-05,
      "loss": 1.3679,
      "theoretical_loss": 3.3381102206469704,
      "tokens_seen": 3134390272
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0772316950852555e-05,
      "loss": 1.3588,
      "theoretical_loss": 3.338104947365869,
      "tokens_seen": 3134455808
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.075225677031094e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.338099674225892,
      "tokens_seen": 3134521344
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.073219658976931e-05,
      "loss": 1.2055,
      "theoretical_loss": 3.3380944012270337,
      "tokens_seen": 3134586880
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.071213640922768e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.338089128369286,
      "tokens_seen": 3134652416
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.069207622868606e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.338083855652643,
      "tokens_seen": 3134717952
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0672016048144434e-05,
      "loss": 1.2353,
      "theoretical_loss": 3.3380785830770976,
      "tokens_seen": 3134783488
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.065195586760281e-05,
      "loss": 1.1747,
      "theoretical_loss": 3.3380733106426432,
      "tokens_seen": 3134849024
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0631895687061185e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.3380680383492733,
      "tokens_seen": 3134914560
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0611835506519554e-05,
      "loss": 1.4633,
      "theoretical_loss": 3.338062766196981,
      "tokens_seen": 3134980096
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.059177532597793e-05,
      "loss": 1.2252,
      "theoretical_loss": 3.338057494185759,
      "tokens_seen": 3135045632
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.057171514543631e-05,
      "loss": 1.3513,
      "theoretical_loss": 3.3380522223156017,
      "tokens_seen": 3135111168
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.055165496489469e-05,
      "loss": 1.3981,
      "theoretical_loss": 3.338046950586502,
      "tokens_seen": 3135176704
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0531594784353064e-05,
      "loss": 1.32,
      "theoretical_loss": 3.338041678998452,
      "tokens_seen": 3135242240
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.051153460381143e-05,
      "loss": 1.2975,
      "theoretical_loss": 3.3380364075514466,
      "tokens_seen": 3135307776
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.049147442326981e-05,
      "loss": 1.2276,
      "theoretical_loss": 3.3380311362454784,
      "tokens_seen": 3135373312
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0471414242728185e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.3380258650805406,
      "tokens_seen": 3135438848
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.045135406218656e-05,
      "loss": 1.401,
      "theoretical_loss": 3.3380205940566268,
      "tokens_seen": 3135504384
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0431293881644936e-05,
      "loss": 1.2158,
      "theoretical_loss": 3.33801532317373,
      "tokens_seen": 3135569920
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0411233701103305e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.3380100524318435,
      "tokens_seen": 3135635456
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.039117352056169e-05,
      "loss": 1.187,
      "theoretical_loss": 3.33800478183096,
      "tokens_seen": 3135700992
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0371113340020064e-05,
      "loss": 1.3466,
      "theoretical_loss": 3.3379995113710743,
      "tokens_seen": 3135766528
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.035105315947844e-05,
      "loss": 1.1589,
      "theoretical_loss": 3.3379942410521783,
      "tokens_seen": 3135832064
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0330992978936815e-05,
      "loss": 1.3644,
      "theoretical_loss": 3.337988970874266,
      "tokens_seen": 3135897600
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0310932798395184e-05,
      "loss": 1.1876,
      "theoretical_loss": 3.3379837008373303,
      "tokens_seen": 3135963136
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.029087261785356e-05,
      "loss": 1.2908,
      "theoretical_loss": 3.3379784309413645,
      "tokens_seen": 3136028672
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0270812437311936e-05,
      "loss": 1.1286,
      "theoretical_loss": 3.3379731611863623,
      "tokens_seen": 3136094208
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.025075225677031e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.3379678915723168,
      "tokens_seen": 3136159744
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.023069207622869e-05,
      "loss": 1.2129,
      "theoretical_loss": 3.337962622099221,
      "tokens_seen": 3136225280
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0210631895687057e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.3379573527670683,
      "tokens_seen": 3136290816
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.019057171514544e-05,
      "loss": 1.2485,
      "theoretical_loss": 3.337952083575852,
      "tokens_seen": 3136356352
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0170511534603815e-05,
      "loss": 1.2041,
      "theoretical_loss": 3.3379468145255657,
      "tokens_seen": 3136421888
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.015045135406219e-05,
      "loss": 1.3232,
      "theoretical_loss": 3.3379415456162023,
      "tokens_seen": 3136487424
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0130391173520567e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.337936276847755,
      "tokens_seen": 3136552960
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0110330992978936e-05,
      "loss": 1.2622,
      "theoretical_loss": 3.3379310082202176,
      "tokens_seen": 3136618496
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.009027081243731e-05,
      "loss": 1.1747,
      "theoretical_loss": 3.337925739733583,
      "tokens_seen": 3136684032
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.007021063189569e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.3379204713878448,
      "tokens_seen": 3136749568
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.005015045135406e-05,
      "loss": 1.1638,
      "theoretical_loss": 3.3379152031829955,
      "tokens_seen": 3136815104
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.003009027081243e-05,
      "loss": 1.1809,
      "theoretical_loss": 3.337909935119029,
      "tokens_seen": 3136880640
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0010030090270814e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.3379046671959394,
      "tokens_seen": 3136946176
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.998996990972919e-05,
      "loss": 1.1885,
      "theoretical_loss": 3.3378993994137183,
      "tokens_seen": 3137011712
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9969909729187566e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.33789413177236,
      "tokens_seen": 3137077248
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.994984954864594e-05,
      "loss": 1.152,
      "theoretical_loss": 3.3378888642718576,
      "tokens_seen": 3137142784
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.992978936810431e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.3378835969122047,
      "tokens_seen": 3137208320
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.990972918756269e-05,
      "loss": 1.2261,
      "theoretical_loss": 3.337878329693394,
      "tokens_seen": 3137273856
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.988966900702106e-05,
      "loss": 1.1775,
      "theoretical_loss": 3.337873062615419,
      "tokens_seen": 3137339392
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.986960882647944e-05,
      "loss": 1.2312,
      "theoretical_loss": 3.337867795678273,
      "tokens_seen": 3137404928
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9849548645937814e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.33786252888195,
      "tokens_seen": 3137470464
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.982948846539619e-05,
      "loss": 1.1631,
      "theoretical_loss": 3.337857262226442,
      "tokens_seen": 3137536000
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9809428284854566e-05,
      "loss": 1.231,
      "theoretical_loss": 3.3378519957117434,
      "tokens_seen": 3137601536
    },
    {
      "epoch": 0.9,
      "objective/train/docs_used": 1140026,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2733089923858643,
      "objective/train/theoretical_loss": 3.337846729337847,
      "objective/train/tokens_used": 1508126176,
      "theoretical_loss": 3.337846729337847,
      "tokens_seen": 3137667072
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.978936810431294e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.337846729337847,
      "tokens_seen": 3137667072
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.976930792377132e-05,
      "loss": 1.2051,
      "theoretical_loss": 3.3378414631047457,
      "tokens_seen": 3137732608
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.974924774322969e-05,
      "loss": 1.3251,
      "theoretical_loss": 3.3378361970124337,
      "tokens_seen": 3137798144
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.972918756268806e-05,
      "loss": 1.273,
      "theoretical_loss": 3.337830931060904,
      "tokens_seen": 3137863680
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.970912738214644e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.3378256652501492,
      "tokens_seen": 3137929216
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9689067201604814e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.3378203995801634,
      "tokens_seen": 3137994752
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.966900702106319e-05,
      "loss": 1.2236,
      "theoretical_loss": 3.3378151340509397,
      "tokens_seen": 3138060288
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9648946840521565e-05,
      "loss": 1.2506,
      "theoretical_loss": 3.3378098686624713,
      "tokens_seen": 3138125824
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.962888665997994e-05,
      "loss": 1.2146,
      "theoretical_loss": 3.3378046034147513,
      "tokens_seen": 3138191360
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.960882647943832e-05,
      "loss": 1.2642,
      "theoretical_loss": 3.3377993383077733,
      "tokens_seen": 3138256896
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.958876629889669e-05,
      "loss": 1.3508,
      "theoretical_loss": 3.3377940733415308,
      "tokens_seen": 3138322432
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.956870611835507e-05,
      "loss": 1.287,
      "theoretical_loss": 3.3377888085160166,
      "tokens_seen": 3138387968
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9548645937813444e-05,
      "loss": 1.1982,
      "theoretical_loss": 3.3377835438312244,
      "tokens_seen": 3138453504
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.952858575727181e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.3377782792871473,
      "tokens_seen": 3138519040
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.950852557673019e-05,
      "loss": 1.3182,
      "theoretical_loss": 3.337773014883778,
      "tokens_seen": 3138584576
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9488465396188565e-05,
      "loss": 1.2152,
      "theoretical_loss": 3.3377677506211114,
      "tokens_seen": 3138650112
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.946840521564694e-05,
      "loss": 1.2699,
      "theoretical_loss": 3.3377624864991393,
      "tokens_seen": 3138715648
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.944834503510532e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.3377572225178556,
      "tokens_seen": 3138781184
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.942828485456369e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.3377519586772535,
      "tokens_seen": 3138846720
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.940822467402207e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.3377466949773265,
      "tokens_seen": 3138912256
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9388164493480444e-05,
      "loss": 1.2201,
      "theoretical_loss": 3.3377414314180673,
      "tokens_seen": 3138977792
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.936810431293882e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.33773616799947,
      "tokens_seen": 3139043328
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9348044132397195e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.3377309047215276,
      "tokens_seen": 3139108864
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9327983951855564e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.337725641584233,
      "tokens_seen": 3139174400
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.930792377131394e-05,
      "loss": 1.1949,
      "theoretical_loss": 3.33772037858758,
      "tokens_seen": 3139239936
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9287863590772316e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.337715115731562,
      "tokens_seen": 3139305472
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.92678034102307e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.3377098530161717,
      "tokens_seen": 3139371008
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9247743229689074e-05,
      "loss": 1.2147,
      "theoretical_loss": 3.337704590441403,
      "tokens_seen": 3139436544
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.922768304914744e-05,
      "loss": 1.384,
      "theoretical_loss": 3.337699328007249,
      "tokens_seen": 3139502080
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.920762286860582e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.337694065713703,
      "tokens_seen": 3139567616
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9187562688064195e-05,
      "loss": 1.2075,
      "theoretical_loss": 3.3376888035607584,
      "tokens_seen": 3139633152
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.916750250752257e-05,
      "loss": 1.366,
      "theoretical_loss": 3.337683541548408,
      "tokens_seen": 3139698688
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.914744232698094e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.3376782796766458,
      "tokens_seen": 3139764224
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9127382146439316e-05,
      "loss": 1.2079,
      "theoretical_loss": 3.337673017945465,
      "tokens_seen": 3139829760
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.910732196589769e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.3376677563548585,
      "tokens_seen": 3139895296
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.908726178535607e-05,
      "loss": 1.3092,
      "theoretical_loss": 3.3376624949048197,
      "tokens_seen": 3139960832
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.906720160481445e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.3376572335953423,
      "tokens_seen": 3140026368
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.904714142427282e-05,
      "loss": 1.1864,
      "theoretical_loss": 3.337651972426419,
      "tokens_seen": 3140091904
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9027081243731194e-05,
      "loss": 1.1771,
      "theoretical_loss": 3.337646711398044,
      "tokens_seen": 3140157440
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.900702106318957e-05,
      "loss": 1.3713,
      "theoretical_loss": 3.33764145051021,
      "tokens_seen": 3140222976
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8986960882647946e-05,
      "loss": 1.297,
      "theoretical_loss": 3.33763618976291,
      "tokens_seen": 3140288512
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.896690070210632e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.337630929156138,
      "tokens_seen": 3140354048
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.894684052156469e-05,
      "loss": 1.1341,
      "theoretical_loss": 3.337625668689887,
      "tokens_seen": 3140419584
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.892678034102307e-05,
      "loss": 1.278,
      "theoretical_loss": 3.3376204083641507,
      "tokens_seen": 3140485120
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.890672016048144e-05,
      "loss": 1.2392,
      "theoretical_loss": 3.3376151481789216,
      "tokens_seen": 3140550656
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8886659979939825e-05,
      "loss": 1.2132,
      "theoretical_loss": 3.3376098881341933,
      "tokens_seen": 3140616192
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.88665997993982e-05,
      "loss": 1.2068,
      "theoretical_loss": 3.33760462822996,
      "tokens_seen": 3140681728
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.884653961885657e-05,
      "loss": 1.2037,
      "theoretical_loss": 3.3375993684662135,
      "tokens_seen": 3140747264
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8826479438314946e-05,
      "loss": 1.269,
      "theoretical_loss": 3.3375941088429486,
      "tokens_seen": 3140812800
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.880641925777332e-05,
      "loss": 1.318,
      "theoretical_loss": 3.3375888493601575,
      "tokens_seen": 3140878336
    },
    {
      "epoch": 0.9,
      "objective/train/docs_used": 1141163,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4142014980316162,
      "objective/train/theoretical_loss": 3.337583590017834,
      "objective/train/tokens_used": 1511402976,
      "theoretical_loss": 3.337583590017834,
      "tokens_seen": 3140943872
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.87863590772317e-05,
      "loss": 1.3455,
      "theoretical_loss": 3.337583590017834,
      "tokens_seen": 3140943872
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.876629889669007e-05,
      "loss": 1.2995,
      "theoretical_loss": 3.3375783308159717,
      "tokens_seen": 3141009408
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.874623871614844e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.337573071754563,
      "tokens_seen": 3141074944
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.872617853560682e-05,
      "loss": 1.1986,
      "theoretical_loss": 3.3375678128336026,
      "tokens_seen": 3141140480
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8706118355065194e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.3375625540530827,
      "tokens_seen": 3141206016
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8686058174523576e-05,
      "loss": 1.2499,
      "theoretical_loss": 3.3375572954129966,
      "tokens_seen": 3141271552
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.866599799398195e-05,
      "loss": 1.3153,
      "theoretical_loss": 3.3375520369133382,
      "tokens_seen": 3141337088
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.864593781344032e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.337546778554101,
      "tokens_seen": 3141402624
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.86258776328987e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.337541520335277,
      "tokens_seen": 3141468160
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.860581745235707e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.337536262256861,
      "tokens_seen": 3141533696
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.858575727181545e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.337531004318846,
      "tokens_seen": 3141599232
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8565697091273824e-05,
      "loss": 1.3788,
      "theoretical_loss": 3.3375257465212247,
      "tokens_seen": 3141664768
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.854563691073219e-05,
      "loss": 1.291,
      "theoretical_loss": 3.3375204888639907,
      "tokens_seen": 3141730304
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.852557673019057e-05,
      "loss": 1.2566,
      "theoretical_loss": 3.3375152313471377,
      "tokens_seen": 3141795840
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.850551654964895e-05,
      "loss": 1.2593,
      "theoretical_loss": 3.3375099739706586,
      "tokens_seen": 3141861376
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.848545636910733e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.337504716734547,
      "tokens_seen": 3141926912
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8465396188565696e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.337499459638796,
      "tokens_seen": 3141992448
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.844533600802407e-05,
      "loss": 1.1774,
      "theoretical_loss": 3.337494202683399,
      "tokens_seen": 3142057984
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.842527582748245e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.3374889458683494,
      "tokens_seen": 3142123520
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8405215646940824e-05,
      "loss": 1.3214,
      "theoretical_loss": 3.3374836891936406,
      "tokens_seen": 3142189056
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.83851554663992e-05,
      "loss": 1.1312,
      "theoretical_loss": 3.3374784326592657,
      "tokens_seen": 3142254592
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.836509528585757e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.3374731762652177,
      "tokens_seen": 3142320128
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8345035105315944e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.337467920011491,
      "tokens_seen": 3142385664
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.832497492477433e-05,
      "loss": 1.2093,
      "theoretical_loss": 3.337462663898078,
      "tokens_seen": 3142451200
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.83049147442327e-05,
      "loss": 1.1933,
      "theoretical_loss": 3.337457407924972,
      "tokens_seen": 3142516736
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.828485456369108e-05,
      "loss": 1.249,
      "theoretical_loss": 3.337452152092167,
      "tokens_seen": 3142582272
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.826479438314945e-05,
      "loss": 1.308,
      "theoretical_loss": 3.337446896399656,
      "tokens_seen": 3142647808
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.824473420260782e-05,
      "loss": 1.2132,
      "theoretical_loss": 3.337441640847432,
      "tokens_seen": 3142713344
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.82246740220662e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.337436385435489,
      "tokens_seen": 3142778880
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8204613841524575e-05,
      "loss": 1.214,
      "theoretical_loss": 3.3374311301638198,
      "tokens_seen": 3142844416
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.818455366098295e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.337425875032418,
      "tokens_seen": 3142909952
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.816449348044132e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.337420620041277,
      "tokens_seen": 3142975488
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8144433299899696e-05,
      "loss": 1.1995,
      "theoretical_loss": 3.337415365190389,
      "tokens_seen": 3143041024
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.812437311935808e-05,
      "loss": 1.414,
      "theoretical_loss": 3.337410110479749,
      "tokens_seen": 3143106560
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8104312938816454e-05,
      "loss": 1.3447,
      "theoretical_loss": 3.33740485590935,
      "tokens_seen": 3143172096
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.808425275827483e-05,
      "loss": 1.2407,
      "theoretical_loss": 3.3373996014791842,
      "tokens_seen": 3143237632
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.80641925777332e-05,
      "loss": 1.2571,
      "theoretical_loss": 3.337394347189246,
      "tokens_seen": 3143303168
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.8044132397191574e-05,
      "loss": 1.2274,
      "theoretical_loss": 3.3373890930395285,
      "tokens_seen": 3143368704
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.802407221664995e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.3373838390300246,
      "tokens_seen": 3143434240
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.8004012036108326e-05,
      "loss": 1.2451,
      "theoretical_loss": 3.337378585160728,
      "tokens_seen": 3143499776
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.79839518555667e-05,
      "loss": 1.176,
      "theoretical_loss": 3.3373733314316323,
      "tokens_seen": 3143565312
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.796389167502507e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.3373680778427306,
      "tokens_seen": 3143630848
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7943831494483453e-05,
      "loss": 1.2184,
      "theoretical_loss": 3.3373628243940163,
      "tokens_seen": 3143696384
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.792377131394183e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.337357571085482,
      "tokens_seen": 3143761920
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7903711133400205e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.3373523179171225,
      "tokens_seen": 3143827456
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.788365095285858e-05,
      "loss": 1.3526,
      "theoretical_loss": 3.3373470648889296,
      "tokens_seen": 3143892992
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.786359077231695e-05,
      "loss": 1.2532,
      "theoretical_loss": 3.3373418120008975,
      "tokens_seen": 3143958528
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7843530591775326e-05,
      "loss": 1.2901,
      "theoretical_loss": 3.3373365592530195,
      "tokens_seen": 3144024064
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.78234704112337e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.337331306645289,
      "tokens_seen": 3144089600
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.780341023069208e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.3373260541776992,
      "tokens_seen": 3144155136
    },
    {
      "epoch": 0.91,
      "objective/train/docs_used": 1142279,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2020081281661987,
      "objective/train/theoretical_loss": 3.337320801850243,
      "objective/train/tokens_used": 1514679776,
      "theoretical_loss": 3.337320801850243,
      "tokens_seen": 3144220672
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7783350050150446e-05,
      "loss": 1.2245,
      "theoretical_loss": 3.337320801850243,
      "tokens_seen": 3144220672
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.776328986960883e-05,
      "loss": 1.208,
      "theoretical_loss": 3.3373155496629145,
      "tokens_seen": 3144286208
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7743229689067205e-05,
      "loss": 1.1283,
      "theoretical_loss": 3.3373102976157063,
      "tokens_seen": 3144351744
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.772316950852558e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.3373050457086126,
      "tokens_seen": 3144417280
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7703109327983956e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.337299793941626,
      "tokens_seen": 3144482816
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7683049147442325e-05,
      "loss": 1.1711,
      "theoretical_loss": 3.3372945423147398,
      "tokens_seen": 3144548352
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.76629889669007e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.337289290827948,
      "tokens_seen": 3144613888
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.764292878635908e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.3372840394812435,
      "tokens_seen": 3144679424
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.762286860581745e-05,
      "loss": 1.3013,
      "theoretical_loss": 3.33727878827462,
      "tokens_seen": 3144744960
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.760280842527583e-05,
      "loss": 1.1943,
      "theoretical_loss": 3.3372735372080706,
      "tokens_seen": 3144810496
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.75827482447342e-05,
      "loss": 1.3692,
      "theoretical_loss": 3.337268286281588,
      "tokens_seen": 3144876032
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.756268806419258e-05,
      "loss": 1.2721,
      "theoretical_loss": 3.337263035495167,
      "tokens_seen": 3144941568
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7542627883650956e-05,
      "loss": 1.2203,
      "theoretical_loss": 3.3372577848487994,
      "tokens_seen": 3145007104
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.752256770310933e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.3372525343424795,
      "tokens_seen": 3145072640
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.750250752256771e-05,
      "loss": 1.3444,
      "theoretical_loss": 3.3372472839762004,
      "tokens_seen": 3145138176
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7482447342026076e-05,
      "loss": 1.2555,
      "theoretical_loss": 3.3372420337499555,
      "tokens_seen": 3145203712
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.746238716148445e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.337236783663738,
      "tokens_seen": 3145269248
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.744232698094283e-05,
      "loss": 1.1666,
      "theoretical_loss": 3.3372315337175418,
      "tokens_seen": 3145334784
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7422266800401204e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.3372262839113596,
      "tokens_seen": 3145400320
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.740220661985958e-05,
      "loss": 1.1583,
      "theoretical_loss": 3.337221034245185,
      "tokens_seen": 3145465856
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7382146439317955e-05,
      "loss": 1.2667,
      "theoretical_loss": 3.3372157847190107,
      "tokens_seen": 3145531392
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.736208625877633e-05,
      "loss": 1.3222,
      "theoretical_loss": 3.337210535332831,
      "tokens_seen": 3145596928
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.734202607823471e-05,
      "loss": 1.1587,
      "theoretical_loss": 3.3372052860866392,
      "tokens_seen": 3145662464
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.732196589769308e-05,
      "loss": 1.2351,
      "theoretical_loss": 3.3372000369804278,
      "tokens_seen": 3145728000
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.730190571715146e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.337194788014191,
      "tokens_seen": 3145793536
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.728184553660983e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.3371895391879223,
      "tokens_seen": 3145859072
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.72617853560682e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.337184290501614,
      "tokens_seen": 3145924608
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.724172517552658e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.3371790419552605,
      "tokens_seen": 3145990144
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7221664994984955e-05,
      "loss": 1.1978,
      "theoretical_loss": 3.337173793548854,
      "tokens_seen": 3146055680
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.720160481444334e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.337168545282389,
      "tokens_seen": 3146121216
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7181544633901706e-05,
      "loss": 1.2446,
      "theoretical_loss": 3.3371632971558585,
      "tokens_seen": 3146186752
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.716148445336008e-05,
      "loss": 1.311,
      "theoretical_loss": 3.337158049169256,
      "tokens_seen": 3146252288
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.714142427281846e-05,
      "loss": 1.1678,
      "theoretical_loss": 3.337152801322574,
      "tokens_seen": 3146317824
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7121364092276834e-05,
      "loss": 1.1866,
      "theoretical_loss": 3.337147553615807,
      "tokens_seen": 3146383360
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.710130391173521e-05,
      "loss": 1.2351,
      "theoretical_loss": 3.3371423060489476,
      "tokens_seen": 3146448896
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.708124373119358e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.3371370586219893,
      "tokens_seen": 3146514432
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7061183550651954e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.3371318113349258,
      "tokens_seen": 3146579968
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.704112337011033e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.33712656418775,
      "tokens_seen": 3146645504
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7021063189568706e-05,
      "loss": 1.2186,
      "theoretical_loss": 3.3371213171804555,
      "tokens_seen": 3146711040
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.700100300902708e-05,
      "loss": 1.1894,
      "theoretical_loss": 3.3371160703130354,
      "tokens_seen": 3146776576
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.698094282848546e-05,
      "loss": 1.2474,
      "theoretical_loss": 3.3371108235854834,
      "tokens_seen": 3146842112
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6960882647943833e-05,
      "loss": 1.2163,
      "theoretical_loss": 3.337105576997793,
      "tokens_seen": 3146907648
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.694082246740221e-05,
      "loss": 1.2292,
      "theoretical_loss": 3.337100330549957,
      "tokens_seen": 3146973184
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6920762286860585e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.3370950842419695,
      "tokens_seen": 3147038720
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6900702106318954e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.337089838073823,
      "tokens_seen": 3147104256
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.688064192577733e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.3370845920455112,
      "tokens_seen": 3147169792
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6860581745235706e-05,
      "loss": 1.2198,
      "theoretical_loss": 3.337079346157028,
      "tokens_seen": 3147235328
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.684052156469408e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.337074100408366,
      "tokens_seen": 3147300864
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6820461384152464e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.3370688547995186,
      "tokens_seen": 3147366400
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.680040120361083e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.33706360933048,
      "tokens_seen": 3147431936
    },
    {
      "epoch": 0.91,
      "objective/train/docs_used": 1143525,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2574238777160645,
      "objective/train/theoretical_loss": 3.337058364001243,
      "objective/train/tokens_used": 1517956576,
      "theoretical_loss": 3.337058364001243,
      "tokens_seen": 3147497472
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.678034102306921e-05,
      "loss": 1.262,
      "theoretical_loss": 3.337058364001243,
      "tokens_seen": 3147497472
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6760280842527585e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.3370531188118004,
      "tokens_seen": 3147563008
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.674022066198596e-05,
      "loss": 1.3488,
      "theoretical_loss": 3.3370478737621463,
      "tokens_seen": 3147628544
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6720160481444336e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.3370426288522737,
      "tokens_seen": 3147694080
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6700100300902705e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.3370373840821763,
      "tokens_seen": 3147759616
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.668004012036108e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.3370321394518476,
      "tokens_seen": 3147825152
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.665997993981946e-05,
      "loss": 1.1735,
      "theoretical_loss": 3.3370268949612805,
      "tokens_seen": 3147890688
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.663991975927784e-05,
      "loss": 1.1973,
      "theoretical_loss": 3.3370216506104686,
      "tokens_seen": 3147956224
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6619859578736215e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.337016406399405,
      "tokens_seen": 3148021760
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6599799398194584e-05,
      "loss": 1.2031,
      "theoretical_loss": 3.337011162328083,
      "tokens_seen": 3148087296
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.657973921765296e-05,
      "loss": 1.2275,
      "theoretical_loss": 3.337005918396497,
      "tokens_seen": 3148152832
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6559679037111336e-05,
      "loss": 1.1418,
      "theoretical_loss": 3.337000674604639,
      "tokens_seen": 3148218368
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.653961885656971e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.336995430952503,
      "tokens_seen": 3148283904
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.651955867602809e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.3369901874400827,
      "tokens_seen": 3148349440
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6499498495486456e-05,
      "loss": 1.1954,
      "theoretical_loss": 3.336984944067371,
      "tokens_seen": 3148414976
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.647943831494483e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.336979700834361,
      "tokens_seen": 3148480512
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.645937813440321e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.336974457741047,
      "tokens_seen": 3148546048
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.643931795386159e-05,
      "loss": 1.2498,
      "theoretical_loss": 3.3369692147874215,
      "tokens_seen": 3148611584
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6419257773319966e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.336963971973478,
      "tokens_seen": 3148677120
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6399197592778335e-05,
      "loss": 1.204,
      "theoretical_loss": 3.3369587292992104,
      "tokens_seen": 3148742656
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.637913741223671e-05,
      "loss": 1.2453,
      "theoretical_loss": 3.336953486764612,
      "tokens_seen": 3148808192
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.635907723169509e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.336948244369675,
      "tokens_seen": 3148873728
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.633901705115346e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.3369430021143947,
      "tokens_seen": 3148939264
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.631895687061183e-05,
      "loss": 1.3557,
      "theoretical_loss": 3.336937759998763,
      "tokens_seen": 3149004800
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.629889669007021e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.3369325180227736,
      "tokens_seen": 3149070336
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.627883650952858e-05,
      "loss": 1.286,
      "theoretical_loss": 3.33692727618642,
      "tokens_seen": 3149135872
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6258776328986966e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.336922034489696,
      "tokens_seen": 3149201408
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.623871614844534e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.3369167929325942,
      "tokens_seen": 3149266944
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.621865596790371e-05,
      "loss": 1.3393,
      "theoretical_loss": 3.3369115515151084,
      "tokens_seen": 3149332480
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6198595787362086e-05,
      "loss": 1.2407,
      "theoretical_loss": 3.336906310237232,
      "tokens_seen": 3149398016
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.617853560682046e-05,
      "loss": 1.4112,
      "theoretical_loss": 3.3369010690989582,
      "tokens_seen": 3149463552
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.615847542627884e-05,
      "loss": 1.2202,
      "theoretical_loss": 3.33689582810028,
      "tokens_seen": 3149529088
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6138415245737214e-05,
      "loss": 1.3244,
      "theoretical_loss": 3.336890587241192,
      "tokens_seen": 3149594624
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.611835506519558e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.3368853465216866,
      "tokens_seen": 3149660160
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.609829488465396e-05,
      "loss": 1.2194,
      "theoretical_loss": 3.336880105941757,
      "tokens_seen": 3149725696
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6078234704112334e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.3368748655013976,
      "tokens_seen": 3149791232
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.605817452357072e-05,
      "loss": 1.3712,
      "theoretical_loss": 3.3368696252006003,
      "tokens_seen": 3149856768
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.603811434302909e-05,
      "loss": 1.2365,
      "theoretical_loss": 3.33686438503936,
      "tokens_seen": 3149922304
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.601805416248746e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.336859145017669,
      "tokens_seen": 3149987840
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.599799398194584e-05,
      "loss": 1.1566,
      "theoretical_loss": 3.3368539051355213,
      "tokens_seen": 3150053376
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5977933801404213e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.3368486653929104,
      "tokens_seen": 3150118912
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.595787362086259e-05,
      "loss": 1.3376,
      "theoretical_loss": 3.3368434257898287,
      "tokens_seen": 3150184448
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5937813440320965e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.3368381863262706,
      "tokens_seen": 3150249984
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5917753259779334e-05,
      "loss": 1.2162,
      "theoretical_loss": 3.336832947002229,
      "tokens_seen": 3150315520
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.589769307923771e-05,
      "loss": 1.1898,
      "theoretical_loss": 3.3368277078176973,
      "tokens_seen": 3150381056
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.587763289869609e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.3368224687726693,
      "tokens_seen": 3150446592
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.585757271815447e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.336817229867138,
      "tokens_seen": 3150512128
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5837512537612844e-05,
      "loss": 1.3739,
      "theoretical_loss": 3.3368119911010967,
      "tokens_seen": 3150577664
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.581745235707121e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.336806752474539,
      "tokens_seen": 3150643200
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.579739217652959e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.3368015139874583,
      "tokens_seen": 3150708736
    },
    {
      "epoch": 0.91,
      "objective/train/docs_used": 1144741,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2674669027328491,
      "objective/train/theoretical_loss": 3.3367962756398475,
      "objective/train/tokens_used": 1521233376,
      "theoretical_loss": 3.3367962756398475,
      "tokens_seen": 3150774272
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5777331995987965e-05,
      "loss": 1.1916,
      "theoretical_loss": 3.3367962756398475,
      "tokens_seen": 3150774272
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.575727181544634e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.3367910374317007,
      "tokens_seen": 3150839808
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5737211634904716e-05,
      "loss": 1.3161,
      "theoretical_loss": 3.3367857993630112,
      "tokens_seen": 3150905344
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5717151454363085e-05,
      "loss": 1.1615,
      "theoretical_loss": 3.3367805614337716,
      "tokens_seen": 3150970880
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.569709127382147e-05,
      "loss": 1.1384,
      "theoretical_loss": 3.3367753236439763,
      "tokens_seen": 3151036416
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5677031093279844e-05,
      "loss": 1.2374,
      "theoretical_loss": 3.336770085993618,
      "tokens_seen": 3151101952
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.565697091273822e-05,
      "loss": 1.3545,
      "theoretical_loss": 3.3367648484826904,
      "tokens_seen": 3151167488
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5636910732196595e-05,
      "loss": 1.2162,
      "theoretical_loss": 3.336759611111187,
      "tokens_seen": 3151233024
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5616850551654964e-05,
      "loss": 1.3187,
      "theoretical_loss": 3.3367543738791006,
      "tokens_seen": 3151298560
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.559679037111334e-05,
      "loss": 1.1118,
      "theoretical_loss": 3.336749136786425,
      "tokens_seen": 3151364096
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5576730190571716e-05,
      "loss": 1.2803,
      "theoretical_loss": 3.336743899833154,
      "tokens_seen": 3151429632
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.555667001003009e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.3367386630192803,
      "tokens_seen": 3151495168
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.553660982948846e-05,
      "loss": 1.3682,
      "theoretical_loss": 3.336733426344798,
      "tokens_seen": 3151560704
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5516549648946836e-05,
      "loss": 1.2141,
      "theoretical_loss": 3.3367281898096994,
      "tokens_seen": 3151626240
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.549648946840522e-05,
      "loss": 1.3611,
      "theoretical_loss": 3.336722953413979,
      "tokens_seen": 3151691776
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5476429287863595e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.3367177171576294,
      "tokens_seen": 3151757312
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.545636910732197e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.3367124810406446,
      "tokens_seen": 3151822848
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.543630892678034e-05,
      "loss": 1.2049,
      "theoretical_loss": 3.3367072450630175,
      "tokens_seen": 3151888384
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5416248746238715e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.3367020092247417,
      "tokens_seen": 3151953920
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.539618856569709e-05,
      "loss": 1.202,
      "theoretical_loss": 3.336696773525811,
      "tokens_seen": 3152019456
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.537612838515547e-05,
      "loss": 1.348,
      "theoretical_loss": 3.336691537966218,
      "tokens_seen": 3152084992
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.535606820461384e-05,
      "loss": 1.2311,
      "theoretical_loss": 3.3366863025459566,
      "tokens_seen": 3152150528
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.533600802407221e-05,
      "loss": 1.1958,
      "theoretical_loss": 3.3366810672650202,
      "tokens_seen": 3152216064
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5315947843530594e-05,
      "loss": 1.248,
      "theoretical_loss": 3.336675832123402,
      "tokens_seen": 3152281600
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.529588766298897e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.3366705971210955,
      "tokens_seen": 3152347136
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5275827482447346e-05,
      "loss": 1.2417,
      "theoretical_loss": 3.3366653622580937,
      "tokens_seen": 3152412672
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.525576730190572e-05,
      "loss": 1.2279,
      "theoretical_loss": 3.3366601275343912,
      "tokens_seen": 3152478208
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.523570712136409e-05,
      "loss": 1.1412,
      "theoretical_loss": 3.33665489294998,
      "tokens_seen": 3152543744
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5215646940822466e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.336649658504854,
      "tokens_seen": 3152609280
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.519558676028084e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.336644424199007,
      "tokens_seen": 3152674816
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.517552657973922e-05,
      "loss": 1.3406,
      "theoretical_loss": 3.3366391900324324,
      "tokens_seen": 3152740352
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5155466399197594e-05,
      "loss": 1.2501,
      "theoretical_loss": 3.3366339560051226,
      "tokens_seen": 3152805888
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.513540621865597e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.336628722117072,
      "tokens_seen": 3152871424
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5115346038114345e-05,
      "loss": 1.3459,
      "theoretical_loss": 3.3366234883682733,
      "tokens_seen": 3152936960
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.509528585757272e-05,
      "loss": 1.3986,
      "theoretical_loss": 3.336618254758721,
      "tokens_seen": 3153002496
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.50752256770311e-05,
      "loss": 1.1736,
      "theoretical_loss": 3.3366130212884073,
      "tokens_seen": 3153068032
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.505516549648947e-05,
      "loss": 1.2123,
      "theoretical_loss": 3.336607787957326,
      "tokens_seen": 3153133568
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.503510531594784e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.3366025547654705,
      "tokens_seen": 3153199104
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.501504513540622e-05,
      "loss": 1.2003,
      "theoretical_loss": 3.3365973217128344,
      "tokens_seen": 3153264640
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4994984954864593e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.3365920887994114,
      "tokens_seen": 3153330176
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.497492477432297e-05,
      "loss": 1.247,
      "theoretical_loss": 3.336586856025194,
      "tokens_seen": 3153395712
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4954864593781345e-05,
      "loss": 1.248,
      "theoretical_loss": 3.336581623390176,
      "tokens_seen": 3153461248
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.493480441323972e-05,
      "loss": 1.2997,
      "theoretical_loss": 3.3365763908943515,
      "tokens_seen": 3153526784
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4914744232698097e-05,
      "loss": 1.2299,
      "theoretical_loss": 3.336571158537713,
      "tokens_seen": 3153592320
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.489468405215647e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.3365659263202536,
      "tokens_seen": 3153657856
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.487462387161485e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.3365606942419683,
      "tokens_seen": 3153723392
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.485456369107322e-05,
      "loss": 1.2268,
      "theoretical_loss": 3.336555462302849,
      "tokens_seen": 3153788928
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.483450351053159e-05,
      "loss": 1.359,
      "theoretical_loss": 3.3365502305028896,
      "tokens_seen": 3153854464
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.481444332998997e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.3365449988420837,
      "tokens_seen": 3153920000
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4794383149448345e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.3365397673204242,
      "tokens_seen": 3153985536
    },
    {
      "epoch": 0.91,
      "objective/train/docs_used": 1145990,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.0538361072540283,
      "objective/train/theoretical_loss": 3.336534535937905,
      "objective/train/tokens_used": 1524510176,
      "theoretical_loss": 3.336534535937905,
      "tokens_seen": 3154051072
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.477432296890672e-05,
      "loss": 1.1775,
      "theoretical_loss": 3.336534535937905,
      "tokens_seen": 3154051072
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4754262788365096e-05,
      "loss": 1.3139,
      "theoretical_loss": 3.3365293046945195,
      "tokens_seen": 3154116608
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.473420260782347e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.3365240735902613,
      "tokens_seen": 3154182144
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.471414242728185e-05,
      "loss": 1.1932,
      "theoretical_loss": 3.336518842625123,
      "tokens_seen": 3154247680
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4694082246740224e-05,
      "loss": 1.1654,
      "theoretical_loss": 3.3365136117990986,
      "tokens_seen": 3154313216
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.46740220661986e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.336508381112181,
      "tokens_seen": 3154378752
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.465396188565697e-05,
      "loss": 1.2226,
      "theoretical_loss": 3.3365031505643645,
      "tokens_seen": 3154444288
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4633901705115344e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.336497920155642,
      "tokens_seen": 3154509824
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.461384152457372e-05,
      "loss": 1.2066,
      "theoretical_loss": 3.336492689886007,
      "tokens_seen": 3154575360
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4593781344032096e-05,
      "loss": 1.1455,
      "theoretical_loss": 3.3364874597554524,
      "tokens_seen": 3154640896
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.457372116349048e-05,
      "loss": 1.3764,
      "theoretical_loss": 3.3364822297639725,
      "tokens_seen": 3154706432
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.455366098294885e-05,
      "loss": 1.3628,
      "theoretical_loss": 3.33647699991156,
      "tokens_seen": 3154771968
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.453360080240722e-05,
      "loss": 1.2322,
      "theoretical_loss": 3.3364717701982083,
      "tokens_seen": 3154837504
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.45135406218656e-05,
      "loss": 1.286,
      "theoretical_loss": 3.3364665406239116,
      "tokens_seen": 3154903040
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4493480441323975e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.3364613111886627,
      "tokens_seen": 3154968576
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.447342026078235e-05,
      "loss": 1.252,
      "theoretical_loss": 3.3364560818924547,
      "tokens_seen": 3155034112
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.445336008024072e-05,
      "loss": 1.3872,
      "theoretical_loss": 3.336450852735282,
      "tokens_seen": 3155099648
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4433299899699095e-05,
      "loss": 1.2238,
      "theoretical_loss": 3.3364456237171374,
      "tokens_seen": 3155165184
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.441323971915747e-05,
      "loss": 1.1784,
      "theoretical_loss": 3.3364403948380144,
      "tokens_seen": 3155230720
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.439317953861585e-05,
      "loss": 1.2155,
      "theoretical_loss": 3.336435166097906,
      "tokens_seen": 3155296256
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.437311935807423e-05,
      "loss": 1.3243,
      "theoretical_loss": 3.3364299374968063,
      "tokens_seen": 3155361792
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.43530591775326e-05,
      "loss": 1.1857,
      "theoretical_loss": 3.3364247090347083,
      "tokens_seen": 3155427328
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4332998996990974e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.336419480711605,
      "tokens_seen": 3155492864
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.431293881644935e-05,
      "loss": 1.3773,
      "theoretical_loss": 3.336414252527491,
      "tokens_seen": 3155558400
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4292878635907726e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.336409024482359,
      "tokens_seen": 3155623936
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.42728184553661e-05,
      "loss": 1.2252,
      "theoretical_loss": 3.3364037965762026,
      "tokens_seen": 3155689472
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.425275827482447e-05,
      "loss": 1.2123,
      "theoretical_loss": 3.336398568809015,
      "tokens_seen": 3155755008
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4232698094282846e-05,
      "loss": 1.2288,
      "theoretical_loss": 3.3363933411807896,
      "tokens_seen": 3155820544
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.421263791374122e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.33638811369152,
      "tokens_seen": 3155886080
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4192577733199605e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.3363828863411995,
      "tokens_seen": 3155951616
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.417251755265798e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.336377659129822,
      "tokens_seen": 3156017152
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.415245737211635e-05,
      "loss": 1.2274,
      "theoretical_loss": 3.33637243205738,
      "tokens_seen": 3156082688
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4132397191574725e-05,
      "loss": 1.1872,
      "theoretical_loss": 3.3363672051238678,
      "tokens_seen": 3156148224
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.41123370110331e-05,
      "loss": 1.1664,
      "theoretical_loss": 3.336361978329278,
      "tokens_seen": 3156213760
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.409227683049148e-05,
      "loss": 1.23,
      "theoretical_loss": 3.336356751673605,
      "tokens_seen": 3156279296
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4072216649949846e-05,
      "loss": 1.2076,
      "theoretical_loss": 3.3363515251568416,
      "tokens_seen": 3156344832
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.405215646940822e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.336346298778981,
      "tokens_seen": 3156410368
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.40320962888666e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.3363410725400175,
      "tokens_seen": 3156475904
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.401203610832498e-05,
      "loss": 1.2029,
      "theoretical_loss": 3.3363358464399435,
      "tokens_seen": 3156541440
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3991975927783356e-05,
      "loss": 1.2664,
      "theoretical_loss": 3.336330620478753,
      "tokens_seen": 3156606976
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3971915747241725e-05,
      "loss": 1.1964,
      "theoretical_loss": 3.3363253946564395,
      "tokens_seen": 3156672512
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.39518555667001e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.336320168972996,
      "tokens_seen": 3156738048
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3931795386158477e-05,
      "loss": 1.4383,
      "theoretical_loss": 3.3363149434284165,
      "tokens_seen": 3156803584
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.391173520561685e-05,
      "loss": 1.1866,
      "theoretical_loss": 3.3363097180226937,
      "tokens_seen": 3156869120
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.389167502507523e-05,
      "loss": 1.2071,
      "theoretical_loss": 3.336304492755822,
      "tokens_seen": 3156934656
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.38716148445336e-05,
      "loss": 1.277,
      "theoretical_loss": 3.336299267627794,
      "tokens_seen": 3157000192
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.385155466399197e-05,
      "loss": 1.3284,
      "theoretical_loss": 3.3362940426386034,
      "tokens_seen": 3157065728
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.383149448345035e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.336288817788243,
      "tokens_seen": 3157131264
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.381143430290873e-05,
      "loss": 1.355,
      "theoretical_loss": 3.3362835930767076,
      "tokens_seen": 3157196800
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.379137412236711e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.33627836850399,
      "tokens_seen": 3157262336
    },
    {
      "epoch": 0.91,
      "objective/train/docs_used": 1147172,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3894332647323608,
      "objective/train/theoretical_loss": 3.3362731440700832,
      "objective/train/tokens_used": 1527786976,
      "theoretical_loss": 3.3362731440700832,
      "tokens_seen": 3157327872
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3771313941825476e-05,
      "loss": 1.3515,
      "theoretical_loss": 3.3362731440700832,
      "tokens_seen": 3157327872
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.375125376128385e-05,
      "loss": 1.3709,
      "theoretical_loss": 3.336267919774981,
      "tokens_seen": 3157393408
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.373119358074223e-05,
      "loss": 1.2245,
      "theoretical_loss": 3.3362626956186765,
      "tokens_seen": 3157458944
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3711133400200604e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.3362574716011637,
      "tokens_seen": 3157524480
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.369107321965898e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.336252247722436,
      "tokens_seen": 3157590016
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.367101303911735e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.336247023982486,
      "tokens_seen": 3157655552
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3650952858575724e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.336241800381308,
      "tokens_seen": 3157721088
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.363089267803411e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.3362365769188953,
      "tokens_seen": 3157786624
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.361083249749248e-05,
      "loss": 1.2228,
      "theoretical_loss": 3.336231353595241,
      "tokens_seen": 3157852160
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.359077231695086e-05,
      "loss": 1.1596,
      "theoretical_loss": 3.3362261304103384,
      "tokens_seen": 3157917696
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.357071213640923e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.3362209073641815,
      "tokens_seen": 3157983232
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.35506519558676e-05,
      "loss": 1.1985,
      "theoretical_loss": 3.3362156844567634,
      "tokens_seen": 3158048768
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.353059177532598e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.336210461688078,
      "tokens_seen": 3158114304
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3510531594784355e-05,
      "loss": 1.3681,
      "theoretical_loss": 3.336205239058118,
      "tokens_seen": 3158179840
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.349047141424273e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.3362000165668775,
      "tokens_seen": 3158245376
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.34704112337011e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.3361947942143493,
      "tokens_seen": 3158310912
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3450351053159475e-05,
      "loss": 1.2061,
      "theoretical_loss": 3.336189572000527,
      "tokens_seen": 3158376448
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.343029087261786e-05,
      "loss": 1.2245,
      "theoretical_loss": 3.3361843499254045,
      "tokens_seen": 3158441984
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3410230692076234e-05,
      "loss": 1.1925,
      "theoretical_loss": 3.336179127988975,
      "tokens_seen": 3158507520
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.33901705115346e-05,
      "loss": 1.1961,
      "theoretical_loss": 3.336173906191232,
      "tokens_seen": 3158573056
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.337011033099298e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.3361686845321685,
      "tokens_seen": 3158638592
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3350050150451354e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.3361634630117782,
      "tokens_seen": 3158704128
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.332998996990973e-05,
      "loss": 1.1975,
      "theoretical_loss": 3.336158241630055,
      "tokens_seen": 3158769664
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3309929789368106e-05,
      "loss": 1.2201,
      "theoretical_loss": 3.336153020386991,
      "tokens_seen": 3158835200
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3289869608826475e-05,
      "loss": 1.2286,
      "theoretical_loss": 3.3361477992825814,
      "tokens_seen": 3158900736
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.326980942828485e-05,
      "loss": 1.1963,
      "theoretical_loss": 3.3361425783168186,
      "tokens_seen": 3158966272
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.324974924774323e-05,
      "loss": 1.3,
      "theoretical_loss": 3.3361373574896964,
      "tokens_seen": 3159031808
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.322968906720161e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.3361321368012082,
      "tokens_seen": 3159097344
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3209628886659985e-05,
      "loss": 1.2431,
      "theoretical_loss": 3.336126916251347,
      "tokens_seen": 3159162880
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3189568706118354e-05,
      "loss": 1.2237,
      "theoretical_loss": 3.3361216958401068,
      "tokens_seen": 3159228416
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.316950852557673e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.336116475567481,
      "tokens_seen": 3159293952
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3149448345035105e-05,
      "loss": 1.1768,
      "theoretical_loss": 3.3361112554334627,
      "tokens_seen": 3159359488
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.312938816449348e-05,
      "loss": 1.3441,
      "theoretical_loss": 3.3361060354380454,
      "tokens_seen": 3159425024
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.310932798395186e-05,
      "loss": 1.1866,
      "theoretical_loss": 3.3361008155812226,
      "tokens_seen": 3159490560
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3089267803410226e-05,
      "loss": 1.3557,
      "theoretical_loss": 3.3360955958629877,
      "tokens_seen": 3159556096
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.306920762286861e-05,
      "loss": 1.2246,
      "theoretical_loss": 3.3360903762833347,
      "tokens_seen": 3159621632
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.3049147442326984e-05,
      "loss": 1.2066,
      "theoretical_loss": 3.3360851568422563,
      "tokens_seen": 3159687168
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.302908726178536e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.3360799375397465,
      "tokens_seen": 3159752704
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.3009027081243736e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3360747183757984,
      "tokens_seen": 3159818240
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2988966900702105e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.3360694993504056,
      "tokens_seen": 3159883776
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.296890672016048e-05,
      "loss": 1.291,
      "theoretical_loss": 3.336064280463561,
      "tokens_seen": 3159949312
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2948846539618857e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.3360590617152592,
      "tokens_seen": 3160014848
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.292878635907723e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.3360538431054927,
      "tokens_seen": 3160080384
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.290872617853561e-05,
      "loss": 1.249,
      "theoretical_loss": 3.3360486246342553,
      "tokens_seen": 3160145920
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.288866599799398e-05,
      "loss": 1.1733,
      "theoretical_loss": 3.3360434063015405,
      "tokens_seen": 3160211456
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.286860581745236e-05,
      "loss": 1.3569,
      "theoretical_loss": 3.3360381881073415,
      "tokens_seen": 3160276992
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2848545636910736e-05,
      "loss": 1.3927,
      "theoretical_loss": 3.336032970051652,
      "tokens_seen": 3160342528
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.282848545636911e-05,
      "loss": 1.187,
      "theoretical_loss": 3.336027752134465,
      "tokens_seen": 3160408064
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.280842527582749e-05,
      "loss": 1.1135,
      "theoretical_loss": 3.3360225343557746,
      "tokens_seen": 3160473600
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2788365095285856e-05,
      "loss": 1.3896,
      "theoretical_loss": 3.3360173167155738,
      "tokens_seen": 3160539136
    },
    {
      "epoch": 0.92,
      "objective/train/docs_used": 1148479,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1145994663238525,
      "objective/train/theoretical_loss": 3.3360120992138564,
      "objective/train/tokens_used": 1531063776,
      "theoretical_loss": 3.3360120992138564,
      "tokens_seen": 3160604672
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.276830491474423e-05,
      "loss": 1.1438,
      "theoretical_loss": 3.3360120992138564,
      "tokens_seen": 3160604672
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.274824473420261e-05,
      "loss": 1.2758,
      "theoretical_loss": 3.3360068818506154,
      "tokens_seen": 3160670208
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2728184553660983e-05,
      "loss": 1.1878,
      "theoretical_loss": 3.3360016646258446,
      "tokens_seen": 3160735744
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.270812437311935e-05,
      "loss": 1.173,
      "theoretical_loss": 3.3359964475395376,
      "tokens_seen": 3160801280
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2688064192577735e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.3359912305916875,
      "tokens_seen": 3160866816
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.266800401203611e-05,
      "loss": 1.219,
      "theoretical_loss": 3.3359860137822874,
      "tokens_seen": 3160932352
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.264794383149449e-05,
      "loss": 1.289,
      "theoretical_loss": 3.3359807971113318,
      "tokens_seen": 3160997888
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.262788365095286e-05,
      "loss": 1.1787,
      "theoretical_loss": 3.335975580578814,
      "tokens_seen": 3161063424
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.260782347041123e-05,
      "loss": 1.2942,
      "theoretical_loss": 3.335970364184726,
      "tokens_seen": 3161128960
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.258776328986961e-05,
      "loss": 1.2276,
      "theoretical_loss": 3.3359651479290626,
      "tokens_seen": 3161194496
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.256770310932798e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.3359599318118174,
      "tokens_seen": 3161260032
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.254764292878636e-05,
      "loss": 1.2318,
      "theoretical_loss": 3.3359547158329828,
      "tokens_seen": 3161325568
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2527582748244735e-05,
      "loss": 1.1965,
      "theoretical_loss": 3.3359494999925534,
      "tokens_seen": 3161391104
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.250752256770311e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.335944284290522,
      "tokens_seen": 3161456640
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2487462387161486e-05,
      "loss": 1.2003,
      "theoretical_loss": 3.3359390687268817,
      "tokens_seen": 3161522176
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.246740220661986e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.335933853301627,
      "tokens_seen": 3161587712
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.244734202607824e-05,
      "loss": 1.2995,
      "theoretical_loss": 3.3359286380147504,
      "tokens_seen": 3161653248
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2427281845536614e-05,
      "loss": 1.3452,
      "theoretical_loss": 3.335923422866246,
      "tokens_seen": 3161718784
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.240722166499498e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.335918207856107,
      "tokens_seen": 3161784320
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.238716148445336e-05,
      "loss": 1.3243,
      "theoretical_loss": 3.335912992984327,
      "tokens_seen": 3161849856
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2367101303911734e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.3359077782508995,
      "tokens_seen": 3161915392
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.234704112337011e-05,
      "loss": 1.3339,
      "theoretical_loss": 3.3359025636558175,
      "tokens_seen": 3161980928
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2326980942828486e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.3358973491990747,
      "tokens_seen": 3162046464
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.230692076228686e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.3358921348806647,
      "tokens_seen": 3162112000
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.228686058174524e-05,
      "loss": 1.3976,
      "theoretical_loss": 3.335886920700581,
      "tokens_seen": 3162177536
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.226680040120361e-05,
      "loss": 1.254,
      "theoretical_loss": 3.335881706658817,
      "tokens_seen": 3162243072
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.224674022066199e-05,
      "loss": 1.2119,
      "theoretical_loss": 3.335876492755366,
      "tokens_seen": 3162308608
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2226680040120365e-05,
      "loss": 1.2555,
      "theoretical_loss": 3.3358712789902216,
      "tokens_seen": 3162374144
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2206619859578734e-05,
      "loss": 1.2412,
      "theoretical_loss": 3.335866065363377,
      "tokens_seen": 3162439680
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.218655967903711e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.3358608518748265,
      "tokens_seen": 3162505216
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2166499498495485e-05,
      "loss": 1.189,
      "theoretical_loss": 3.3358556385245626,
      "tokens_seen": 3162570752
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.214643931795386e-05,
      "loss": 1.2102,
      "theoretical_loss": 3.3358504253125796,
      "tokens_seen": 3162636288
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2126379137412244e-05,
      "loss": 1.3473,
      "theoretical_loss": 3.33584521223887,
      "tokens_seen": 3162701824
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.210631895687061e-05,
      "loss": 1.3251,
      "theoretical_loss": 3.335839999303428,
      "tokens_seen": 3162767360
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.208625877632899e-05,
      "loss": 1.2249,
      "theoretical_loss": 3.335834786506247,
      "tokens_seen": 3162832896
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2066198595787364e-05,
      "loss": 1.197,
      "theoretical_loss": 3.33582957384732,
      "tokens_seen": 3162898432
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.204613841524574e-05,
      "loss": 1.2428,
      "theoretical_loss": 3.3358243613266407,
      "tokens_seen": 3162963968
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2026078234704116e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.335819148944203,
      "tokens_seen": 3163029504
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.2006018054162485e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.3358139366999997,
      "tokens_seen": 3163095040
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.198595787362086e-05,
      "loss": 1.3888,
      "theoretical_loss": 3.335808724594025,
      "tokens_seen": 3163160576
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1965897693079237e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.3358035126262715,
      "tokens_seen": 3163226112
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.194583751253762e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.335798300796734,
      "tokens_seen": 3163291648
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.192577733199599e-05,
      "loss": 1.1194,
      "theoretical_loss": 3.335793089105404,
      "tokens_seen": 3163357184
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1905717151454364e-05,
      "loss": 1.2015,
      "theoretical_loss": 3.335787877552277,
      "tokens_seen": 3163422720
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.188565697091274e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.3357826661373453,
      "tokens_seen": 3163488256
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1865596790371115e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.3357774548606023,
      "tokens_seen": 3163553792
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.184553660982949e-05,
      "loss": 1.235,
      "theoretical_loss": 3.335772243722042,
      "tokens_seen": 3163619328
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.182547642928786e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.335767032721658,
      "tokens_seen": 3163684864
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1805416248746236e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.335761821859443,
      "tokens_seen": 3163750400
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.178535606820461e-05,
      "loss": 1.2549,
      "theoretical_loss": 3.335756611135391,
      "tokens_seen": 3163815936
    },
    {
      "epoch": 0.92,
      "objective/train/docs_used": 1149760,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3347498178482056,
      "objective/train/theoretical_loss": 3.3357514005494955,
      "objective/train/tokens_used": 1534340576,
      "theoretical_loss": 3.3357514005494955,
      "tokens_seen": 3163881472
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.176529588766299e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.3357514005494955,
      "tokens_seen": 3163881472
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.174523570712137e-05,
      "loss": 1.1763,
      "theoretical_loss": 3.3357461901017498,
      "tokens_seen": 3163947008
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.172517552657974e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.3357409797921473,
      "tokens_seen": 3164012544
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1705115346038115e-05,
      "loss": 1.2481,
      "theoretical_loss": 3.335735769620682,
      "tokens_seen": 3164078080
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.168505516549649e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.335730559587347,
      "tokens_seen": 3164143616
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.166499498495487e-05,
      "loss": 1.2242,
      "theoretical_loss": 3.3357253496921357,
      "tokens_seen": 3164209152
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.164493480441324e-05,
      "loss": 1.176,
      "theoretical_loss": 3.3357201399350416,
      "tokens_seen": 3164274688
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.162487462387161e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.3357149303160583,
      "tokens_seen": 3164340224
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.160481444332999e-05,
      "loss": 1.1745,
      "theoretical_loss": 3.3357097208351787,
      "tokens_seen": 3164405760
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.158475426278836e-05,
      "loss": 1.2409,
      "theoretical_loss": 3.3357045114923976,
      "tokens_seen": 3164471296
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1564694082246746e-05,
      "loss": 1.2745,
      "theoretical_loss": 3.335699302287707,
      "tokens_seen": 3164536832
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.154463390170512e-05,
      "loss": 1.1979,
      "theoretical_loss": 3.335694093221101,
      "tokens_seen": 3164602368
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.152457372116349e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.335688884292574,
      "tokens_seen": 3164667904
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1504513540621866e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.3356836755021178,
      "tokens_seen": 3164733440
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.148445336008024e-05,
      "loss": 1.1411,
      "theoretical_loss": 3.335678466849727,
      "tokens_seen": 3164798976
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.146439317953862e-05,
      "loss": 1.228,
      "theoretical_loss": 3.3356732583353947,
      "tokens_seen": 3164864512
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1444332998996994e-05,
      "loss": 1.2168,
      "theoretical_loss": 3.3356680499591147,
      "tokens_seen": 3164930048
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.142427281845536e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.33566284172088,
      "tokens_seen": 3164995584
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.140421263791374e-05,
      "loss": 1.2297,
      "theoretical_loss": 3.3356576336206842,
      "tokens_seen": 3165061120
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.138415245737212e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.3356524256585214,
      "tokens_seen": 3165126656
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.13640922768305e-05,
      "loss": 1.1925,
      "theoretical_loss": 3.335647217834384,
      "tokens_seen": 3165192192
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.134403209628887e-05,
      "loss": 1.2672,
      "theoretical_loss": 3.3356420101482667,
      "tokens_seen": 3165257728
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.132397191574724e-05,
      "loss": 1.153,
      "theoretical_loss": 3.335636802600162,
      "tokens_seen": 3165323264
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.130391173520562e-05,
      "loss": 1.1988,
      "theoretical_loss": 3.3356315951900637,
      "tokens_seen": 3165388800
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.128385155466399e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.3356263879179653,
      "tokens_seen": 3165454336
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.126379137412237e-05,
      "loss": 1.1913,
      "theoretical_loss": 3.3356211807838605,
      "tokens_seen": 3165519872
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.124373119358074e-05,
      "loss": 1.1645,
      "theoretical_loss": 3.3356159737877427,
      "tokens_seen": 3165585408
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1223671013039114e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.335610766929605,
      "tokens_seen": 3165650944
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.120361083249749e-05,
      "loss": 1.2076,
      "theoretical_loss": 3.335605560209441,
      "tokens_seen": 3165716480
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.118355065195587e-05,
      "loss": 1.2008,
      "theoretical_loss": 3.3356003536272447,
      "tokens_seen": 3165782016
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.116349047141425e-05,
      "loss": 1.2819,
      "theoretical_loss": 3.3355951471830094,
      "tokens_seen": 3165847552
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.114343029087262e-05,
      "loss": 1.1804,
      "theoretical_loss": 3.335589940876728,
      "tokens_seen": 3165913088
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.112337011033099e-05,
      "loss": 1.2733,
      "theoretical_loss": 3.335584734708395,
      "tokens_seen": 3165978624
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.110330992978937e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.335579528678003,
      "tokens_seen": 3166044160
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1083249749247744e-05,
      "loss": 1.2149,
      "theoretical_loss": 3.3355743227855457,
      "tokens_seen": 3166109696
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.106318956870612e-05,
      "loss": 1.2791,
      "theoretical_loss": 3.3355691170310164,
      "tokens_seen": 3166175232
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.104312938816449e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.3355639114144093,
      "tokens_seen": 3166240768
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1023069207622865e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.3355587059357177,
      "tokens_seen": 3166306304
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.100300902708125e-05,
      "loss": 1.2571,
      "theoretical_loss": 3.3355535005949344,
      "tokens_seen": 3166371840
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.098294884653962e-05,
      "loss": 1.2016,
      "theoretical_loss": 3.3355482953920537,
      "tokens_seen": 3166437376
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0962888665998e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.335543090327069,
      "tokens_seen": 3166502912
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.094282848545637e-05,
      "loss": 1.2303,
      "theoretical_loss": 3.335537885399973,
      "tokens_seen": 3166568448
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0922768304914744e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.33553268061076,
      "tokens_seen": 3166633984
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.090270812437312e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.335527475959423,
      "tokens_seen": 3166699520
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0882647943831495e-05,
      "loss": 1.3382,
      "theoretical_loss": 3.335522271445956,
      "tokens_seen": 3166765056
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.086258776328987e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.3355170670703522,
      "tokens_seen": 3166830592
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.084252758274824e-05,
      "loss": 1.2261,
      "theoretical_loss": 3.335511862832605,
      "tokens_seen": 3166896128
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0822467402206616e-05,
      "loss": 1.2053,
      "theoretical_loss": 3.335506658732708,
      "tokens_seen": 3166961664
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0802407221665e-05,
      "loss": 1.2507,
      "theoretical_loss": 3.3355014547706547,
      "tokens_seen": 3167027200
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0782347041123374e-05,
      "loss": 1.1478,
      "theoretical_loss": 3.3354962509464388,
      "tokens_seen": 3167092736
    },
    {
      "epoch": 0.92,
      "objective/train/docs_used": 1150963,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2981247901916504,
      "objective/train/theoretical_loss": 3.335491047260054,
      "objective/train/tokens_used": 1537617376,
      "theoretical_loss": 3.335491047260054,
      "tokens_seen": 3167158272
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.076228686058175e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.335491047260054,
      "tokens_seen": 3167158272
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.074222668004012e-05,
      "loss": 1.3542,
      "theoretical_loss": 3.335485843711493,
      "tokens_seen": 3167223808
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0722166499498495e-05,
      "loss": 1.2101,
      "theoretical_loss": 3.3354806403007493,
      "tokens_seen": 3167289344
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.070210631895687e-05,
      "loss": 1.2149,
      "theoretical_loss": 3.3354754370278172,
      "tokens_seen": 3167354880
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.068204613841525e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.3354702338926896,
      "tokens_seen": 3167420416
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.066198595787362e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.33546503089536,
      "tokens_seen": 3167485952
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.064192577733199e-05,
      "loss": 1.3479,
      "theoretical_loss": 3.3354598280358227,
      "tokens_seen": 3167551488
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0621865596790374e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.33545462531407,
      "tokens_seen": 3167617024
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.060180541624875e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.335449422730097,
      "tokens_seen": 3167682560
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0581745235707126e-05,
      "loss": 1.2071,
      "theoretical_loss": 3.335444220283895,
      "tokens_seen": 3167748096
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.05616850551655e-05,
      "loss": 1.2345,
      "theoretical_loss": 3.3354390179754594,
      "tokens_seen": 3167813632
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.054162487462387e-05,
      "loss": 1.182,
      "theoretical_loss": 3.335433815804783,
      "tokens_seen": 3167879168
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0521564694082246e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.335428613771859,
      "tokens_seen": 3167944704
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.050150451354062e-05,
      "loss": 1.1854,
      "theoretical_loss": 3.3354234118766817,
      "tokens_seen": 3168010240
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0481444332999e-05,
      "loss": 1.3149,
      "theoretical_loss": 3.3354182101192436,
      "tokens_seen": 3168075776
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.046138415245737e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.335413008499539,
      "tokens_seen": 3168141312
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.044132397191575e-05,
      "loss": 1.2084,
      "theoretical_loss": 3.335407807017561,
      "tokens_seen": 3168206848
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0421263791374125e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.335402605673303,
      "tokens_seen": 3168272384
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.04012036108325e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.335397404466759,
      "tokens_seen": 3168337920
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.038114343029088e-05,
      "loss": 1.3029,
      "theoretical_loss": 3.335392203397922,
      "tokens_seen": 3168403456
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0361083249749246e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.3353870024667858,
      "tokens_seen": 3168468992
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.034102306920762e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.335381801673344,
      "tokens_seen": 3168534528
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0320962888666e-05,
      "loss": 1.4167,
      "theoretical_loss": 3.33537660101759,
      "tokens_seen": 3168600064
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.030090270812437e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.3353714004995174,
      "tokens_seen": 3168665600
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.028084252758275e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.335366200119119,
      "tokens_seen": 3168731136
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.026078234704112e-05,
      "loss": 1.298,
      "theoretical_loss": 3.3353609998763893,
      "tokens_seen": 3168796672
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.02407221664995e-05,
      "loss": 1.1669,
      "theoretical_loss": 3.3353557997713215,
      "tokens_seen": 3168862208
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0220661985957876e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.335350599803909,
      "tokens_seen": 3168927744
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.020060180541625e-05,
      "loss": 1.3765,
      "theoretical_loss": 3.3353453999741447,
      "tokens_seen": 3168993280
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.018054162487463e-05,
      "loss": 1.2997,
      "theoretical_loss": 3.3353402002820234,
      "tokens_seen": 3169058816
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0160481444333e-05,
      "loss": 1.3257,
      "theoretical_loss": 3.3353350007275377,
      "tokens_seen": 3169124352
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.014042126379137e-05,
      "loss": 1.3571,
      "theoretical_loss": 3.335329801310681,
      "tokens_seen": 3169189888
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.012036108324975e-05,
      "loss": 1.3639,
      "theoretical_loss": 3.3353246020314478,
      "tokens_seen": 3169255424
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0100300902708124e-05,
      "loss": 1.279,
      "theoretical_loss": 3.33531940288983,
      "tokens_seen": 3169320960
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.00802407221665e-05,
      "loss": 1.2148,
      "theoretical_loss": 3.3353142038858232,
      "tokens_seen": 3169386496
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.0060180541624876e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.335309005019419,
      "tokens_seen": 3169452032
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.004012036108325e-05,
      "loss": 1.2083,
      "theoretical_loss": 3.3353038062906117,
      "tokens_seen": 3169517568
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.002006018054163e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.335298607699395,
      "tokens_seen": 3169583104
    },
    {
      "epoch": 0.92,
      "learning_rate": 4e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.3352934092457627,
      "tokens_seen": 3169648640
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.997993981945838e-05,
      "loss": 1.2224,
      "theoretical_loss": 3.3352882109297073,
      "tokens_seen": 3169714176
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.995987963891675e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.3352830127512227,
      "tokens_seen": 3169779712
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9939819458375124e-05,
      "loss": 1.187,
      "theoretical_loss": 3.335277814710303,
      "tokens_seen": 3169845248
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.99197592778335e-05,
      "loss": 1.3572,
      "theoretical_loss": 3.335272616806941,
      "tokens_seen": 3169910784
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9899699097291875e-05,
      "loss": 1.397,
      "theoretical_loss": 3.3352674190411307,
      "tokens_seen": 3169976320
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.987963891675026e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.335262221412865,
      "tokens_seen": 3170041856
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.985957873620863e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.3352570239221384,
      "tokens_seen": 3170107392
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9839518555667e-05,
      "loss": 1.1672,
      "theoretical_loss": 3.3352518265689435,
      "tokens_seen": 3170172928
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.981945837512538e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.335246629353274,
      "tokens_seen": 3170238464
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9799398194583754e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.335241432275124,
      "tokens_seen": 3170304000
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9779338014042123e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.335236235334486,
      "tokens_seen": 3170369536
    },
    {
      "epoch": 0.92,
      "objective/train/docs_used": 1152259,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.282792329788208,
      "objective/train/theoretical_loss": 3.335231038531355,
      "objective/train/tokens_used": 1540894176,
      "theoretical_loss": 3.335231038531355,
      "tokens_seen": 3170435072
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.97592778335005e-05,
      "loss": 1.2144,
      "theoretical_loss": 3.335231038531355,
      "tokens_seen": 3170435072
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9739217652958875e-05,
      "loss": 1.3341,
      "theoretical_loss": 3.335225841865723,
      "tokens_seen": 3170500608
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.971915747241725e-05,
      "loss": 1.4197,
      "theoretical_loss": 3.335220645337584,
      "tokens_seen": 3170566144
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.969909729187563e-05,
      "loss": 1.1527,
      "theoretical_loss": 3.3352154489469323,
      "tokens_seen": 3170631680
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9679037111334e-05,
      "loss": 1.2158,
      "theoretical_loss": 3.33521025269376,
      "tokens_seen": 3170697216
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.965897693079238e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.3352050565780624,
      "tokens_seen": 3170762752
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9638916750250754e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.3351998605998316,
      "tokens_seen": 3170828288
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.961885656970913e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.3351946647590616,
      "tokens_seen": 3170893824
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9598796389167506e-05,
      "loss": 1.1965,
      "theoretical_loss": 3.3351894690557455,
      "tokens_seen": 3170959360
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9578736208625875e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.3351842734898773,
      "tokens_seen": 3171024896
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.955867602808425e-05,
      "loss": 1.2452,
      "theoretical_loss": 3.3351790780614508,
      "tokens_seen": 3171090432
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9538615847542626e-05,
      "loss": 1.2613,
      "theoretical_loss": 3.335173882770459,
      "tokens_seen": 3171155968
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9518555667001e-05,
      "loss": 1.2146,
      "theoretical_loss": 3.3351686876168958,
      "tokens_seen": 3171221504
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9498495486459385e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.335163492600754,
      "tokens_seen": 3171287040
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9478435305917754e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.335158297722028,
      "tokens_seen": 3171352576
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.945837512537613e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.335153102980711,
      "tokens_seen": 3171418112
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9438314944834505e-05,
      "loss": 1.2763,
      "theoretical_loss": 3.335147908376796,
      "tokens_seen": 3171483648
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.941825476429288e-05,
      "loss": 1.2466,
      "theoretical_loss": 3.335142713910278,
      "tokens_seen": 3171549184
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.939819458375126e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.3351375195811483,
      "tokens_seen": 3171614720
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9378134403209626e-05,
      "loss": 1.2489,
      "theoretical_loss": 3.3351323253894027,
      "tokens_seen": 3171680256
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9358074222668e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.335127131335033,
      "tokens_seen": 3171745792
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.933801404212638e-05,
      "loss": 1.2081,
      "theoretical_loss": 3.335121937418034,
      "tokens_seen": 3171811328
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.931795386158476e-05,
      "loss": 1.271,
      "theoretical_loss": 3.335116743638398,
      "tokens_seen": 3171876864
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9297893681043136e-05,
      "loss": 1.2297,
      "theoretical_loss": 3.3351115499961197,
      "tokens_seen": 3171942400
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9277833500501505e-05,
      "loss": 1.1892,
      "theoretical_loss": 3.335106356491192,
      "tokens_seen": 3172007936
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.925777331995988e-05,
      "loss": 1.2346,
      "theoretical_loss": 3.3351011631236083,
      "tokens_seen": 3172073472
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9237713139418256e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.3350959698933624,
      "tokens_seen": 3172139008
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.921765295887663e-05,
      "loss": 1.3634,
      "theoretical_loss": 3.335090776800448,
      "tokens_seen": 3172204544
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.919759277833501e-05,
      "loss": 1.1991,
      "theoretical_loss": 3.3350855838448585,
      "tokens_seen": 3172270080
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.917753259779338e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.335080391026587,
      "tokens_seen": 3172335616
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.915747241725175e-05,
      "loss": 1.2042,
      "theoretical_loss": 3.3350751983456277,
      "tokens_seen": 3172401152
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.913741223671013e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.3350700058019735,
      "tokens_seen": 3172466688
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.911735205616851e-05,
      "loss": 1.1589,
      "theoretical_loss": 3.3350648133956184,
      "tokens_seen": 3172532224
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.909729187562689e-05,
      "loss": 1.2833,
      "theoretical_loss": 3.335059621126556,
      "tokens_seen": 3172597760
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.9077231695085256e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.3350544289947797,
      "tokens_seen": 3172663296
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.905717151454363e-05,
      "loss": 1.3292,
      "theoretical_loss": 3.3350492370002827,
      "tokens_seen": 3172728832
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.903711133400201e-05,
      "loss": 1.3853,
      "theoretical_loss": 3.335044045143059,
      "tokens_seen": 3172794368
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.901705115346038e-05,
      "loss": 1.1974,
      "theoretical_loss": 3.3350388534231015,
      "tokens_seen": 3172859904
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.899699097291875e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.335033661840405,
      "tokens_seen": 3172925440
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.897693079237713e-05,
      "loss": 1.29,
      "theoretical_loss": 3.335028470394961,
      "tokens_seen": 3172990976
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8956870611835504e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.3350232790867653,
      "tokens_seen": 3173056512
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8936810431293886e-05,
      "loss": 1.2265,
      "theoretical_loss": 3.3350180879158096,
      "tokens_seen": 3173122048
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.891675025075226e-05,
      "loss": 1.2382,
      "theoretical_loss": 3.335012896882089,
      "tokens_seen": 3173187584
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.889669007021063e-05,
      "loss": 1.2474,
      "theoretical_loss": 3.335007705985596,
      "tokens_seen": 3173253120
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.887662988966901e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.335002515226324,
      "tokens_seen": 3173318656
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.885656970912738e-05,
      "loss": 1.179,
      "theoretical_loss": 3.334997324604267,
      "tokens_seen": 3173384192
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.883650952858576e-05,
      "loss": 1.3581,
      "theoretical_loss": 3.3349921341194184,
      "tokens_seen": 3173449728
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8816449348044134e-05,
      "loss": 1.2506,
      "theoretical_loss": 3.3349869437717725,
      "tokens_seen": 3173515264
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8796389167502503e-05,
      "loss": 1.2539,
      "theoretical_loss": 3.3349817535613218,
      "tokens_seen": 3173580800
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.877632898696088e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.33497656348806,
      "tokens_seen": 3173646336
    },
    {
      "epoch": 0.92,
      "objective/train/docs_used": 1153509,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.4432129859924316,
      "objective/train/theoretical_loss": 3.334971373551981,
      "objective/train/tokens_used": 1544170976,
      "theoretical_loss": 3.334971373551981,
      "tokens_seen": 3173711872
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.875626880641926e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.334971373551981,
      "tokens_seen": 3173711872
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.873620862587764e-05,
      "loss": 1.22,
      "theoretical_loss": 3.3349661837530777,
      "tokens_seen": 3173777408
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.871614844533601e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.3349609940913445,
      "tokens_seen": 3173842944
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.869608826479438e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.3349558045667744,
      "tokens_seen": 3173908480
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.867602808425276e-05,
      "loss": 1.2465,
      "theoretical_loss": 3.3349506151793613,
      "tokens_seen": 3173974016
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8655967903711134e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.3349454259290985,
      "tokens_seen": 3174039552
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.863590772316951e-05,
      "loss": 1.3327,
      "theoretical_loss": 3.3349402368159797,
      "tokens_seen": 3174105088
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8615847542627886e-05,
      "loss": 1.1824,
      "theoretical_loss": 3.334935047839998,
      "tokens_seen": 3174170624
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8595787362086255e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.3349298590011474,
      "tokens_seen": 3174236160
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.857572718154463e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.3349246702994213,
      "tokens_seen": 3174301696
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.855566700100301e-05,
      "loss": 1.2855,
      "theoretical_loss": 3.3349194817348136,
      "tokens_seen": 3174367232
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.853560682046139e-05,
      "loss": 1.1609,
      "theoretical_loss": 3.334914293307317,
      "tokens_seen": 3174432768
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8515546639919765e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.334909105016926,
      "tokens_seen": 3174498304
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8495486459378134e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.334903916863633,
      "tokens_seen": 3174563840
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.847542627883651e-05,
      "loss": 1.1971,
      "theoretical_loss": 3.334898728847433,
      "tokens_seen": 3174629376
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8455366098294885e-05,
      "loss": 1.2087,
      "theoretical_loss": 3.3348935409683187,
      "tokens_seen": 3174694912
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.843530591775326e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.3348883532262836,
      "tokens_seen": 3174760448
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.841524573721164e-05,
      "loss": 1.2814,
      "theoretical_loss": 3.334883165621321,
      "tokens_seen": 3174825984
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8395185556670006e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.3348779781534255,
      "tokens_seen": 3174891520
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.837512537612839e-05,
      "loss": 1.262,
      "theoretical_loss": 3.3348727908225895,
      "tokens_seen": 3174957056
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8355065195586764e-05,
      "loss": 1.1667,
      "theoretical_loss": 3.3348676036288074,
      "tokens_seen": 3175022592
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.833500501504514e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.334862416572072,
      "tokens_seen": 3175088128
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8314944834503516e-05,
      "loss": 1.1941,
      "theoretical_loss": 3.3348572296523775,
      "tokens_seen": 3175153664
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8294884653961885e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.3348520428697173,
      "tokens_seen": 3175219200
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.827482447342026e-05,
      "loss": 1.212,
      "theoretical_loss": 3.3348468562240847,
      "tokens_seen": 3175284736
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8254764292878636e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.3348416697154732,
      "tokens_seen": 3175350272
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.823470411233701e-05,
      "loss": 1.2434,
      "theoretical_loss": 3.334836483343877,
      "tokens_seen": 3175415808
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.821464393179538e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.334831297109289,
      "tokens_seen": 3175481344
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.819458375125376e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.3348261110117026,
      "tokens_seen": 3175546880
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.817452357071214e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.3348209250511123,
      "tokens_seen": 3175612416
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8154463390170515e-05,
      "loss": 1.2059,
      "theoretical_loss": 3.3348157392275106,
      "tokens_seen": 3175677952
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.813440320962889e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.334810553540892,
      "tokens_seen": 3175743488
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.811434302908726e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.3348053679912493,
      "tokens_seen": 3175809024
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.8094282848545636e-05,
      "loss": 1.3618,
      "theoretical_loss": 3.334800182578576,
      "tokens_seen": 3175874560
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.807422266800401e-05,
      "loss": 1.2222,
      "theoretical_loss": 3.334794997302866,
      "tokens_seen": 3175940096
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.805416248746239e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.3347898121641135,
      "tokens_seen": 3176005632
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.803410230692076e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.334784627162311,
      "tokens_seen": 3176071168
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.801404212637913e-05,
      "loss": 1.3293,
      "theoretical_loss": 3.334779442297452,
      "tokens_seen": 3176136704
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.7993981945837515e-05,
      "loss": 1.1686,
      "theoretical_loss": 3.3347742575695314,
      "tokens_seen": 3176202240
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.797392176529589e-05,
      "loss": 1.1868,
      "theoretical_loss": 3.334769072978541,
      "tokens_seen": 3176267776
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7953861584754266e-05,
      "loss": 1.164,
      "theoretical_loss": 3.334763888524476,
      "tokens_seen": 3176333312
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.793380140421264e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.334758704207329,
      "tokens_seen": 3176398848
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.791374122367101e-05,
      "loss": 1.1949,
      "theoretical_loss": 3.334753520027093,
      "tokens_seen": 3176464384
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.789368104312939e-05,
      "loss": 1.1931,
      "theoretical_loss": 3.334748335983763,
      "tokens_seen": 3176529920
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.787362086258776e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.3347431520773316,
      "tokens_seen": 3176595456
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.785356068204614e-05,
      "loss": 1.2451,
      "theoretical_loss": 3.3347379683077927,
      "tokens_seen": 3176660992
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7833500501504514e-05,
      "loss": 1.2177,
      "theoretical_loss": 3.33473278467514,
      "tokens_seen": 3176726528
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.781344032096289e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.3347276011793667,
      "tokens_seen": 3176792064
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7793380140421266e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.3347224178204664,
      "tokens_seen": 3176857600
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.777331995987964e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.3347172345984326,
      "tokens_seen": 3176923136
    },
    {
      "epoch": 0.93,
      "objective/train/docs_used": 1154349,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2232264280319214,
      "objective/train/theoretical_loss": 3.3347120515132596,
      "objective/train/tokens_used": 1547447776,
      "theoretical_loss": 3.3347120515132596,
      "tokens_seen": 3176988672
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.775325977933802e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.3347120515132596,
      "tokens_seen": 3176988672
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.773319959879639e-05,
      "loss": 1.2147,
      "theoretical_loss": 3.3347068685649397,
      "tokens_seen": 3177054208
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.771313941825476e-05,
      "loss": 1.3657,
      "theoretical_loss": 3.3347016857534677,
      "tokens_seen": 3177119744
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.769307923771314e-05,
      "loss": 1.2965,
      "theoretical_loss": 3.3346965030788365,
      "tokens_seen": 3177185280
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7673019057171514e-05,
      "loss": 1.1945,
      "theoretical_loss": 3.33469132054104,
      "tokens_seen": 3177250816
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.765295887662989e-05,
      "loss": 1.3389,
      "theoretical_loss": 3.334686138140071,
      "tokens_seen": 3177316352
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.763289869608826e-05,
      "loss": 1.2169,
      "theoretical_loss": 3.334680955875924,
      "tokens_seen": 3177381888
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.761283851554664e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.334675773748592,
      "tokens_seen": 3177447424
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.759277833500502e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.334670591758069,
      "tokens_seen": 3177512960
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.757271815446339e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.334665409904348,
      "tokens_seen": 3177578496
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.755265797392177e-05,
      "loss": 1.2498,
      "theoretical_loss": 3.334660228187423,
      "tokens_seen": 3177644032
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.753259779338014e-05,
      "loss": 1.2791,
      "theoretical_loss": 3.3346550466072875,
      "tokens_seen": 3177709568
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7512537612838514e-05,
      "loss": 1.2031,
      "theoretical_loss": 3.334649865163935,
      "tokens_seen": 3177775104
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.749247743229689e-05,
      "loss": 1.3,
      "theoretical_loss": 3.334644683857359,
      "tokens_seen": 3177840640
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7472417251755265e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.3346395026875535,
      "tokens_seen": 3177906176
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.745235707121364e-05,
      "loss": 1.2035,
      "theoretical_loss": 3.3346343216545113,
      "tokens_seen": 3177971712
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.743229689067202e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.334629140758227,
      "tokens_seen": 3178037248
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.741223671013039e-05,
      "loss": 1.1984,
      "theoretical_loss": 3.3346239599986927,
      "tokens_seen": 3178102784
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.739217652958877e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.3346187793759032,
      "tokens_seen": 3178168320
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7372116349047144e-05,
      "loss": 1.2096,
      "theoretical_loss": 3.334613598889852,
      "tokens_seen": 3178233856
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.735205616850552e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.3346084185405322,
      "tokens_seen": 3178299392
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.733199598796389e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.3346032383279374,
      "tokens_seen": 3178364928
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7311935807422265e-05,
      "loss": 1.1661,
      "theoretical_loss": 3.3345980582520616,
      "tokens_seen": 3178430464
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.729187562688064e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.3345928783128977,
      "tokens_seen": 3178496000
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7271815446339016e-05,
      "loss": 1.1627,
      "theoretical_loss": 3.33458769851044,
      "tokens_seen": 3178561536
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.72517552657974e-05,
      "loss": 1.268,
      "theoretical_loss": 3.3345825188446816,
      "tokens_seen": 3178627072
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.723169508525577e-05,
      "loss": 1.1921,
      "theoretical_loss": 3.3345773393156164,
      "tokens_seen": 3178692608
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7211634904714144e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.334572159923238,
      "tokens_seen": 3178758144
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.719157472417252e-05,
      "loss": 1.2662,
      "theoretical_loss": 3.334566980667539,
      "tokens_seen": 3178823680
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7171514543630895e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.334561801548514,
      "tokens_seen": 3178889216
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.715145436308927e-05,
      "loss": 1.2912,
      "theoretical_loss": 3.334556622566157,
      "tokens_seen": 3178954752
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.713139418254764e-05,
      "loss": 1.1587,
      "theoretical_loss": 3.33455144372046,
      "tokens_seen": 3179020288
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7111334002006016e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.3345462650114177,
      "tokens_seen": 3179085824
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.709127382146439e-05,
      "loss": 1.2594,
      "theoretical_loss": 3.3345410864390237,
      "tokens_seen": 3179151360
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.707121364092277e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.3345359080032715,
      "tokens_seen": 3179216896
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.705115346038115e-05,
      "loss": 1.2698,
      "theoretical_loss": 3.334530729704154,
      "tokens_seen": 3179282432
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.703109327983952e-05,
      "loss": 1.2054,
      "theoretical_loss": 3.3345255515416654,
      "tokens_seen": 3179347968
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.7011033099297895e-05,
      "loss": 1.354,
      "theoretical_loss": 3.334520373515799,
      "tokens_seen": 3179413504
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.699097291875627e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.3345151956265493,
      "tokens_seen": 3179479040
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6970912738214646e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.3345100178739084,
      "tokens_seen": 3179544576
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.695085255767302e-05,
      "loss": 1.1916,
      "theoretical_loss": 3.3345048402578708,
      "tokens_seen": 3179610112
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.693079237713139e-05,
      "loss": 1.3089,
      "theoretical_loss": 3.33449966277843,
      "tokens_seen": 3179675648
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.691073219658977e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.334494485435579,
      "tokens_seen": 3179741184
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.689067201604814e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.3344893082293123,
      "tokens_seen": 3179806720
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6870611835506525e-05,
      "loss": 1.3164,
      "theoretical_loss": 3.334484131159623,
      "tokens_seen": 3179872256
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.68505516549649e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.3344789542265048,
      "tokens_seen": 3179937792
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.683049147442327e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.334473777429951,
      "tokens_seen": 3180003328
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6810431293881646e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.3344686007699553,
      "tokens_seen": 3180068864
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.679037111334002e-05,
      "loss": 1.2594,
      "theoretical_loss": 3.3344634242465117,
      "tokens_seen": 3180134400
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.67703109327984e-05,
      "loss": 1.3712,
      "theoretical_loss": 3.334458247859613,
      "tokens_seen": 3180199936
    },
    {
      "epoch": 0.93,
      "objective/train/docs_used": 1155539,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3495402336120605,
      "objective/train/theoretical_loss": 3.3344530716092535,
      "objective/train/tokens_used": 1550724576,
      "theoretical_loss": 3.3344530716092535,
      "tokens_seen": 3180265472
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6750250752256767e-05,
      "loss": 1.3193,
      "theoretical_loss": 3.3344530716092535,
      "tokens_seen": 3180265472
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.673019057171514e-05,
      "loss": 1.343,
      "theoretical_loss": 3.334447895495426,
      "tokens_seen": 3180331008
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.671013039117352e-05,
      "loss": 1.2165,
      "theoretical_loss": 3.334442719518125,
      "tokens_seen": 3180396544
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.66900702106319e-05,
      "loss": 1.206,
      "theoretical_loss": 3.334437543677344,
      "tokens_seen": 3180462080
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6670010030090277e-05,
      "loss": 1.3021,
      "theoretical_loss": 3.334432367973076,
      "tokens_seen": 3180527616
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6649949849548646e-05,
      "loss": 1.3859,
      "theoretical_loss": 3.3344271924053146,
      "tokens_seen": 3180593152
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.662988966900702e-05,
      "loss": 1.1776,
      "theoretical_loss": 3.334422016974054,
      "tokens_seen": 3180658688
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.66098294884654e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.334416841679287,
      "tokens_seen": 3180724224
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.658976930792377e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.3344116665210084,
      "tokens_seen": 3180789760
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.656970912738215e-05,
      "loss": 1.1545,
      "theoretical_loss": 3.33440649149921,
      "tokens_seen": 3180855296
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.654964894684052e-05,
      "loss": 1.1856,
      "theoretical_loss": 3.3344013166138873,
      "tokens_seen": 3180920832
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6529588766298894e-05,
      "loss": 1.241,
      "theoretical_loss": 3.3343961418650325,
      "tokens_seen": 3180986368
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.650952858575727e-05,
      "loss": 1.191,
      "theoretical_loss": 3.3343909672526397,
      "tokens_seen": 3181051904
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.648946840521565e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.334385792776702,
      "tokens_seen": 3181117440
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.646940822467403e-05,
      "loss": 1.3139,
      "theoretical_loss": 3.3343806184372147,
      "tokens_seen": 3181182976
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.64493480441324e-05,
      "loss": 1.2336,
      "theoretical_loss": 3.334375444234169,
      "tokens_seen": 3181248512
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.642928786359077e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.33437027016756,
      "tokens_seen": 3181314048
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.640922768304915e-05,
      "loss": 1.3485,
      "theoretical_loss": 3.334365096237381,
      "tokens_seen": 3181379584
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6389167502507524e-05,
      "loss": 1.206,
      "theoretical_loss": 3.3343599224436256,
      "tokens_seen": 3181445120
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.63691073219659e-05,
      "loss": 1.2412,
      "theoretical_loss": 3.3343547487862875,
      "tokens_seen": 3181510656
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.634904714142427e-05,
      "loss": 1.252,
      "theoretical_loss": 3.3343495752653594,
      "tokens_seen": 3181576192
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6328986960882645e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.334344401880836,
      "tokens_seen": 3181641728
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.630892678034103e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.334339228632711,
      "tokens_seen": 3181707264
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.62888665997994e-05,
      "loss": 1.2271,
      "theoretical_loss": 3.3343340555209764,
      "tokens_seen": 3181772800
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.626880641925778e-05,
      "loss": 1.2566,
      "theoretical_loss": 3.334328882545628,
      "tokens_seen": 3181838336
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.624874623871615e-05,
      "loss": 1.221,
      "theoretical_loss": 3.3343237097066574,
      "tokens_seen": 3181903872
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6228686058174524e-05,
      "loss": 1.2555,
      "theoretical_loss": 3.3343185370040596,
      "tokens_seen": 3181969408
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.62086258776329e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.3343133644378273,
      "tokens_seen": 3182034944
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6188565697091275e-05,
      "loss": 1.303,
      "theoretical_loss": 3.3343081920079545,
      "tokens_seen": 3182100480
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.616850551654965e-05,
      "loss": 1.2302,
      "theoretical_loss": 3.334303019714435,
      "tokens_seen": 3182166016
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.614844533600802e-05,
      "loss": 1.3526,
      "theoretical_loss": 3.334297847557262,
      "tokens_seen": 3182231552
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.61283851554664e-05,
      "loss": 1.1723,
      "theoretical_loss": 3.3342926755364295,
      "tokens_seen": 3182297088
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.610832497492478e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3342875036519306,
      "tokens_seen": 3182362624
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6088264794383154e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.3342823319037596,
      "tokens_seen": 3182428160
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.606820461384152e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.334277160291909,
      "tokens_seen": 3182493696
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.60481444332999e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.3342719888163734,
      "tokens_seen": 3182559232
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.6028084252758275e-05,
      "loss": 1.3828,
      "theoretical_loss": 3.334266817477146,
      "tokens_seen": 3182624768
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.600802407221665e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.3342616462742205,
      "tokens_seen": 3182690304
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5987963891675026e-05,
      "loss": 1.28,
      "theoretical_loss": 3.3342564752075905,
      "tokens_seen": 3182755840
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5967903711133395e-05,
      "loss": 1.2742,
      "theoretical_loss": 3.3342513042772493,
      "tokens_seen": 3182821376
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.594784353059177e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.334246133483191,
      "tokens_seen": 3182886912
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5927783350050154e-05,
      "loss": 1.2238,
      "theoretical_loss": 3.334240962825409,
      "tokens_seen": 3182952448
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.590772316950853e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.3342357923038968,
      "tokens_seen": 3183017984
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5887662988966905e-05,
      "loss": 1.3255,
      "theoretical_loss": 3.334230621918648,
      "tokens_seen": 3183083520
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5867602808425274e-05,
      "loss": 1.2388,
      "theoretical_loss": 3.3342254516696563,
      "tokens_seen": 3183149056
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.584754262788365e-05,
      "loss": 1.22,
      "theoretical_loss": 3.3342202815569153,
      "tokens_seen": 3183214592
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5827482447342026e-05,
      "loss": 1.237,
      "theoretical_loss": 3.334215111580418,
      "tokens_seen": 3183280128
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.58074222668004e-05,
      "loss": 1.2283,
      "theoretical_loss": 3.3342099417401596,
      "tokens_seen": 3183345664
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.578736208625878e-05,
      "loss": 1.3675,
      "theoretical_loss": 3.334204772036132,
      "tokens_seen": 3183411200
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5767301905717147e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.33419960246833,
      "tokens_seen": 3183476736
    },
    {
      "epoch": 0.93,
      "objective/train/docs_used": 1156704,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2980313301086426,
      "objective/train/theoretical_loss": 3.334194433036746,
      "objective/train/tokens_used": 1554001376,
      "theoretical_loss": 3.334194433036746,
      "tokens_seen": 3183542272
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.574724172517553e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.334194433036746,
      "tokens_seen": 3183542272
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5727181544633905e-05,
      "loss": 1.2307,
      "theoretical_loss": 3.334189263741375,
      "tokens_seen": 3183607808
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.570712136409228e-05,
      "loss": 1.2814,
      "theoretical_loss": 3.3341840945822097,
      "tokens_seen": 3183673344
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5687061183550657e-05,
      "loss": 1.264,
      "theoretical_loss": 3.334178925559244,
      "tokens_seen": 3183738880
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5667001003009026e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.334173756672471,
      "tokens_seen": 3183804416
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.56469408224674e-05,
      "loss": 1.2419,
      "theoretical_loss": 3.3341685879218854,
      "tokens_seen": 3183869952
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.562688064192578e-05,
      "loss": 1.1863,
      "theoretical_loss": 3.3341634193074796,
      "tokens_seen": 3183935488
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.560682046138415e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.334158250829248,
      "tokens_seen": 3184001024
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.558676028084253e-05,
      "loss": 1.1897,
      "theoretical_loss": 3.334153082487184,
      "tokens_seen": 3184066560
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5566700100300904e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.3341479142812807,
      "tokens_seen": 3184132096
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.554663991975928e-05,
      "loss": 1.1599,
      "theoretical_loss": 3.3341427462115325,
      "tokens_seen": 3184197632
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5526579739217656e-05,
      "loss": 1.3222,
      "theoretical_loss": 3.334137578277933,
      "tokens_seen": 3184263168
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.550651955867603e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.3341324104804753,
      "tokens_seen": 3184328704
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.548645937813441e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.334127242819153,
      "tokens_seen": 3184394240
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.546639919759278e-05,
      "loss": 1.2733,
      "theoretical_loss": 3.33412207529396,
      "tokens_seen": 3184459776
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.544633901705115e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.3341169079048902,
      "tokens_seen": 3184525312
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.542627883650953e-05,
      "loss": 1.2709,
      "theoretical_loss": 3.3341117406519367,
      "tokens_seen": 3184590848
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5406218655967904e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.334106573535093,
      "tokens_seen": 3184656384
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.538615847542627e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.3341014065543533,
      "tokens_seen": 3184721920
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5366098294884656e-05,
      "loss": 1.2395,
      "theoretical_loss": 3.3340962397097105,
      "tokens_seen": 3184787456
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.534603811434303e-05,
      "loss": 1.2537,
      "theoretical_loss": 3.334091073001159,
      "tokens_seen": 3184852992
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.532597793380141e-05,
      "loss": 1.24,
      "theoretical_loss": 3.334085906428692,
      "tokens_seen": 3184918528
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.530591775325978e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.334080739992303,
      "tokens_seen": 3184984064
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.528585757271815e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.334075573691986,
      "tokens_seen": 3185049600
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.526579739217653e-05,
      "loss": 1.1736,
      "theoretical_loss": 3.334070407527734,
      "tokens_seen": 3185115136
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5245737211634904e-05,
      "loss": 1.204,
      "theoretical_loss": 3.334065241499541,
      "tokens_seen": 3185180672
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.522567703109328e-05,
      "loss": 1.3741,
      "theoretical_loss": 3.3340600756074013,
      "tokens_seen": 3185246208
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5205616850551655e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.3340549098513073,
      "tokens_seen": 3185311744
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.518555667001003e-05,
      "loss": 1.2723,
      "theoretical_loss": 3.334049744231253,
      "tokens_seen": 3185377280
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.516549648946841e-05,
      "loss": 1.4063,
      "theoretical_loss": 3.3340445787472324,
      "tokens_seen": 3185442816
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.514543630892678e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3340394133992386,
      "tokens_seen": 3185508352
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.512537612838516e-05,
      "loss": 1.232,
      "theoretical_loss": 3.334034248187266,
      "tokens_seen": 3185573888
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5105315947843534e-05,
      "loss": 1.3486,
      "theoretical_loss": 3.3340290831113073,
      "tokens_seen": 3185639424
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.50852557673019e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.334023918171357,
      "tokens_seen": 3185704960
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.506519558676028e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.3340187533674075,
      "tokens_seen": 3185770496
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5045135406218655e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.3340135886994537,
      "tokens_seen": 3185836032
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.502507522567703e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.3340084241674885,
      "tokens_seen": 3185901568
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.5005015045135406e-05,
      "loss": 1.3337,
      "theoretical_loss": 3.334003259771506,
      "tokens_seen": 3185967104
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.498495486459378e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.3339980955114994,
      "tokens_seen": 3186032640
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.496489468405216e-05,
      "loss": 1.26,
      "theoretical_loss": 3.3339929313874626,
      "tokens_seen": 3186098176
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4944834503510534e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.333987767399389,
      "tokens_seen": 3186163712
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.492477432296891e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.3339826035472724,
      "tokens_seen": 3186229248
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4904714142427285e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.333977439831106,
      "tokens_seen": 3186294784
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4884653961885654e-05,
      "loss": 1.325,
      "theoretical_loss": 3.3339722762508845,
      "tokens_seen": 3186360320
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.486459378134403e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.3339671128066004,
      "tokens_seen": 3186425856
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4844533600802406e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.3339619494982475,
      "tokens_seen": 3186491392
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.482447342026078e-05,
      "loss": 1.3368,
      "theoretical_loss": 3.3339567863258197,
      "tokens_seen": 3186556928
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4804413239719164e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.3339516232893107,
      "tokens_seen": 3186622464
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.478435305917753e-05,
      "loss": 1.076,
      "theoretical_loss": 3.3339464603887143,
      "tokens_seen": 3186688000
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.476429287863591e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.3339412976240235,
      "tokens_seen": 3186753536
    },
    {
      "epoch": 0.93,
      "objective/train/docs_used": 1158346,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2709064483642578,
      "objective/train/theoretical_loss": 3.3339361349952323,
      "objective/train/tokens_used": 1557278176,
      "theoretical_loss": 3.3339361349952323,
      "tokens_seen": 3186819072
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4744232698094285e-05,
      "loss": 1.272,
      "theoretical_loss": 3.3339361349952323,
      "tokens_seen": 3186819072
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.472417251755266e-05,
      "loss": 1.2636,
      "theoretical_loss": 3.3339309725023343,
      "tokens_seen": 3186884608
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4704112337011036e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.333925810145323,
      "tokens_seen": 3186950144
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4684052156469406e-05,
      "loss": 1.217,
      "theoretical_loss": 3.333920647924192,
      "tokens_seen": 3187015680
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.466399197592778e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.3339154858389355,
      "tokens_seen": 3187081216
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.464393179538616e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.3339103238895467,
      "tokens_seen": 3187146752
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.462387161484454e-05,
      "loss": 1.2361,
      "theoretical_loss": 3.333905162076019,
      "tokens_seen": 3187212288
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.460381143430291e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.333900000398346,
      "tokens_seen": 3187277824
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4583751253761284e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.333894838856522,
      "tokens_seen": 3187343360
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.456369107321966e-05,
      "loss": 1.2766,
      "theoretical_loss": 3.33388967745054,
      "tokens_seen": 3187408896
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4543630892678036e-05,
      "loss": 1.2068,
      "theoretical_loss": 3.333884516180394,
      "tokens_seen": 3187474432
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.452357071213641e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.3338793550460775,
      "tokens_seen": 3187539968
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.450351053159478e-05,
      "loss": 1.3271,
      "theoretical_loss": 3.333874194047584,
      "tokens_seen": 3187605504
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.448345035105316e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.3338690331849072,
      "tokens_seen": 3187671040
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.446339017051153e-05,
      "loss": 1.3184,
      "theoretical_loss": 3.3338638724580414,
      "tokens_seen": 3187736576
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.444332998996991e-05,
      "loss": 1.3118,
      "theoretical_loss": 3.333858711866979,
      "tokens_seen": 3187802112
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.442326980942829e-05,
      "loss": 1.2715,
      "theoretical_loss": 3.3338535514117145,
      "tokens_seen": 3187867648
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.440320962888666e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.333848391092241,
      "tokens_seen": 3187933184
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4383149448345036e-05,
      "loss": 1.3667,
      "theoretical_loss": 3.3338432309085526,
      "tokens_seen": 3187998720
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.436308926780341e-05,
      "loss": 1.3637,
      "theoretical_loss": 3.3338380708606428,
      "tokens_seen": 3188064256
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.434302908726179e-05,
      "loss": 1.2311,
      "theoretical_loss": 3.333832910948505,
      "tokens_seen": 3188129792
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.432296890672016e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.3338277511721337,
      "tokens_seen": 3188195328
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.430290872617853e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.333822591531521,
      "tokens_seen": 3188260864
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.428284854563691e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.3338174320266623,
      "tokens_seen": 3188326400
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4262788365095284e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.3338122726575494,
      "tokens_seen": 3188391936
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4242728184553666e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.3338071134241773,
      "tokens_seen": 3188457472
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.422266800401204e-05,
      "loss": 1.229,
      "theoretical_loss": 3.3338019543265394,
      "tokens_seen": 3188523008
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.420260782347041e-05,
      "loss": 1.2089,
      "theoretical_loss": 3.333796795364629,
      "tokens_seen": 3188588544
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.418254764292879e-05,
      "loss": 1.2151,
      "theoretical_loss": 3.33379163653844,
      "tokens_seen": 3188654080
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.416248746238716e-05,
      "loss": 1.3197,
      "theoretical_loss": 3.333786477847966,
      "tokens_seen": 3188719616
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.414242728184554e-05,
      "loss": 1.2131,
      "theoretical_loss": 3.3337813192932004,
      "tokens_seen": 3188785152
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.4122367101303914e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.333776160874137,
      "tokens_seen": 3188850688
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.410230692076228e-05,
      "loss": 1.2109,
      "theoretical_loss": 3.33377100259077,
      "tokens_seen": 3188916224
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.408224674022066e-05,
      "loss": 1.2029,
      "theoretical_loss": 3.333765844443092,
      "tokens_seen": 3188981760
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.406218655967904e-05,
      "loss": 1.2033,
      "theoretical_loss": 3.333760686431097,
      "tokens_seen": 3189047296
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.404212637913742e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.333755528554779,
      "tokens_seen": 3189112832
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.402206619859579e-05,
      "loss": 1.1803,
      "theoretical_loss": 3.3337503708141316,
      "tokens_seen": 3189178368
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.400200601805416e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.333745213209148,
      "tokens_seen": 3189243904
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.398194583751254e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.3337400557398222,
      "tokens_seen": 3189309440
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3961885656970914e-05,
      "loss": 1.1744,
      "theoretical_loss": 3.333734898406148,
      "tokens_seen": 3189374976
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.394182547642929e-05,
      "loss": 1.2601,
      "theoretical_loss": 3.333729741208119,
      "tokens_seen": 3189440512
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.392176529588766e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.333724584145728,
      "tokens_seen": 3189506048
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3901705115346034e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.3337194272189694,
      "tokens_seen": 3189571584
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.388164493480441e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.333714270427837,
      "tokens_seen": 3189637120
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.386158475426279e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.3337091137723243,
      "tokens_seen": 3189702656
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.384152457372117e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.3337039572524247,
      "tokens_seen": 3189768192
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.382146439317954e-05,
      "loss": 1.1622,
      "theoretical_loss": 3.333698800868132,
      "tokens_seen": 3189833728
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.380140421263791e-05,
      "loss": 1.3705,
      "theoretical_loss": 3.33369364461944,
      "tokens_seen": 3189899264
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.378134403209629e-05,
      "loss": 1.286,
      "theoretical_loss": 3.3336884885063416,
      "tokens_seen": 3189964800
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3761283851554665e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.3336833325288318,
      "tokens_seen": 3190030336
    },
    {
      "epoch": 0.93,
      "objective/train/docs_used": 1159129,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2354068756103516,
      "objective/train/theoretical_loss": 3.3336781766869032,
      "objective/train/tokens_used": 1560554976,
      "theoretical_loss": 3.3336781766869032,
      "tokens_seen": 3190095872
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.374122367101304e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.3336781766869032,
      "tokens_seen": 3190095872
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.372116349047141e-05,
      "loss": 1.1774,
      "theoretical_loss": 3.33367302098055,
      "tokens_seen": 3190161408
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3701103309929785e-05,
      "loss": 1.2721,
      "theoretical_loss": 3.333667865409765,
      "tokens_seen": 3190226944
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.368104312938817e-05,
      "loss": 1.2423,
      "theoretical_loss": 3.3336627099745426,
      "tokens_seen": 3190292480
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3660982948846544e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.3336575546748763,
      "tokens_seen": 3190358016
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.364092276830492e-05,
      "loss": 1.2745,
      "theoretical_loss": 3.33365239951076,
      "tokens_seen": 3190423552
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.362086258776329e-05,
      "loss": 1.3171,
      "theoretical_loss": 3.333647244482187,
      "tokens_seen": 3190489088
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3600802407221664e-05,
      "loss": 1.2114,
      "theoretical_loss": 3.333642089589151,
      "tokens_seen": 3190554624
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.358074222668004e-05,
      "loss": 1.225,
      "theoretical_loss": 3.3336369348316457,
      "tokens_seen": 3190620160
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3560682046138416e-05,
      "loss": 1.2766,
      "theoretical_loss": 3.3336317802096644,
      "tokens_seen": 3190685696
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.354062186559679e-05,
      "loss": 1.2729,
      "theoretical_loss": 3.3336266257232015,
      "tokens_seen": 3190751232
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.352056168505516e-05,
      "loss": 1.2037,
      "theoretical_loss": 3.3336214713722505,
      "tokens_seen": 3190816768
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3500501504513543e-05,
      "loss": 1.1866,
      "theoretical_loss": 3.333616317156804,
      "tokens_seen": 3190882304
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.348044132397192e-05,
      "loss": 1.2633,
      "theoretical_loss": 3.3336111630768572,
      "tokens_seen": 3190947840
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3460381143430295e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.333606009132403,
      "tokens_seen": 3191013376
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.344032096288867e-05,
      "loss": 1.3164,
      "theoretical_loss": 3.333600855323435,
      "tokens_seen": 3191078912
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.342026078234704e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.3335957016499465,
      "tokens_seen": 3191144448
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3400200601805416e-05,
      "loss": 1.1958,
      "theoretical_loss": 3.333590548111932,
      "tokens_seen": 3191209984
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.338014042126379e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.3335853947093845,
      "tokens_seen": 3191275520
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.336008024072217e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.3335802414422977,
      "tokens_seen": 3191341056
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.334002006018054e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.333575088310666,
      "tokens_seen": 3191406592
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.331995987963891e-05,
      "loss": 1.222,
      "theoretical_loss": 3.333569935314482,
      "tokens_seen": 3191472128
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3299899699097295e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.3335647824537404,
      "tokens_seen": 3191537664
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.327983951855567e-05,
      "loss": 1.187,
      "theoretical_loss": 3.333559629728434,
      "tokens_seen": 3191603200
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3259779338014046e-05,
      "loss": 1.2778,
      "theoretical_loss": 3.333554477138557,
      "tokens_seen": 3191668736
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.323971915747242e-05,
      "loss": 1.2241,
      "theoretical_loss": 3.333549324684103,
      "tokens_seen": 3191734272
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.321965897693079e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.3335441723650647,
      "tokens_seen": 3191799808
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.319959879638917e-05,
      "loss": 1.2769,
      "theoretical_loss": 3.3335390201814374,
      "tokens_seen": 3191865344
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.317953861584754e-05,
      "loss": 1.3967,
      "theoretical_loss": 3.3335338681332134,
      "tokens_seen": 3191930880
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.315947843530592e-05,
      "loss": 1.1088,
      "theoretical_loss": 3.3335287162203873,
      "tokens_seen": 3191996416
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.313941825476429e-05,
      "loss": 1.2912,
      "theoretical_loss": 3.3335235644429524,
      "tokens_seen": 3192061952
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.311935807422267e-05,
      "loss": 1.1715,
      "theoretical_loss": 3.333518412800902,
      "tokens_seen": 3192127488
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3099297893681046e-05,
      "loss": 1.2432,
      "theoretical_loss": 3.33351326129423,
      "tokens_seen": 3192193024
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.307923771313942e-05,
      "loss": 1.2865,
      "theoretical_loss": 3.3335081099229305,
      "tokens_seen": 3192258560
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.30591775325978e-05,
      "loss": 1.2417,
      "theoretical_loss": 3.3335029586869966,
      "tokens_seen": 3192324096
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.3039117352056166e-05,
      "loss": 1.2539,
      "theoretical_loss": 3.3334978075864226,
      "tokens_seen": 3192389632
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.301905717151454e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.333492656621201,
      "tokens_seen": 3192455168
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.299899699097292e-05,
      "loss": 1.2059,
      "theoretical_loss": 3.333487505791327,
      "tokens_seen": 3192520704
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.2978936810431294e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.333482355096793,
      "tokens_seen": 3192586240
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.295887662988967e-05,
      "loss": 1.2672,
      "theoretical_loss": 3.3334772045375933,
      "tokens_seen": 3192651776
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.2938816449348045e-05,
      "loss": 1.3394,
      "theoretical_loss": 3.333472054113721,
      "tokens_seen": 3192717312
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.291875626880642e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.3334669038251707,
      "tokens_seen": 3192782848
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.28986960882648e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.3334617536719353,
      "tokens_seen": 3192848384
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.287863590772317e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.3334566036540085,
      "tokens_seen": 3192913920
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.285857572718155e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.333451453771384,
      "tokens_seen": 3192979456
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.283851554663992e-05,
      "loss": 1.206,
      "theoretical_loss": 3.3334463040240565,
      "tokens_seen": 3193044992
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.281845536609829e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.3334411544120184,
      "tokens_seen": 3193110528
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.279839518555667e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.3334360049352636,
      "tokens_seen": 3193176064
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2778335005015045e-05,
      "loss": 1.1807,
      "theoretical_loss": 3.333430855593786,
      "tokens_seen": 3193241600
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.275827482447342e-05,
      "loss": 1.2656,
      "theoretical_loss": 3.333425706387579,
      "tokens_seen": 3193307136
    },
    {
      "epoch": 0.94,
      "objective/train/docs_used": 1160256,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.259645938873291,
      "objective/train/theoretical_loss": 3.333420557316637,
      "objective/train/tokens_used": 1563831776,
      "theoretical_loss": 3.333420557316637,
      "tokens_seen": 3193372672
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2738214643931796e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.333420557316637,
      "tokens_seen": 3193372672
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.271815446339017e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.3334154083809526,
      "tokens_seen": 3193438208
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.269809428284855e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.3334102595805204,
      "tokens_seen": 3193503744
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2678034102306924e-05,
      "loss": 1.2857,
      "theoretical_loss": 3.3334051109153338,
      "tokens_seen": 3193569280
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.26579739217653e-05,
      "loss": 1.3352,
      "theoretical_loss": 3.3333999623853865,
      "tokens_seen": 3193634816
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.263791374122367e-05,
      "loss": 1.3427,
      "theoretical_loss": 3.3333948139906715,
      "tokens_seen": 3193700352
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2617853560682044e-05,
      "loss": 1.1445,
      "theoretical_loss": 3.3333896657311834,
      "tokens_seen": 3193765888
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.259779338014042e-05,
      "loss": 1.2172,
      "theoretical_loss": 3.3333845176069152,
      "tokens_seen": 3193831424
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2577733199598796e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.333379369617861,
      "tokens_seen": 3193896960
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.255767301905718e-05,
      "loss": 1.2057,
      "theoretical_loss": 3.3333742217640143,
      "tokens_seen": 3193962496
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.253761283851555e-05,
      "loss": 1.2388,
      "theoretical_loss": 3.333369074045369,
      "tokens_seen": 3194028032
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2517552657973923e-05,
      "loss": 1.2882,
      "theoretical_loss": 3.3333639264619186,
      "tokens_seen": 3194093568
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.24974924774323e-05,
      "loss": 1.2046,
      "theoretical_loss": 3.3333587790136567,
      "tokens_seen": 3194159104
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2477432296890675e-05,
      "loss": 1.213,
      "theoretical_loss": 3.333353631700577,
      "tokens_seen": 3194224640
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2457372116349044e-05,
      "loss": 1.2268,
      "theoretical_loss": 3.333348484522673,
      "tokens_seen": 3194290176
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.243731193580742e-05,
      "loss": 1.2195,
      "theoretical_loss": 3.333343337479939,
      "tokens_seen": 3194355712
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2417251755265796e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.333338190572368,
      "tokens_seen": 3194421248
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.239719157472417e-05,
      "loss": 1.2452,
      "theoretical_loss": 3.333333043799954,
      "tokens_seen": 3194486784
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.237713139418255e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.3333278971626905,
      "tokens_seen": 3194552320
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.235707121364092e-05,
      "loss": 1.3843,
      "theoretical_loss": 3.3333227506605714,
      "tokens_seen": 3194617856
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.23370110330993e-05,
      "loss": 1.3511,
      "theoretical_loss": 3.3333176042935904,
      "tokens_seen": 3194683392
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2316950852557675e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.333312458061741,
      "tokens_seen": 3194748928
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.229689067201605e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.333307311965017,
      "tokens_seen": 3194814464
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2276830491474426e-05,
      "loss": 1.2086,
      "theoretical_loss": 3.3333021660034117,
      "tokens_seen": 3194880000
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2256770310932795e-05,
      "loss": 1.2055,
      "theoretical_loss": 3.3332970201769196,
      "tokens_seen": 3194945536
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.223671013039117e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.3332918744855338,
      "tokens_seen": 3195011072
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.221664994984955e-05,
      "loss": 1.266,
      "theoretical_loss": 3.333286728929248,
      "tokens_seen": 3195076608
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.219658976930792e-05,
      "loss": 1.1838,
      "theoretical_loss": 3.3332815835080556,
      "tokens_seen": 3195142144
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2176529588766305e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.333276438221951,
      "tokens_seen": 3195207680
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2156469408224674e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.3332712930709274,
      "tokens_seen": 3195273216
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.213640922768305e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.3332661480549786,
      "tokens_seen": 3195338752
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2116349047141426e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.333261003174098,
      "tokens_seen": 3195404288
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.20962888665998e-05,
      "loss": 1.1317,
      "theoretical_loss": 3.3332558584282803,
      "tokens_seen": 3195469824
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.207622868605818e-05,
      "loss": 1.2064,
      "theoretical_loss": 3.333250713817518,
      "tokens_seen": 3195535360
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.2056168505516546e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.333245569341805,
      "tokens_seen": 3195600896
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.203610832497492e-05,
      "loss": 1.2109,
      "theoretical_loss": 3.3332404250011356,
      "tokens_seen": 3195666432
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.20160481444333e-05,
      "loss": 1.3276,
      "theoretical_loss": 3.333235280795503,
      "tokens_seen": 3195731968
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.199598796389168e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.333230136724901,
      "tokens_seen": 3195797504
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1975927783350056e-05,
      "loss": 1.3568,
      "theoretical_loss": 3.333224992789323,
      "tokens_seen": 3195863040
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1955867602808425e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.3332198489887634,
      "tokens_seen": 3195928576
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.19358074222668e-05,
      "loss": 1.2124,
      "theoretical_loss": 3.3332147053232153,
      "tokens_seen": 3195994112
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.191574724172518e-05,
      "loss": 1.2309,
      "theoretical_loss": 3.3332095617926725,
      "tokens_seen": 3196059648
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.189568706118355e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.3332044183971288,
      "tokens_seen": 3196125184
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.187562688064193e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.333199275136578,
      "tokens_seen": 3196190720
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.18555667001003e-05,
      "loss": 1.2225,
      "theoretical_loss": 3.333194132011013,
      "tokens_seen": 3196256256
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.183550651955867e-05,
      "loss": 1.3153,
      "theoretical_loss": 3.3331889890204285,
      "tokens_seen": 3196321792
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.181544633901705e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.3331838461648178,
      "tokens_seen": 3196387328
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.179538615847543e-05,
      "loss": 1.2266,
      "theoretical_loss": 3.3331787034441747,
      "tokens_seen": 3196452864
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.177532597793381e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.3331735608584925,
      "tokens_seen": 3196518400
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1755265797392176e-05,
      "loss": 1.1787,
      "theoretical_loss": 3.333168418407765,
      "tokens_seen": 3196583936
    },
    {
      "epoch": 0.94,
      "objective/train/docs_used": 1161767,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1227366924285889,
      "objective/train/theoretical_loss": 3.3331632760919865,
      "objective/train/tokens_used": 1567108576,
      "theoretical_loss": 3.3331632760919865,
      "tokens_seen": 3196649472
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.173520561685055e-05,
      "loss": 1.2256,
      "theoretical_loss": 3.3331632760919865,
      "tokens_seen": 3196649472
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.171514543630893e-05,
      "loss": 1.1919,
      "theoretical_loss": 3.3331581339111502,
      "tokens_seen": 3196715008
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1695085255767304e-05,
      "loss": 1.1696,
      "theoretical_loss": 3.3331529918652496,
      "tokens_seen": 3196780544
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.167502507522567e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.333147849954279,
      "tokens_seen": 3196846080
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.165496489468405e-05,
      "loss": 1.1894,
      "theoretical_loss": 3.333142708178231,
      "tokens_seen": 3196911616
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1634904714142424e-05,
      "loss": 1.163,
      "theoretical_loss": 3.3331375665371006,
      "tokens_seen": 3196977152
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.161484453360081e-05,
      "loss": 1.2177,
      "theoretical_loss": 3.3331324250308807,
      "tokens_seen": 3197042688
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.159478435305918e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.3331272836595653,
      "tokens_seen": 3197108224
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.157472417251755e-05,
      "loss": 1.2079,
      "theoretical_loss": 3.3331221424231483,
      "tokens_seen": 3197173760
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.155466399197593e-05,
      "loss": 1.1817,
      "theoretical_loss": 3.3331170013216225,
      "tokens_seen": 3197239296
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1534603811434303e-05,
      "loss": 1.3461,
      "theoretical_loss": 3.3331118603549825,
      "tokens_seen": 3197304832
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.151454363089268e-05,
      "loss": 1.2436,
      "theoretical_loss": 3.333106719523222,
      "tokens_seen": 3197370368
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1494483450351055e-05,
      "loss": 1.1733,
      "theoretical_loss": 3.3331015788263336,
      "tokens_seen": 3197435904
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1474423269809424e-05,
      "loss": 1.2231,
      "theoretical_loss": 3.3330964382643122,
      "tokens_seen": 3197501440
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.14543630892678e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.3330912978371514,
      "tokens_seen": 3197566976
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.143430290872618e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.333086157544844,
      "tokens_seen": 3197632512
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.141424272818456e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.333081017387385,
      "tokens_seen": 3197698048
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1394182547642934e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.3330758773647666,
      "tokens_seen": 3197763584
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.13741223671013e-05,
      "loss": 1.2164,
      "theoretical_loss": 3.3330707374769837,
      "tokens_seen": 3197829120
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.135406218655968e-05,
      "loss": 1.3039,
      "theoretical_loss": 3.3330655977240293,
      "tokens_seen": 3197894656
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1334002006018055e-05,
      "loss": 1.2146,
      "theoretical_loss": 3.333060458105898,
      "tokens_seen": 3197960192
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.131394182547643e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.333055318622582,
      "tokens_seen": 3198025728
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1293881644934806e-05,
      "loss": 1.3548,
      "theoretical_loss": 3.3330501792740765,
      "tokens_seen": 3198091264
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1273821464393175e-05,
      "loss": 1.232,
      "theoretical_loss": 3.3330450400603744,
      "tokens_seen": 3198156800
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.125376128385155e-05,
      "loss": 1.2076,
      "theoretical_loss": 3.3330399009814693,
      "tokens_seen": 3198222336
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.123370110330993e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.3330347620373555,
      "tokens_seen": 3198287872
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.12136409227683e-05,
      "loss": 1.2514,
      "theoretical_loss": 3.3330296232280263,
      "tokens_seen": 3198353408
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1193580742226685e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.3330244845534756,
      "tokens_seen": 3198418944
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1173520561685054e-05,
      "loss": 1.2536,
      "theoretical_loss": 3.333019346013697,
      "tokens_seen": 3198484480
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.115346038114343e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.3330142076086844,
      "tokens_seen": 3198550016
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.1133400200601806e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.3330090693384307,
      "tokens_seen": 3198615552
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.111334002006018e-05,
      "loss": 1.1543,
      "theoretical_loss": 3.3330039312029305,
      "tokens_seen": 3198681088
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.109327983951856e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.3329987932021776,
      "tokens_seen": 3198746624
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.107321965897693e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.332993655336165,
      "tokens_seen": 3198812160
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.105315947843531e-05,
      "loss": 1.2507,
      "theoretical_loss": 3.3329885176048863,
      "tokens_seen": 3198877696
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.103309929789368e-05,
      "loss": 1.1671,
      "theoretical_loss": 3.332983380008336,
      "tokens_seen": 3198943232
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.101303911735206e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.3329782425465075,
      "tokens_seen": 3199008768
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.099297893681043e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3329731052193945,
      "tokens_seen": 3199074304
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0972918756268805e-05,
      "loss": 1.2433,
      "theoretical_loss": 3.3329679680269906,
      "tokens_seen": 3199139840
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.095285857572718e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.332962830969289,
      "tokens_seen": 3199205376
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.093279839518556e-05,
      "loss": 1.1851,
      "theoretical_loss": 3.332957694046285,
      "tokens_seen": 3199270912
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.091273821464393e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.3329525572579706,
      "tokens_seen": 3199336448
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.089267803410231e-05,
      "loss": 1.3198,
      "theoretical_loss": 3.33294742060434,
      "tokens_seen": 3199401984
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0872617853560684e-05,
      "loss": 1.2211,
      "theoretical_loss": 3.3329422840853877,
      "tokens_seen": 3199467520
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.085255767301906e-05,
      "loss": 1.1736,
      "theoretical_loss": 3.3329371477011067,
      "tokens_seen": 3199533056
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.083249749247743e-05,
      "loss": 1.2601,
      "theoretical_loss": 3.3329320114514904,
      "tokens_seen": 3199598592
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.081243731193581e-05,
      "loss": 1.1871,
      "theoretical_loss": 3.332926875336533,
      "tokens_seen": 3199664128
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.079237713139418e-05,
      "loss": 1.2228,
      "theoretical_loss": 3.3329217393562285,
      "tokens_seen": 3199729664
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0772316950852556e-05,
      "loss": 1.3689,
      "theoretical_loss": 3.33291660351057,
      "tokens_seen": 3199795200
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.075225677031093e-05,
      "loss": 1.2252,
      "theoretical_loss": 3.3329114677995517,
      "tokens_seen": 3199860736
    },
    {
      "epoch": 0.94,
      "objective/train/docs_used": 1162972,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.224173665046692,
      "objective/train/theoretical_loss": 3.3329063322231667,
      "objective/train/tokens_used": 1570385376,
      "theoretical_loss": 3.3329063322231667,
      "tokens_seen": 3199926272
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.073219658976931e-05,
      "loss": 1.263,
      "theoretical_loss": 3.3329063322231667,
      "tokens_seen": 3199926272
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0712136409227684e-05,
      "loss": 1.1724,
      "theoretical_loss": 3.3329011967814095,
      "tokens_seen": 3199991808
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.069207622868606e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.332896061474273,
      "tokens_seen": 3200057344
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0672016048144435e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.3328909263017517,
      "tokens_seen": 3200122880
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0651955867602804e-05,
      "loss": 1.3151,
      "theoretical_loss": 3.3328857912638385,
      "tokens_seen": 3200188416
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.063189568706119e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.332880656360528,
      "tokens_seen": 3200253952
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.061183550651956e-05,
      "loss": 1.298,
      "theoretical_loss": 3.3328755215918133,
      "tokens_seen": 3200319488
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.059177532597793e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.3328703869576883,
      "tokens_seen": 3200385024
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.057171514543631e-05,
      "loss": 1.238,
      "theoretical_loss": 3.3328652524581464,
      "tokens_seen": 3200450560
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.055165496489468e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.332860118093182,
      "tokens_seen": 3200516096
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.053159478435306e-05,
      "loss": 1.2758,
      "theoretical_loss": 3.3328549838627883,
      "tokens_seen": 3200581632
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0511534603811435e-05,
      "loss": 1.1879,
      "theoretical_loss": 3.332849849766959,
      "tokens_seen": 3200647168
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.049147442326981e-05,
      "loss": 1.2132,
      "theoretical_loss": 3.332844715805688,
      "tokens_seen": 3200712704
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0471414242728183e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.3328395819789693,
      "tokens_seen": 3200778240
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0451354062186562e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.332834448286796,
      "tokens_seen": 3200843776
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0431293881644935e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.332829314729162,
      "tokens_seen": 3200909312
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.041123370110331e-05,
      "loss": 1.1728,
      "theoretical_loss": 3.3328241813060617,
      "tokens_seen": 3200974848
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0391173520561686e-05,
      "loss": 1.2975,
      "theoretical_loss": 3.3328190480174875,
      "tokens_seen": 3201040384
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.037111334002006e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.3328139148634346,
      "tokens_seen": 3201105920
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0351053159478438e-05,
      "loss": 1.2212,
      "theoretical_loss": 3.332808781843896,
      "tokens_seen": 3201171456
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.033099297893681e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.3328036489588646,
      "tokens_seen": 3201236992
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0310932798395186e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.3327985162083356,
      "tokens_seen": 3201302528
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0290872617853562e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.332793383592302,
      "tokens_seen": 3201368064
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0270812437311934e-05,
      "loss": 1.2185,
      "theoretical_loss": 3.3327882511107574,
      "tokens_seen": 3201433600
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0250752256770314e-05,
      "loss": 1.3755,
      "theoretical_loss": 3.332783118763696,
      "tokens_seen": 3201499136
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0230692076228686e-05,
      "loss": 1.2786,
      "theoretical_loss": 3.332777986551111,
      "tokens_seen": 3201564672
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0210631895687062e-05,
      "loss": 1.211,
      "theoretical_loss": 3.332772854472996,
      "tokens_seen": 3201630208
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0190571715145438e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.3327677225293457,
      "tokens_seen": 3201695744
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0170511534603813e-05,
      "loss": 1.2581,
      "theoretical_loss": 3.332762590720153,
      "tokens_seen": 3201761280
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.015045135406219e-05,
      "loss": 1.1953,
      "theoretical_loss": 3.332757459045412,
      "tokens_seen": 3201826816
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.013039117352056e-05,
      "loss": 1.2296,
      "theoretical_loss": 3.332752327505116,
      "tokens_seen": 3201892352
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0110330992978937e-05,
      "loss": 1.3123,
      "theoretical_loss": 3.3327471960992594,
      "tokens_seen": 3201957888
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.009027081243731e-05,
      "loss": 1.2276,
      "theoretical_loss": 3.332742064827835,
      "tokens_seen": 3202023424
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.007021063189569e-05,
      "loss": 1.3513,
      "theoretical_loss": 3.3327369336908372,
      "tokens_seen": 3202088960
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0050150451354065e-05,
      "loss": 1.2609,
      "theoretical_loss": 3.3327318026882597,
      "tokens_seen": 3202154496
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0030090270812437e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.332726671820096,
      "tokens_seen": 3202220032
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.0010030090270813e-05,
      "loss": 1.1589,
      "theoretical_loss": 3.33272154108634,
      "tokens_seen": 3202285568
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9989969909729185e-05,
      "loss": 1.1905,
      "theoretical_loss": 3.3327164104869853,
      "tokens_seen": 3202351104
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9969909729187564e-05,
      "loss": 1.1578,
      "theoretical_loss": 3.3327112800220258,
      "tokens_seen": 3202416640
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.994984954864594e-05,
      "loss": 1.3288,
      "theoretical_loss": 3.3327061496914547,
      "tokens_seen": 3202482176
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9929789368104313e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.3327010194952664,
      "tokens_seen": 3202547712
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.990972918756269e-05,
      "loss": 1.2673,
      "theoretical_loss": 3.332695889433454,
      "tokens_seen": 3202613248
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.988966900702106e-05,
      "loss": 1.1942,
      "theoretical_loss": 3.3326907595060122,
      "tokens_seen": 3202678784
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.986960882647944e-05,
      "loss": 1.0726,
      "theoretical_loss": 3.332685629712934,
      "tokens_seen": 3202744320
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9849548645937816e-05,
      "loss": 1.1859,
      "theoretical_loss": 3.3326805000542135,
      "tokens_seen": 3202809856
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9829488465396188e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.3326753705298438,
      "tokens_seen": 3202875392
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9809428284854564e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.332670241139819,
      "tokens_seen": 3202940928
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.978936810431294e-05,
      "loss": 1.2537,
      "theoretical_loss": 3.332665111884133,
      "tokens_seen": 3203006464
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9769307923771316e-05,
      "loss": 1.2067,
      "theoretical_loss": 3.332659982762779,
      "tokens_seen": 3203072000
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.974924774322969e-05,
      "loss": 1.2316,
      "theoretical_loss": 3.3326548537757517,
      "tokens_seen": 3203137536
    },
    {
      "epoch": 0.94,
      "objective/train/docs_used": 1164087,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2418830394744873,
      "objective/train/theoretical_loss": 3.3326497249230442,
      "objective/train/tokens_used": 1573662176,
      "theoretical_loss": 3.3326497249230442,
      "tokens_seen": 3203203072
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9729187562688064e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.3326497249230442,
      "tokens_seen": 3203203072
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.970912738214644e-05,
      "loss": 1.237,
      "theoretical_loss": 3.33264459620465,
      "tokens_seen": 3203268608
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9689067201604815e-05,
      "loss": 1.1192,
      "theoretical_loss": 3.332639467620563,
      "tokens_seen": 3203334144
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.966900702106319e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.3326343391707773,
      "tokens_seen": 3203399680
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9648946840521564e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.3326292108552864,
      "tokens_seen": 3203465216
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.962888665997994e-05,
      "loss": 1.2324,
      "theoretical_loss": 3.332624082674084,
      "tokens_seen": 3203530752
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9608826479438315e-05,
      "loss": 1.1505,
      "theoretical_loss": 3.332618954627164,
      "tokens_seen": 3203596288
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.958876629889669e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.3326138267145202,
      "tokens_seen": 3203661824
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9568706118355067e-05,
      "loss": 1.26,
      "theoretical_loss": 3.332608698936146,
      "tokens_seen": 3203727360
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.954864593781344e-05,
      "loss": 1.1703,
      "theoretical_loss": 3.3326035712920348,
      "tokens_seen": 3203792896
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9528585757271815e-05,
      "loss": 1.2581,
      "theoretical_loss": 3.332598443782181,
      "tokens_seen": 3203858432
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9508525576730194e-05,
      "loss": 1.2226,
      "theoretical_loss": 3.3325933164065784,
      "tokens_seen": 3203923968
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9488465396188567e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.3325881891652207,
      "tokens_seen": 3203989504
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9468405215646942e-05,
      "loss": 1.2379,
      "theoretical_loss": 3.332583062058101,
      "tokens_seen": 3204055040
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9448345035105315e-05,
      "loss": 1.3421,
      "theoretical_loss": 3.332577935085214,
      "tokens_seen": 3204120576
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.942828485456369e-05,
      "loss": 1.278,
      "theoretical_loss": 3.332572808246552,
      "tokens_seen": 3204186112
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.940822467402207e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.332567681542111,
      "tokens_seen": 3204251648
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9388164493480442e-05,
      "loss": 1.2609,
      "theoretical_loss": 3.332562554971882,
      "tokens_seen": 3204317184
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9368104312938818e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.332557428535861,
      "tokens_seen": 3204382720
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.934804413239719e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.3325523022340406,
      "tokens_seen": 3204448256
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9327983951855566e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.3325471760664147,
      "tokens_seen": 3204513792
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9307923771313942e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.3325420500329774,
      "tokens_seen": 3204579328
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9287863590772318e-05,
      "loss": 1.1489,
      "theoretical_loss": 3.332536924133722,
      "tokens_seen": 3204644864
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9267803410230693e-05,
      "loss": 1.3658,
      "theoretical_loss": 3.3325317983686427,
      "tokens_seen": 3204710400
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9247743229689066e-05,
      "loss": 1.1399,
      "theoretical_loss": 3.332526672737733,
      "tokens_seen": 3204775936
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9227683049147445e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.3325215472409866,
      "tokens_seen": 3204841472
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9207622868605817e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.332516421878397,
      "tokens_seen": 3204907008
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9187562688064193e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.332511296649959,
      "tokens_seen": 3204972544
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.916750250752257e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.3325061715556648,
      "tokens_seen": 3205038080
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.914744232698094e-05,
      "loss": 1.2182,
      "theoretical_loss": 3.3325010465955094,
      "tokens_seen": 3205103616
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.912738214643932e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.3324959217694854,
      "tokens_seen": 3205169152
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9107321965897693e-05,
      "loss": 1.2308,
      "theoretical_loss": 3.332490797077588,
      "tokens_seen": 3205234688
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.908726178535607e-05,
      "loss": 1.197,
      "theoretical_loss": 3.3324856725198098,
      "tokens_seen": 3205300224
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9067201604814445e-05,
      "loss": 1.264,
      "theoretical_loss": 3.332480548096145,
      "tokens_seen": 3205365760
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9047141424272817e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.3324754238065872,
      "tokens_seen": 3205431296
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.9027081243731196e-05,
      "loss": 1.3183,
      "theoretical_loss": 3.3324702996511304,
      "tokens_seen": 3205496832
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.900702106318957e-05,
      "loss": 1.1337,
      "theoretical_loss": 3.3324651756297676,
      "tokens_seen": 3205562368
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8986960882647944e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.332460051742494,
      "tokens_seen": 3205627904
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8966900702106317e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.3324549279893017,
      "tokens_seen": 3205693440
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8946840521564696e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.3324498043701856,
      "tokens_seen": 3205758976
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8926780341023072e-05,
      "loss": 1.2266,
      "theoretical_loss": 3.3324446808851387,
      "tokens_seen": 3205824512
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8906720160481444e-05,
      "loss": 1.206,
      "theoretical_loss": 3.3324395575341557,
      "tokens_seen": 3205890048
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.888665997993982e-05,
      "loss": 1.1295,
      "theoretical_loss": 3.3324344343172294,
      "tokens_seen": 3205955584
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8866599799398192e-05,
      "loss": 1.2087,
      "theoretical_loss": 3.3324293112343537,
      "tokens_seen": 3206021120
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.884653961885657e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.332424188285523,
      "tokens_seen": 3206086656
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8826479438314947e-05,
      "loss": 1.1982,
      "theoretical_loss": 3.3324190654707304,
      "tokens_seen": 3206152192
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.880641925777332e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.33241394278997,
      "tokens_seen": 3206217728
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8786359077231696e-05,
      "loss": 1.203,
      "theoretical_loss": 3.332408820243235,
      "tokens_seen": 3206283264
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8766298896690068e-05,
      "loss": 1.269,
      "theoretical_loss": 3.33240369783052,
      "tokens_seen": 3206348800
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8746238716148447e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.332398575551818,
      "tokens_seen": 3206414336
    },
    {
      "epoch": 0.94,
      "objective/train/docs_used": 1165309,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2051477432250977,
      "objective/train/theoretical_loss": 3.3323934534071236,
      "objective/train/tokens_used": 1576938976,
      "theoretical_loss": 3.3323934534071236,
      "tokens_seen": 3206479872
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8726178535606823e-05,
      "loss": 1.2075,
      "theoretical_loss": 3.3323934534071236,
      "tokens_seen": 3206479872
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8706118355065195e-05,
      "loss": 1.1416,
      "theoretical_loss": 3.3323883313964298,
      "tokens_seen": 3206545408
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.868605817452357e-05,
      "loss": 1.2903,
      "theoretical_loss": 3.3323832095197305,
      "tokens_seen": 3206610944
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8665997993981947e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.3323780877770197,
      "tokens_seen": 3206676480
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8645937813440323e-05,
      "loss": 1.2224,
      "theoretical_loss": 3.332372966168291,
      "tokens_seen": 3206742016
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8625877632898695e-05,
      "loss": 1.2216,
      "theoretical_loss": 3.332367844693538,
      "tokens_seen": 3206807552
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.860581745235707e-05,
      "loss": 1.1797,
      "theoretical_loss": 3.332362723352755,
      "tokens_seen": 3206873088
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8585757271815447e-05,
      "loss": 1.1926,
      "theoretical_loss": 3.332357602145935,
      "tokens_seen": 3206938624
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8565697091273823e-05,
      "loss": 1.1943,
      "theoretical_loss": 3.3323524810730722,
      "tokens_seen": 3207004160
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.85456369107322e-05,
      "loss": 1.2024,
      "theoretical_loss": 3.3323473601341607,
      "tokens_seen": 3207069696
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.852557673019057e-05,
      "loss": 1.2243,
      "theoretical_loss": 3.3323422393291935,
      "tokens_seen": 3207135232
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8505516549648947e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.332337118658165,
      "tokens_seen": 3207200768
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8485456369107322e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.3323319981210684,
      "tokens_seen": 3207266304
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8465396188565698e-05,
      "loss": 1.1754,
      "theoretical_loss": 3.3323268777178976,
      "tokens_seen": 3207331840
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8445336008024074e-05,
      "loss": 1.2122,
      "theoretical_loss": 3.332321757448647,
      "tokens_seen": 3207397376
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8425275827482446e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.3323166373133093,
      "tokens_seen": 3207462912
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8405215646940822e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.3323115173118794,
      "tokens_seen": 3207528448
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.83851554663992e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.3323063974443503,
      "tokens_seen": 3207593984
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8365095285857574e-05,
      "loss": 1.304,
      "theoretical_loss": 3.3323012777107155,
      "tokens_seen": 3207659520
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.834503510531595e-05,
      "loss": 1.1445,
      "theoretical_loss": 3.3322961581109696,
      "tokens_seen": 3207725056
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8324974924774322e-05,
      "loss": 1.2071,
      "theoretical_loss": 3.332291038645106,
      "tokens_seen": 3207790592
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8304914744232698e-05,
      "loss": 1.2105,
      "theoretical_loss": 3.3322859193131187,
      "tokens_seen": 3207856128
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8284854563691077e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.332280800115001,
      "tokens_seen": 3207921664
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.826479438314945e-05,
      "loss": 1.3182,
      "theoretical_loss": 3.3322756810507466,
      "tokens_seen": 3207987200
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8244734202607825e-05,
      "loss": 1.2623,
      "theoretical_loss": 3.3322705621203497,
      "tokens_seen": 3208052736
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8224674022066197e-05,
      "loss": 1.3058,
      "theoretical_loss": 3.332265443323804,
      "tokens_seen": 3208118272
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8204613841524573e-05,
      "loss": 1.2667,
      "theoretical_loss": 3.3322603246611027,
      "tokens_seen": 3208183808
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.818455366098295e-05,
      "loss": 1.1057,
      "theoretical_loss": 3.3322552061322406,
      "tokens_seen": 3208249344
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8164493480441325e-05,
      "loss": 1.2134,
      "theoretical_loss": 3.3322500877372105,
      "tokens_seen": 3208314880
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.81444332998997e-05,
      "loss": 1.1234,
      "theoretical_loss": 3.332244969476007,
      "tokens_seen": 3208380416
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8124373119358073e-05,
      "loss": 1.1379,
      "theoretical_loss": 3.332239851348623,
      "tokens_seen": 3208445952
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8104312938816452e-05,
      "loss": 1.247,
      "theoretical_loss": 3.332234733355053,
      "tokens_seen": 3208511488
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8084252758274825e-05,
      "loss": 1.2907,
      "theoretical_loss": 3.33222961549529,
      "tokens_seen": 3208577024
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.80641925777332e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.332224497769329,
      "tokens_seen": 3208642560
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8044132397191576e-05,
      "loss": 1.2156,
      "theoretical_loss": 3.332219380177163,
      "tokens_seen": 3208708096
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.802407221664995e-05,
      "loss": 1.2252,
      "theoretical_loss": 3.332214262718785,
      "tokens_seen": 3208773632
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.8004012036108328e-05,
      "loss": 1.1185,
      "theoretical_loss": 3.3322091453941898,
      "tokens_seen": 3208839168
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.79839518555667e-05,
      "loss": 1.2376,
      "theoretical_loss": 3.332204028203371,
      "tokens_seen": 3208904704
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.7963891675025076e-05,
      "loss": 1.3043,
      "theoretical_loss": 3.332198911146323,
      "tokens_seen": 3208970240
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.7943831494483452e-05,
      "loss": 1.224,
      "theoretical_loss": 3.3321937942230377,
      "tokens_seen": 3209035776
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.7923771313941824e-05,
      "loss": 1.234,
      "theoretical_loss": 3.3321886774335105,
      "tokens_seen": 3209101312
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.7903711133400203e-05,
      "loss": 1.2187,
      "theoretical_loss": 3.332183560777735,
      "tokens_seen": 3209166848
    },
    {
      "epoch": 0.94,
      "learning_rate": 2.7883650952858576e-05,
      "loss": 1.1539,
      "theoretical_loss": 3.3321784442557045,
      "tokens_seen": 3209232384
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.786359077231695e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.332173327867413,
      "tokens_seen": 3209297920
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7843530591775324e-05,
      "loss": 1.1879,
      "theoretical_loss": 3.3321682116128537,
      "tokens_seen": 3209363456
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7823470411233703e-05,
      "loss": 1.1664,
      "theoretical_loss": 3.3321630954920214,
      "tokens_seen": 3209428992
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.780341023069208e-05,
      "loss": 1.3595,
      "theoretical_loss": 3.3321579795049097,
      "tokens_seen": 3209494528
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.778335005015045e-05,
      "loss": 1.1858,
      "theoretical_loss": 3.3321528636515114,
      "tokens_seen": 3209560064
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7763289869608827e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.3321477479318213,
      "tokens_seen": 3209625600
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.77432296890672e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.3321426323458323,
      "tokens_seen": 3209691136
    },
    {
      "epoch": 0.95,
      "objective/train/docs_used": 1166593,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2713661193847656,
      "objective/train/theoretical_loss": 3.332137516893539,
      "objective/train/tokens_used": 1580215776,
      "theoretical_loss": 3.332137516893539,
      "tokens_seen": 3209756672
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.772316950852558e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.332137516893539,
      "tokens_seen": 3209756672
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7703109327983955e-05,
      "loss": 1.2206,
      "theoretical_loss": 3.3321324015749347,
      "tokens_seen": 3209822208
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7683049147442327e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.3321272863900138,
      "tokens_seen": 3209887744
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7662988966900703e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.332122171338769,
      "tokens_seen": 3209953280
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7642928786359075e-05,
      "loss": 1.1127,
      "theoretical_loss": 3.332117056421195,
      "tokens_seen": 3210018816
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7622868605817454e-05,
      "loss": 1.1807,
      "theoretical_loss": 3.3321119416372853,
      "tokens_seen": 3210084352
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.760280842527583e-05,
      "loss": 1.2449,
      "theoretical_loss": 3.3321068269870335,
      "tokens_seen": 3210149888
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7582748244734203e-05,
      "loss": 1.1555,
      "theoretical_loss": 3.3321017124704335,
      "tokens_seen": 3210215424
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.756268806419258e-05,
      "loss": 1.1704,
      "theoretical_loss": 3.3320965980874795,
      "tokens_seen": 3210280960
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7542627883650954e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.3320914838381643,
      "tokens_seen": 3210346496
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.752256770310933e-05,
      "loss": 1.2056,
      "theoretical_loss": 3.3320863697224823,
      "tokens_seen": 3210412032
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7502507522567702e-05,
      "loss": 1.2291,
      "theoretical_loss": 3.3320812557404276,
      "tokens_seen": 3210477568
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7482447342026078e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.3320761418919935,
      "tokens_seen": 3210543104
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7462387161484454e-05,
      "loss": 1.1847,
      "theoretical_loss": 3.332071028177174,
      "tokens_seen": 3210608640
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.744232698094283e-05,
      "loss": 1.115,
      "theoretical_loss": 3.3320659145959626,
      "tokens_seen": 3210674176
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7422266800401205e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.3320608011483532,
      "tokens_seen": 3210739712
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7402206619859578e-05,
      "loss": 1.1393,
      "theoretical_loss": 3.3320556878343397,
      "tokens_seen": 3210805248
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7382146439317954e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.3320505746539157,
      "tokens_seen": 3210870784
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.736208625877633e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.332045461607075,
      "tokens_seen": 3210936320
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7342026078234705e-05,
      "loss": 1.2277,
      "theoretical_loss": 3.3320403486938117,
      "tokens_seen": 3211001856
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.732196589769308e-05,
      "loss": 1.1811,
      "theoretical_loss": 3.3320352359141197,
      "tokens_seen": 3211067392
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7301905717151453e-05,
      "loss": 1.1838,
      "theoretical_loss": 3.332030123267992,
      "tokens_seen": 3211132928
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.728184553660983e-05,
      "loss": 1.2732,
      "theoretical_loss": 3.332025010755423,
      "tokens_seen": 3211198464
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7261785356068205e-05,
      "loss": 1.2297,
      "theoretical_loss": 3.3320198983764064,
      "tokens_seen": 3211264000
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.724172517552658e-05,
      "loss": 1.1829,
      "theoretical_loss": 3.332014786130936,
      "tokens_seen": 3211329536
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7221664994984957e-05,
      "loss": 1.3253,
      "theoretical_loss": 3.3320096740190053,
      "tokens_seen": 3211395072
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.720160481444333e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.3320045620406082,
      "tokens_seen": 3211460608
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7181544633901705e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.3319994501957386,
      "tokens_seen": 3211526144
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.716148445336008e-05,
      "loss": 1.3422,
      "theoretical_loss": 3.3319943384843906,
      "tokens_seen": 3211591680
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7141424272818456e-05,
      "loss": 1.196,
      "theoretical_loss": 3.331989226906557,
      "tokens_seen": 3211657216
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7121364092276832e-05,
      "loss": 1.2292,
      "theoretical_loss": 3.331984115462233,
      "tokens_seen": 3211722752
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7101303911735205e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.3319790041514112,
      "tokens_seen": 3211788288
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.708124373119358e-05,
      "loss": 1.1779,
      "theoretical_loss": 3.331973892974086,
      "tokens_seen": 3211853824
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7061183550651956e-05,
      "loss": 1.264,
      "theoretical_loss": 3.331968781930251,
      "tokens_seen": 3211919360
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7041123370110332e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.3319636710198997,
      "tokens_seen": 3211984896
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.7021063189568708e-05,
      "loss": 1.2371,
      "theoretical_loss": 3.3319585602430264,
      "tokens_seen": 3212050432
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.700100300902708e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.3319534495996246,
      "tokens_seen": 3212115968
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6980942828485456e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.3319483390896885,
      "tokens_seen": 3212181504
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6960882647943832e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.3319432287132114,
      "tokens_seen": 3212247040
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6940822467402208e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.331938118470187,
      "tokens_seen": 3212312576
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6920762286860583e-05,
      "loss": 1.219,
      "theoretical_loss": 3.3319330083606093,
      "tokens_seen": 3212378112
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6900702106318956e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.3319278983844725,
      "tokens_seen": 3212443648
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6880641925777335e-05,
      "loss": 1.1639,
      "theoretical_loss": 3.33192278854177,
      "tokens_seen": 3212509184
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6860581745235707e-05,
      "loss": 1.2732,
      "theoretical_loss": 3.3319176788324953,
      "tokens_seen": 3212574720
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6840521564694083e-05,
      "loss": 1.3217,
      "theoretical_loss": 3.3319125692566427,
      "tokens_seen": 3212640256
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6820461384152456e-05,
      "loss": 1.2477,
      "theoretical_loss": 3.331907459814206,
      "tokens_seen": 3212705792
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.680040120361083e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.3319023505051786,
      "tokens_seen": 3212771328
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.678034102306921e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.3318972413295547,
      "tokens_seen": 3212836864
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6760280842527583e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.3318921322873276,
      "tokens_seen": 3212902400
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.674022066198596e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.3318870233784916,
      "tokens_seen": 3212967936
    },
    {
      "epoch": 0.95,
      "objective/train/docs_used": 1167631,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.370660662651062,
      "objective/train/theoretical_loss": 3.33188191460304,
      "objective/train/tokens_used": 1583492576,
      "theoretical_loss": 3.33188191460304,
      "tokens_seen": 3213033472
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.672016048144433e-05,
      "loss": 1.337,
      "theoretical_loss": 3.33188191460304,
      "tokens_seen": 3213033472
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6700100300902707e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.3318768059609676,
      "tokens_seen": 3213099008
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6680040120361086e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.3318716974522666,
      "tokens_seen": 3213164544
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.665997993981946e-05,
      "loss": 1.3907,
      "theoretical_loss": 3.331866589076932,
      "tokens_seen": 3213230080
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6639919759277834e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.3318614808349576,
      "tokens_seen": 3213295616
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6619859578736207e-05,
      "loss": 1.2363,
      "theoretical_loss": 3.3318563727263366,
      "tokens_seen": 3213361152
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6599799398194586e-05,
      "loss": 1.2536,
      "theoretical_loss": 3.331851264751063,
      "tokens_seen": 3213426688
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.657973921765296e-05,
      "loss": 1.262,
      "theoretical_loss": 3.331846156909131,
      "tokens_seen": 3213492224
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6559679037111334e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.331841049200534,
      "tokens_seen": 3213557760
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.653961885656971e-05,
      "loss": 1.1555,
      "theoretical_loss": 3.3318359416252656,
      "tokens_seen": 3213623296
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6519558676028082e-05,
      "loss": 1.229,
      "theoretical_loss": 3.3318308341833203,
      "tokens_seen": 3213688832
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.649949849548646e-05,
      "loss": 1.2031,
      "theoretical_loss": 3.3318257268746914,
      "tokens_seen": 3213754368
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6479438314944837e-05,
      "loss": 1.2461,
      "theoretical_loss": 3.3318206196993723,
      "tokens_seen": 3213819904
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.645937813440321e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.3318155126573576,
      "tokens_seen": 3213885440
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6439317953861585e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.331810405748641,
      "tokens_seen": 3213950976
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6419257773319958e-05,
      "loss": 1.2075,
      "theoretical_loss": 3.3318052989732156,
      "tokens_seen": 3214016512
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6399197592778337e-05,
      "loss": 1.2814,
      "theoretical_loss": 3.3318001923310763,
      "tokens_seen": 3214082048
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.637913741223671e-05,
      "loss": 1.2443,
      "theoretical_loss": 3.3317950858222156,
      "tokens_seen": 3214147584
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6359077231695085e-05,
      "loss": 1.2158,
      "theoretical_loss": 3.331789979446629,
      "tokens_seen": 3214213120
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.633901705115346e-05,
      "loss": 1.3457,
      "theoretical_loss": 3.3317848732043083,
      "tokens_seen": 3214278656
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6318956870611837e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.3317797670952487,
      "tokens_seen": 3214344192
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6298896690070213e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.331774661119444,
      "tokens_seen": 3214409728
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6278836509528585e-05,
      "loss": 1.1321,
      "theoretical_loss": 3.331769555276887,
      "tokens_seen": 3214475264
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.625877632898696e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.3317644495675722,
      "tokens_seen": 3214540800
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6238716148445337e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.331759343991494,
      "tokens_seen": 3214606336
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6218655967903712e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.331754238548645,
      "tokens_seen": 3214671872
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6198595787362088e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.3317491332390192,
      "tokens_seen": 3214737408
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.617853560682046e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.3317440280626114,
      "tokens_seen": 3214802944
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6158475426278836e-05,
      "loss": 1.2481,
      "theoretical_loss": 3.3317389230194143,
      "tokens_seen": 3214868480
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6138415245737212e-05,
      "loss": 1.1967,
      "theoretical_loss": 3.3317338181094227,
      "tokens_seen": 3214934016
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6118355065195588e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.3317287133326294,
      "tokens_seen": 3214999552
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6098294884653964e-05,
      "loss": 1.0884,
      "theoretical_loss": 3.331723608689029,
      "tokens_seen": 3215065088
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6078234704112336e-05,
      "loss": 1.1659,
      "theoretical_loss": 3.331718504178615,
      "tokens_seen": 3215130624
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6058174523570712e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.331713399801381,
      "tokens_seen": 3215196160
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6038114343029088e-05,
      "loss": 1.2156,
      "theoretical_loss": 3.331708295557321,
      "tokens_seen": 3215261696
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.6018054162487464e-05,
      "loss": 1.1582,
      "theoretical_loss": 3.331703191446429,
      "tokens_seen": 3215327232
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.599799398194584e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.3316980874686988,
      "tokens_seen": 3215392768
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5977933801404212e-05,
      "loss": 1.227,
      "theoretical_loss": 3.331692983624124,
      "tokens_seen": 3215458304
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5957873620862588e-05,
      "loss": 1.1656,
      "theoretical_loss": 3.331687879912698,
      "tokens_seen": 3215523840
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5937813440320963e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.3316827763344157,
      "tokens_seen": 3215589376
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.591775325977934e-05,
      "loss": 1.2341,
      "theoretical_loss": 3.3316776728892696,
      "tokens_seen": 3215654912
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5897693079237715e-05,
      "loss": 1.1757,
      "theoretical_loss": 3.331672569577255,
      "tokens_seen": 3215720448
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5877632898696087e-05,
      "loss": 1.295,
      "theoretical_loss": 3.3316674663983648,
      "tokens_seen": 3215785984
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5857572718154463e-05,
      "loss": 1.2113,
      "theoretical_loss": 3.331662363352592,
      "tokens_seen": 3215851520
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.583751253761284e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.3316572604399326,
      "tokens_seen": 3215917056
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5817452357071215e-05,
      "loss": 1.1693,
      "theoretical_loss": 3.3316521576603786,
      "tokens_seen": 3215982592
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.579739217652959e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.3316470550139243,
      "tokens_seen": 3216048128
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5777331995987963e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.3316419525005636,
      "tokens_seen": 3216113664
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5757271815446342e-05,
      "loss": 1.2329,
      "theoretical_loss": 3.33163685012029,
      "tokens_seen": 3216179200
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5737211634904715e-05,
      "loss": 1.1829,
      "theoretical_loss": 3.3316317478730983,
      "tokens_seen": 3216244736
    },
    {
      "epoch": 0.95,
      "objective/train/docs_used": 1168814,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.0978078842163086,
      "objective/train/theoretical_loss": 3.3316266457589814,
      "objective/train/tokens_used": 1586769376,
      "theoretical_loss": 3.3316266457589814,
      "tokens_seen": 3216310272
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.571715145436309e-05,
      "loss": 1.108,
      "theoretical_loss": 3.3316266457589814,
      "tokens_seen": 3216310272
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5697091273821463e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.3316215437779335,
      "tokens_seen": 3216375808
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.567703109327984e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.331616441929948,
      "tokens_seen": 3216441344
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5656970912738218e-05,
      "loss": 1.2291,
      "theoretical_loss": 3.3316113402150194,
      "tokens_seen": 3216506880
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.563691073219659e-05,
      "loss": 1.1842,
      "theoretical_loss": 3.3316062386331406,
      "tokens_seen": 3216572416
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5616850551654966e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.3316011371843066,
      "tokens_seen": 3216637952
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5596790371113338e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.33159603586851,
      "tokens_seen": 3216703488
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5576730190571714e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.3315909346857455,
      "tokens_seen": 3216769024
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5556670010030093e-05,
      "loss": 1.315,
      "theoretical_loss": 3.3315858336360065,
      "tokens_seen": 3216834560
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5536609829488466e-05,
      "loss": 1.235,
      "theoretical_loss": 3.331580732719287,
      "tokens_seen": 3216900096
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.551654964894684e-05,
      "loss": 1.2065,
      "theoretical_loss": 3.3315756319355807,
      "tokens_seen": 3216965632
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5496489468405214e-05,
      "loss": 1.268,
      "theoretical_loss": 3.3315705312848816,
      "tokens_seen": 3217031168
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5476429287863593e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.331565430767183,
      "tokens_seen": 3217096704
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.545636910732197e-05,
      "loss": 1.3172,
      "theoretical_loss": 3.3315603303824792,
      "tokens_seen": 3217162240
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.543630892678034e-05,
      "loss": 1.2748,
      "theoretical_loss": 3.331555230130764,
      "tokens_seen": 3217227776
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5416248746238717e-05,
      "loss": 1.1847,
      "theoretical_loss": 3.3315501300120314,
      "tokens_seen": 3217293312
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.539618856569709e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.331545030026275,
      "tokens_seen": 3217358848
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.537612838515547e-05,
      "loss": 1.1727,
      "theoretical_loss": 3.3315399301734887,
      "tokens_seen": 3217424384
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.535606820461384e-05,
      "loss": 1.1749,
      "theoretical_loss": 3.3315348304536654,
      "tokens_seen": 3217489920
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5336008024072217e-05,
      "loss": 1.233,
      "theoretical_loss": 3.3315297308668006,
      "tokens_seen": 3217555456
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5315947843530593e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.3315246314128872,
      "tokens_seen": 3217620992
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5295887662988965e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.3315195320919186,
      "tokens_seen": 3217686528
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5275827482447344e-05,
      "loss": 1.241,
      "theoretical_loss": 3.3315144329038895,
      "tokens_seen": 3217752064
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5255767301905717e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.3315093338487936,
      "tokens_seen": 3217817600
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5235707121364092e-05,
      "loss": 1.1386,
      "theoretical_loss": 3.331504234926624,
      "tokens_seen": 3217883136
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5215646940822468e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.331499136137375,
      "tokens_seen": 3217948672
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5195586760280844e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.331494037481041,
      "tokens_seen": 3218014208
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.517552657973922e-05,
      "loss": 1.1898,
      "theoretical_loss": 3.331488938957615,
      "tokens_seen": 3218079744
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5155466399197592e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.331483840567091,
      "tokens_seen": 3218145280
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5135406218655968e-05,
      "loss": 1.1658,
      "theoretical_loss": 3.331478742309463,
      "tokens_seen": 3218210816
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5115346038114344e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.331473644184725,
      "tokens_seen": 3218276352
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.509528585757272e-05,
      "loss": 1.254,
      "theoretical_loss": 3.33146854619287,
      "tokens_seen": 3218341888
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5075225677031095e-05,
      "loss": 1.2163,
      "theoretical_loss": 3.331463448333893,
      "tokens_seen": 3218407424
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5055165496489468e-05,
      "loss": 1.193,
      "theoretical_loss": 3.331458350607787,
      "tokens_seen": 3218472960
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5035105315947844e-05,
      "loss": 1.3643,
      "theoretical_loss": 3.331453253014546,
      "tokens_seen": 3218538496
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.5015045135406216e-05,
      "loss": 1.2357,
      "theoretical_loss": 3.331448155554164,
      "tokens_seen": 3218604032
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4994984954864595e-05,
      "loss": 1.1489,
      "theoretical_loss": 3.3314430582266352,
      "tokens_seen": 3218669568
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.497492477432297e-05,
      "loss": 1.1916,
      "theoretical_loss": 3.3314379610319524,
      "tokens_seen": 3218735104
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4954864593781343e-05,
      "loss": 1.1412,
      "theoretical_loss": 3.33143286397011,
      "tokens_seen": 3218800640
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.493480441323972e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.3314277670411023,
      "tokens_seen": 3218866176
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4914744232698095e-05,
      "loss": 1.3153,
      "theoretical_loss": 3.331422670244922,
      "tokens_seen": 3218931712
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.489468405215647e-05,
      "loss": 1.251,
      "theoretical_loss": 3.3314175735815645,
      "tokens_seen": 3218997248
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4874623871614847e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.331412477051022,
      "tokens_seen": 3219062784
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.485456369107322e-05,
      "loss": 1.2163,
      "theoretical_loss": 3.3314073806532893,
      "tokens_seen": 3219128320
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4834503510531595e-05,
      "loss": 1.1822,
      "theoretical_loss": 3.33140228438836,
      "tokens_seen": 3219193856
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.481444332998997e-05,
      "loss": 1.4299,
      "theoretical_loss": 3.3313971882562283,
      "tokens_seen": 3219259392
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4794383149448346e-05,
      "loss": 1.1915,
      "theoretical_loss": 3.3313920922568876,
      "tokens_seen": 3219324928
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4774322968906722e-05,
      "loss": 1.2312,
      "theoretical_loss": 3.3313869963903318,
      "tokens_seen": 3219390464
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4754262788365095e-05,
      "loss": 1.1726,
      "theoretical_loss": 3.3313819006565546,
      "tokens_seen": 3219456000
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.473420260782347e-05,
      "loss": 1.1859,
      "theoretical_loss": 3.33137680505555,
      "tokens_seen": 3219521536
    },
    {
      "epoch": 0.95,
      "objective/train/docs_used": 1169560,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2335251569747925,
      "objective/train/theoretical_loss": 3.331371709587312,
      "objective/train/tokens_used": 1590046176,
      "theoretical_loss": 3.331371709587312,
      "tokens_seen": 3219587072
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4714142427281846e-05,
      "loss": 1.2528,
      "theoretical_loss": 3.331371709587312,
      "tokens_seen": 3219587072
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4694082246740222e-05,
      "loss": 1.1925,
      "theoretical_loss": 3.331366614251834,
      "tokens_seen": 3219652608
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4674022066198598e-05,
      "loss": 1.2368,
      "theoretical_loss": 3.33136151904911,
      "tokens_seen": 3219718144
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.465396188565697e-05,
      "loss": 1.2101,
      "theoretical_loss": 3.3313564239791345,
      "tokens_seen": 3219783680
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.463390170511535e-05,
      "loss": 1.2169,
      "theoretical_loss": 3.3313513290419006,
      "tokens_seen": 3219849216
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.461384152457372e-05,
      "loss": 1.2076,
      "theoretical_loss": 3.3313462342374023,
      "tokens_seen": 3219914752
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4593781344032097e-05,
      "loss": 1.2444,
      "theoretical_loss": 3.3313411395656334,
      "tokens_seen": 3219980288
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.457372116349047e-05,
      "loss": 1.2414,
      "theoretical_loss": 3.331336045026588,
      "tokens_seen": 3220045824
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4553660982948846e-05,
      "loss": 1.2076,
      "theoretical_loss": 3.3313309506202597,
      "tokens_seen": 3220111360
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4533600802407225e-05,
      "loss": 1.1913,
      "theoretical_loss": 3.331325856346642,
      "tokens_seen": 3220176896
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4513540621865597e-05,
      "loss": 1.2707,
      "theoretical_loss": 3.33132076220573,
      "tokens_seen": 3220242432
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4493480441323973e-05,
      "loss": 1.2394,
      "theoretical_loss": 3.331315668197516,
      "tokens_seen": 3220307968
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4473420260782345e-05,
      "loss": 1.1672,
      "theoretical_loss": 3.3313105743219946,
      "tokens_seen": 3220373504
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.445336008024072e-05,
      "loss": 1.1821,
      "theoretical_loss": 3.33130548057916,
      "tokens_seen": 3220439040
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.44332998996991e-05,
      "loss": 1.1903,
      "theoretical_loss": 3.331300386969005,
      "tokens_seen": 3220504576
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4413239719157473e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.3312952934915248,
      "tokens_seen": 3220570112
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.439317953861585e-05,
      "loss": 1.353,
      "theoretical_loss": 3.3312902001467117,
      "tokens_seen": 3220635648
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.437311935807422e-05,
      "loss": 1.2409,
      "theoretical_loss": 3.331285106934561,
      "tokens_seen": 3220701184
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4353059177532597e-05,
      "loss": 1.326,
      "theoretical_loss": 3.3312800138550656,
      "tokens_seen": 3220766720
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4332998996990976e-05,
      "loss": 1.3305,
      "theoretical_loss": 3.3312749209082195,
      "tokens_seen": 3220832256
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.431293881644935e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.331269828094017,
      "tokens_seen": 3220897792
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4292878635907724e-05,
      "loss": 1.234,
      "theoretical_loss": 3.331264735412452,
      "tokens_seen": 3220963328
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4272818455366097e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.3312596428635173,
      "tokens_seen": 3221028864
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4252758274824476e-05,
      "loss": 1.1977,
      "theoretical_loss": 3.3312545504472073,
      "tokens_seen": 3221094400
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4232698094282848e-05,
      "loss": 1.1636,
      "theoretical_loss": 3.3312494581635166,
      "tokens_seen": 3221159936
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4212637913741224e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.331244366012438,
      "tokens_seen": 3221225472
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.41925777331996e-05,
      "loss": 1.3567,
      "theoretical_loss": 3.331239273993966,
      "tokens_seen": 3221291008
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4172517552657972e-05,
      "loss": 1.1214,
      "theoretical_loss": 3.331234182108094,
      "tokens_seen": 3221356544
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.415245737211635e-05,
      "loss": 1.2211,
      "theoretical_loss": 3.331229090354816,
      "tokens_seen": 3221422080
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4132397191574724e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.331223998734126,
      "tokens_seen": 3221487616
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.41123370110331e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.331218907246018,
      "tokens_seen": 3221553152
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4092276830491475e-05,
      "loss": 1.1383,
      "theoretical_loss": 3.331213815890486,
      "tokens_seen": 3221618688
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4072216649949848e-05,
      "loss": 1.3732,
      "theoretical_loss": 3.3312087246675226,
      "tokens_seen": 3221684224
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4052156469408227e-05,
      "loss": 1.2028,
      "theoretical_loss": 3.331203633577123,
      "tokens_seen": 3221749760
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.40320962888666e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.3311985426192807,
      "tokens_seen": 3221815296
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.4012036108324975e-05,
      "loss": 1.1739,
      "theoretical_loss": 3.331193451793989,
      "tokens_seen": 3221880832
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.399197592778335e-05,
      "loss": 1.3457,
      "theoretical_loss": 3.3311883611012423,
      "tokens_seen": 3221946368
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3971915747241727e-05,
      "loss": 1.2242,
      "theoretical_loss": 3.3311832705410347,
      "tokens_seen": 3222011904
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3951855566700103e-05,
      "loss": 1.2255,
      "theoretical_loss": 3.3311781801133593,
      "tokens_seen": 3222077440
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3931795386158475e-05,
      "loss": 1.3101,
      "theoretical_loss": 3.3311730898182104,
      "tokens_seen": 3222142976
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.391173520561685e-05,
      "loss": 1.289,
      "theoretical_loss": 3.331167999655582,
      "tokens_seen": 3222208512
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3891675025075223e-05,
      "loss": 1.2284,
      "theoretical_loss": 3.3311629096254673,
      "tokens_seen": 3222274048
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3871614844533602e-05,
      "loss": 1.2047,
      "theoretical_loss": 3.3311578197278613,
      "tokens_seen": 3222339584
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3851554663991978e-05,
      "loss": 1.16,
      "theoretical_loss": 3.331152729962757,
      "tokens_seen": 3222405120
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.383149448345035e-05,
      "loss": 1.1904,
      "theoretical_loss": 3.331147640330148,
      "tokens_seen": 3222470656
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3811434302908726e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.3311425508300285,
      "tokens_seen": 3222536192
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.37913741223671e-05,
      "loss": 1.1807,
      "theoretical_loss": 3.3311374614623928,
      "tokens_seen": 3222601728
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3771313941825478e-05,
      "loss": 1.2372,
      "theoretical_loss": 3.3311323722272346,
      "tokens_seen": 3222667264
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3751253761283854e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.3311272831245473,
      "tokens_seen": 3222732800
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3731193580742226e-05,
      "loss": 1.2351,
      "theoretical_loss": 3.3311221941543248,
      "tokens_seen": 3222798336
    },
    {
      "epoch": 0.95,
      "objective/train/docs_used": 1170779,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3256371021270752,
      "objective/train/theoretical_loss": 3.3311171053165616,
      "objective/train/tokens_used": 1593322976,
      "theoretical_loss": 3.3311171053165616,
      "tokens_seen": 3222863872
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3711133400200602e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.3311171053165616,
      "tokens_seen": 3222863872
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3691073219658978e-05,
      "loss": 1.2555,
      "theoretical_loss": 3.3311120166112507,
      "tokens_seen": 3222929408
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3671013039117353e-05,
      "loss": 1.2298,
      "theoretical_loss": 3.3311069280383867,
      "tokens_seen": 3222994944
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.365095285857573e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.331101839597963,
      "tokens_seen": 3223060480
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.36308926780341e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.3310967512899734,
      "tokens_seen": 3223126016
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3610832497492477e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.331091663114412,
      "tokens_seen": 3223191552
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3590772316950853e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.331086575071273,
      "tokens_seen": 3223257088
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.357071213640923e-05,
      "loss": 1.3187,
      "theoretical_loss": 3.3310814871605494,
      "tokens_seen": 3223322624
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3550651955867605e-05,
      "loss": 1.2346,
      "theoretical_loss": 3.331076399382236,
      "tokens_seen": 3223388160
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3530591775325977e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.331071311736326,
      "tokens_seen": 3223453696
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3510531594784353e-05,
      "loss": 1.2377,
      "theoretical_loss": 3.3310662242228135,
      "tokens_seen": 3223519232
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.349047141424273e-05,
      "loss": 1.1226,
      "theoretical_loss": 3.3310611368416922,
      "tokens_seen": 3223584768
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3470411233701105e-05,
      "loss": 1.2154,
      "theoretical_loss": 3.331056049592956,
      "tokens_seen": 3223650304
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3450351053159477e-05,
      "loss": 1.341,
      "theoretical_loss": 3.331050962476599,
      "tokens_seen": 3223715840
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3430290872617853e-05,
      "loss": 1.2315,
      "theoretical_loss": 3.3310458754926153,
      "tokens_seen": 3223781376
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3410230692076232e-05,
      "loss": 1.2106,
      "theoretical_loss": 3.331040788640998,
      "tokens_seen": 3223846912
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3390170511534604e-05,
      "loss": 1.0443,
      "theoretical_loss": 3.3310357019217416,
      "tokens_seen": 3223912448
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.337011033099298e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.3310306153348397,
      "tokens_seen": 3223977984
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3350050150451353e-05,
      "loss": 1.1925,
      "theoretical_loss": 3.331025528880286,
      "tokens_seen": 3224043520
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.332998996990973e-05,
      "loss": 1.257,
      "theoretical_loss": 3.3310204425580747,
      "tokens_seen": 3224109056
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3309929789368108e-05,
      "loss": 1.2346,
      "theoretical_loss": 3.3310153563681992,
      "tokens_seen": 3224174592
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.328986960882648e-05,
      "loss": 1.2147,
      "theoretical_loss": 3.331010270310654,
      "tokens_seen": 3224240128
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3269809428284856e-05,
      "loss": 1.2232,
      "theoretical_loss": 3.3310051843854325,
      "tokens_seen": 3224305664
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3249749247743228e-05,
      "loss": 1.201,
      "theoretical_loss": 3.331000098592529,
      "tokens_seen": 3224371200
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3229689067201604e-05,
      "loss": 1.3445,
      "theoretical_loss": 3.330995012931937,
      "tokens_seen": 3224436736
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3209628886659983e-05,
      "loss": 1.3363,
      "theoretical_loss": 3.33098992740365,
      "tokens_seen": 3224502272
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3189568706118356e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.3309848420076626,
      "tokens_seen": 3224567808
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.316950852557673e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.3309797567439685,
      "tokens_seen": 3224633344
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3149448345035104e-05,
      "loss": 1.3179,
      "theoretical_loss": 3.3309746716125614,
      "tokens_seen": 3224698880
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3129388164493483e-05,
      "loss": 1.2071,
      "theoretical_loss": 3.3309695866134357,
      "tokens_seen": 3224764416
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3109327983951855e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.330964501746584,
      "tokens_seen": 3224829952
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.308926780341023e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.330959417012002,
      "tokens_seen": 3224895488
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.3069207622868607e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.3309543324096813,
      "tokens_seen": 3224961024
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.304914744232698e-05,
      "loss": 1.125,
      "theoretical_loss": 3.3309492479396177,
      "tokens_seen": 3225026560
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.302908726178536e-05,
      "loss": 1.3379,
      "theoretical_loss": 3.3309441636018047,
      "tokens_seen": 3225092096
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.300902708124373e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.3309390793962352,
      "tokens_seen": 3225157632
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.2988966900702107e-05,
      "loss": 1.1996,
      "theoretical_loss": 3.330933995322904,
      "tokens_seen": 3225223168
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.2968906720160482e-05,
      "loss": 1.2715,
      "theoretical_loss": 3.330928911381805,
      "tokens_seen": 3225288704
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.2948846539618855e-05,
      "loss": 1.3601,
      "theoretical_loss": 3.3309238275729314,
      "tokens_seen": 3225354240
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.2928786359077234e-05,
      "loss": 1.28,
      "theoretical_loss": 3.3309187438962775,
      "tokens_seen": 3225419776
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.2908726178535606e-05,
      "loss": 1.174,
      "theoretical_loss": 3.3309136603518374,
      "tokens_seen": 3225485312
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.2888665997993982e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.3309085769396045,
      "tokens_seen": 3225550848
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.2868605817452358e-05,
      "loss": 1.2035,
      "theoretical_loss": 3.330903493659573,
      "tokens_seen": 3225616384
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.2848545636910734e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.3308984105117365,
      "tokens_seen": 3225681920
    },
    {
      "epoch": 0.95,
      "learning_rate": 2.282848545636911e-05,
      "loss": 1.2118,
      "theoretical_loss": 3.3308933274960895,
      "tokens_seen": 3225747456
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2808425275827482e-05,
      "loss": 1.3057,
      "theoretical_loss": 3.3308882446126247,
      "tokens_seen": 3225812992
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2788365095285858e-05,
      "loss": 1.2642,
      "theoretical_loss": 3.3308831618613373,
      "tokens_seen": 3225878528
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.276830491474423e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.33087807924222,
      "tokens_seen": 3225944064
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.274824473420261e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.330872996755268,
      "tokens_seen": 3226009600
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2728184553660985e-05,
      "loss": 1.193,
      "theoretical_loss": 3.330867914400474,
      "tokens_seen": 3226075136
    },
    {
      "epoch": 0.96,
      "objective/train/docs_used": 1172070,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.288637638092041,
      "objective/train/theoretical_loss": 3.330862832177832,
      "objective/train/tokens_used": 1596599776,
      "theoretical_loss": 3.330862832177832,
      "tokens_seen": 3226140672
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2708124373119358e-05,
      "loss": 1.2529,
      "theoretical_loss": 3.330862832177832,
      "tokens_seen": 3226140672
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2688064192577733e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.3308577500873366,
      "tokens_seen": 3226206208
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2668004012036106e-05,
      "loss": 1.1646,
      "theoretical_loss": 3.330852668128981,
      "tokens_seen": 3226271744
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2647943831494485e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.3308475863027596,
      "tokens_seen": 3226337280
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.262788365095286e-05,
      "loss": 1.3478,
      "theoretical_loss": 3.330842504608666,
      "tokens_seen": 3226402816
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2607823470411233e-05,
      "loss": 1.2386,
      "theoretical_loss": 3.3308374230466944,
      "tokens_seen": 3226468352
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.258776328986961e-05,
      "loss": 1.2556,
      "theoretical_loss": 3.330832341616838,
      "tokens_seen": 3226533888
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2567703109327985e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.330827260319091,
      "tokens_seen": 3226599424
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.254764292878636e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.3308221791534476,
      "tokens_seen": 3226664960
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2527582748244736e-05,
      "loss": 1.2261,
      "theoretical_loss": 3.3308170981199012,
      "tokens_seen": 3226730496
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.250752256770311e-05,
      "loss": 1.2299,
      "theoretical_loss": 3.330812017218446,
      "tokens_seen": 3226796032
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2487462387161485e-05,
      "loss": 1.1564,
      "theoretical_loss": 3.330806936449076,
      "tokens_seen": 3226861568
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.246740220661986e-05,
      "loss": 1.26,
      "theoretical_loss": 3.3308018558117847,
      "tokens_seen": 3226927104
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2447342026078236e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.3307967753065664,
      "tokens_seen": 3226992640
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.242728184553661e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.3307916949334144,
      "tokens_seen": 3227058176
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2407221664994984e-05,
      "loss": 1.2644,
      "theoretical_loss": 3.330786614692323,
      "tokens_seen": 3227123712
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.238716148445336e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.330781534583286,
      "tokens_seen": 3227189248
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2367101303911736e-05,
      "loss": 1.3089,
      "theoretical_loss": 3.3307764546062977,
      "tokens_seen": 3227254784
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2347041123370112e-05,
      "loss": 1.2569,
      "theoretical_loss": 3.330771374761351,
      "tokens_seen": 3227320320
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2326980942828484e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.3307662950484413,
      "tokens_seen": 3227385856
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.230692076228686e-05,
      "loss": 1.1319,
      "theoretical_loss": 3.330761215467561,
      "tokens_seen": 3227451392
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.228686058174524e-05,
      "loss": 1.1851,
      "theoretical_loss": 3.3307561360187043,
      "tokens_seen": 3227516928
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.226680040120361e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.3307510567018657,
      "tokens_seen": 3227582464
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2246740220661987e-05,
      "loss": 1.2468,
      "theoretical_loss": 3.3307459775170387,
      "tokens_seen": 3227648000
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.222668004012036e-05,
      "loss": 1.2416,
      "theoretical_loss": 3.3307408984642173,
      "tokens_seen": 3227713536
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2206619859578736e-05,
      "loss": 1.241,
      "theoretical_loss": 3.330735819543395,
      "tokens_seen": 3227779072
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2186559679037115e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.330730740754566,
      "tokens_seen": 3227844608
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2166499498495487e-05,
      "loss": 1.2791,
      "theoretical_loss": 3.330725662097725,
      "tokens_seen": 3227910144
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2146439317953863e-05,
      "loss": 1.3062,
      "theoretical_loss": 3.330720583572864,
      "tokens_seen": 3227975680
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2126379137412235e-05,
      "loss": 1.386,
      "theoretical_loss": 3.3307155051799784,
      "tokens_seen": 3228041216
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.210631895687061e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.3307104269190617,
      "tokens_seen": 3228106752
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.208625877632899e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.330705348790108,
      "tokens_seen": 3228172288
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.2066198595787363e-05,
      "loss": 1.213,
      "theoretical_loss": 3.330700270793111,
      "tokens_seen": 3228237824
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.204613841524574e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.330695192928064,
      "tokens_seen": 3228303360
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.202607823470411e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.3306901151949617,
      "tokens_seen": 3228368896
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.200601805416249e-05,
      "loss": 1.1635,
      "theoretical_loss": 3.3306850375937977,
      "tokens_seen": 3228434432
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1985957873620862e-05,
      "loss": 1.1789,
      "theoretical_loss": 3.330679960124566,
      "tokens_seen": 3228499968
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1965897693079238e-05,
      "loss": 1.244,
      "theoretical_loss": 3.3306748827872603,
      "tokens_seen": 3228565504
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1945837512537614e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.3306698055818744,
      "tokens_seen": 3228631040
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1925777331995986e-05,
      "loss": 1.3707,
      "theoretical_loss": 3.330664728508403,
      "tokens_seen": 3228696576
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1905717151454366e-05,
      "loss": 1.2204,
      "theoretical_loss": 3.330659651566839,
      "tokens_seen": 3228762112
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1885656970912738e-05,
      "loss": 1.3536,
      "theoretical_loss": 3.330654574757177,
      "tokens_seen": 3228827648
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1865596790371114e-05,
      "loss": 1.3062,
      "theoretical_loss": 3.33064949807941,
      "tokens_seen": 3228893184
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.184553660982949e-05,
      "loss": 1.2264,
      "theoretical_loss": 3.330644421533533,
      "tokens_seen": 3228958720
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1825476429287862e-05,
      "loss": 1.1594,
      "theoretical_loss": 3.3306393451195393,
      "tokens_seen": 3229024256
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.180541624874624e-05,
      "loss": 1.1646,
      "theoretical_loss": 3.330634268837423,
      "tokens_seen": 3229089792
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1785356068204614e-05,
      "loss": 1.2088,
      "theoretical_loss": 3.3306291926871774,
      "tokens_seen": 3229155328
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.176529588766299e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.330624116668797,
      "tokens_seen": 3229220864
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1745235707121365e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.3306190407822758,
      "tokens_seen": 3229286400
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1725175526579738e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.3306139650276076,
      "tokens_seen": 3229351936
    },
    {
      "epoch": 0.96,
      "objective/train/docs_used": 1173331,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1166080236434937,
      "objective/train/theoretical_loss": 3.330608889404786,
      "objective/train/tokens_used": 1599876576,
      "theoretical_loss": 3.330608889404786,
      "tokens_seen": 3229417472
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1705115346038117e-05,
      "loss": 1.1923,
      "theoretical_loss": 3.330608889404786,
      "tokens_seen": 3229417472
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.168505516549649e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.330603813913805,
      "tokens_seen": 3229483008
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1664994984954865e-05,
      "loss": 1.2347,
      "theoretical_loss": 3.3305987385546585,
      "tokens_seen": 3229548544
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1644934804413237e-05,
      "loss": 1.2215,
      "theoretical_loss": 3.330593663327341,
      "tokens_seen": 3229614080
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1624874623871617e-05,
      "loss": 1.3385,
      "theoretical_loss": 3.330588588231845,
      "tokens_seen": 3229679616
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1604814443329992e-05,
      "loss": 1.143,
      "theoretical_loss": 3.330583513268166,
      "tokens_seen": 3229745152
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1584754262788365e-05,
      "loss": 1.2193,
      "theoretical_loss": 3.330578438436297,
      "tokens_seen": 3229810688
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.156469408224674e-05,
      "loss": 1.1068,
      "theoretical_loss": 3.3305733637362316,
      "tokens_seen": 3229876224
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1544633901705113e-05,
      "loss": 1.1759,
      "theoretical_loss": 3.3305682891679647,
      "tokens_seen": 3229941760
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1524573721163492e-05,
      "loss": 1.2228,
      "theoretical_loss": 3.3305632147314896,
      "tokens_seen": 3230007296
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1504513540621868e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.3305581404268003,
      "tokens_seen": 3230072832
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.148445336008024e-05,
      "loss": 1.271,
      "theoretical_loss": 3.330553066253891,
      "tokens_seen": 3230138368
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1464393179538616e-05,
      "loss": 1.1847,
      "theoretical_loss": 3.3305479922127548,
      "tokens_seen": 3230203904
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.144433299899699e-05,
      "loss": 1.3259,
      "theoretical_loss": 3.330542918303386,
      "tokens_seen": 3230269440
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1424272818455368e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.3305378445257787,
      "tokens_seen": 3230334976
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1404212637913744e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.3305327708799273,
      "tokens_seen": 3230400512
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1384152457372116e-05,
      "loss": 1.1842,
      "theoretical_loss": 3.3305276973658247,
      "tokens_seen": 3230466048
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1364092276830492e-05,
      "loss": 1.2463,
      "theoretical_loss": 3.330522623983465,
      "tokens_seen": 3230531584
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1344032096288868e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.3305175507328424,
      "tokens_seen": 3230597120
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1323971915747243e-05,
      "loss": 1.2282,
      "theoretical_loss": 3.330512477613951,
      "tokens_seen": 3230662656
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1303911735205616e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.3305074046267844,
      "tokens_seen": 3230728192
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.128385155466399e-05,
      "loss": 1.2149,
      "theoretical_loss": 3.3305023317713363,
      "tokens_seen": 3230793728
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1263791374122367e-05,
      "loss": 1.1811,
      "theoretical_loss": 3.3304972590476014,
      "tokens_seen": 3230859264
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1243731193580743e-05,
      "loss": 1.3527,
      "theoretical_loss": 3.3304921864555723,
      "tokens_seen": 3230924800
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.122367101303912e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.330487113995244,
      "tokens_seen": 3230990336
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.120361083249749e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.3304820416666105,
      "tokens_seen": 3231055872
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1183550651955867e-05,
      "loss": 1.213,
      "theoretical_loss": 3.330476969469665,
      "tokens_seen": 3231121408
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1163490471414243e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.3304718974044016,
      "tokens_seen": 3231186944
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.114343029087262e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.3304668254708143,
      "tokens_seen": 3231252480
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1123370110330994e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.3304617536688967,
      "tokens_seen": 3231318016
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1103309929789367e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.3304566819986436,
      "tokens_seen": 3231383552
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1083249749247743e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.3304516104600483,
      "tokens_seen": 3231449088
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1063189568706122e-05,
      "loss": 1.1985,
      "theoretical_loss": 3.3304465390531046,
      "tokens_seen": 3231514624
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1043129388164494e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.3304414677778067,
      "tokens_seen": 3231580160
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.102306920762287e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.3304363966341484,
      "tokens_seen": 3231645696
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.1003009027081242e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.3304313256221234,
      "tokens_seen": 3231711232
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0982948846539618e-05,
      "loss": 1.2141,
      "theoretical_loss": 3.330426254741726,
      "tokens_seen": 3231776768
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0962888665997994e-05,
      "loss": 1.3376,
      "theoretical_loss": 3.33042118399295,
      "tokens_seen": 3231842304
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.094282848545637e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.3304161133757892,
      "tokens_seen": 3231907840
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0922768304914746e-05,
      "loss": 1.3423,
      "theoretical_loss": 3.3304110428902374,
      "tokens_seen": 3231973376
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0902708124373118e-05,
      "loss": 1.1437,
      "theoretical_loss": 3.3304059725362887,
      "tokens_seen": 3232038912
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0882647943831494e-05,
      "loss": 1.1814,
      "theoretical_loss": 3.330400902313937,
      "tokens_seen": 3232104448
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.086258776328987e-05,
      "loss": 1.2261,
      "theoretical_loss": 3.3303958322231764,
      "tokens_seen": 3232169984
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0842527582748245e-05,
      "loss": 1.1142,
      "theoretical_loss": 3.3303907622640003,
      "tokens_seen": 3232235520
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.082246740220662e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.330385692436403,
      "tokens_seen": 3232301056
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0802407221664994e-05,
      "loss": 1.3445,
      "theoretical_loss": 3.3303806227403783,
      "tokens_seen": 3232366592
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0782347041123373e-05,
      "loss": 1.2306,
      "theoretical_loss": 3.3303755531759207,
      "tokens_seen": 3232432128
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0762286860581745e-05,
      "loss": 1.3416,
      "theoretical_loss": 3.3303704837430232,
      "tokens_seen": 3232497664
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.074222668004012e-05,
      "loss": 1.2135,
      "theoretical_loss": 3.33036541444168,
      "tokens_seen": 3232563200
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0722166499498497e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.3303603452718855,
      "tokens_seen": 3232628736
    },
    {
      "epoch": 0.96,
      "objective/train/docs_used": 1174536,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3326020240783691,
      "objective/train/theoretical_loss": 3.3303552762336333,
      "objective/train/tokens_used": 1603153376,
      "theoretical_loss": 3.3303552762336333,
      "tokens_seen": 3232694272
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.070210631895687e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.3303552762336333,
      "tokens_seen": 3232694272
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.068204613841525e-05,
      "loss": 1.2334,
      "theoretical_loss": 3.3303502073269167,
      "tokens_seen": 3232759808
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.066198595787362e-05,
      "loss": 1.2357,
      "theoretical_loss": 3.3303451385517304,
      "tokens_seen": 3232825344
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0641925777331997e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.3303400699080683,
      "tokens_seen": 3232890880
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.062186559679037e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.330335001395924,
      "tokens_seen": 3232956416
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0601805416248745e-05,
      "loss": 1.1879,
      "theoretical_loss": 3.3303299330152916,
      "tokens_seen": 3233021952
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0581745235707124e-05,
      "loss": 1.2302,
      "theoretical_loss": 3.330324864766165,
      "tokens_seen": 3233087488
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0561685055165496e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.330319796648538,
      "tokens_seen": 3233153024
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0541624874623872e-05,
      "loss": 1.2316,
      "theoretical_loss": 3.330314728662405,
      "tokens_seen": 3233218560
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0521564694082245e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.3303096608077594,
      "tokens_seen": 3233284096
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0501504513540624e-05,
      "loss": 1.201,
      "theoretical_loss": 3.3303045930845947,
      "tokens_seen": 3233349632
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0481444332999e-05,
      "loss": 1.3043,
      "theoretical_loss": 3.330299525492906,
      "tokens_seen": 3233415168
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0461384152457372e-05,
      "loss": 1.1994,
      "theoretical_loss": 3.3302944580326868,
      "tokens_seen": 3233480704
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0441323971915748e-05,
      "loss": 1.2269,
      "theoretical_loss": 3.33028939070393,
      "tokens_seen": 3233546240
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.042126379137412e-05,
      "loss": 1.243,
      "theoretical_loss": 3.330284323506631,
      "tokens_seen": 3233611776
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.04012036108325e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.3302792564407833,
      "tokens_seen": 3233677312
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0381143430290875e-05,
      "loss": 1.0714,
      "theoretical_loss": 3.3302741895063797,
      "tokens_seen": 3233742848
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0361083249749248e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.330269122703416,
      "tokens_seen": 3233808384
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0341023069207623e-05,
      "loss": 1.2078,
      "theoretical_loss": 3.3302640560318846,
      "tokens_seen": 3233873920
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0320962888665996e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.3302589894917802,
      "tokens_seen": 3233939456
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0300902708124375e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.3302539230830965,
      "tokens_seen": 3234004992
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.028084252758275e-05,
      "loss": 1.2246,
      "theoretical_loss": 3.3302488568058277,
      "tokens_seen": 3234070528
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0260782347041123e-05,
      "loss": 1.2353,
      "theoretical_loss": 3.330243790659967,
      "tokens_seen": 3234136064
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.02407221664995e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.3302387246455094,
      "tokens_seen": 3234201600
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0220661985957875e-05,
      "loss": 1.2309,
      "theoretical_loss": 3.330233658762448,
      "tokens_seen": 3234267136
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.020060180541625e-05,
      "loss": 1.1797,
      "theoretical_loss": 3.3302285930107765,
      "tokens_seen": 3234332672
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0180541624874623e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.3302235273904897,
      "tokens_seen": 3234398208
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0160481444333e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.3302184619015813,
      "tokens_seen": 3234463744
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0140421263791374e-05,
      "loss": 1.2414,
      "theoretical_loss": 3.330213396544045,
      "tokens_seen": 3234529280
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.012036108324975e-05,
      "loss": 1.1367,
      "theoretical_loss": 3.3302083313178743,
      "tokens_seen": 3234594816
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0100300902708126e-05,
      "loss": 1.3469,
      "theoretical_loss": 3.330203266223064,
      "tokens_seen": 3234660352
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.00802407221665e-05,
      "loss": 1.2163,
      "theoretical_loss": 3.330198201259608,
      "tokens_seen": 3234725888
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0060180541624874e-05,
      "loss": 1.3408,
      "theoretical_loss": 3.3301931364274995,
      "tokens_seen": 3234791424
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.004012036108325e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.3301880717267327,
      "tokens_seen": 3234856960
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.0020060180541626e-05,
      "loss": 1.1877,
      "theoretical_loss": 3.3301830071573018,
      "tokens_seen": 3234922496
    },
    {
      "epoch": 0.96,
      "learning_rate": 2e-05,
      "loss": 1.2399,
      "theoretical_loss": 3.3301779427192004,
      "tokens_seen": 3234988032
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9979939819458374e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.330172878412423,
      "tokens_seen": 3235053568
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.995987963891675e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.330167814236963,
      "tokens_seen": 3235119104
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.993981945837513e-05,
      "loss": 1.2429,
      "theoretical_loss": 3.3301627501928146,
      "tokens_seen": 3235184640
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.99197592778335e-05,
      "loss": 1.1607,
      "theoretical_loss": 3.3301576862799713,
      "tokens_seen": 3235250176
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9899699097291877e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.330152622498428,
      "tokens_seen": 3235315712
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.987963891675025e-05,
      "loss": 1.3565,
      "theoretical_loss": 3.3301475588481773,
      "tokens_seen": 3235381248
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9859578736208625e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.3301424953292145,
      "tokens_seen": 3235446784
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9839518555667e-05,
      "loss": 1.1964,
      "theoretical_loss": 3.330137431941532,
      "tokens_seen": 3235512320
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9819458375125377e-05,
      "loss": 1.1642,
      "theoretical_loss": 3.3301323686851254,
      "tokens_seen": 3235577856
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9799398194583753e-05,
      "loss": 1.2044,
      "theoretical_loss": 3.3301273055599876,
      "tokens_seen": 3235643392
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9779338014042125e-05,
      "loss": 1.405,
      "theoretical_loss": 3.3301222425661123,
      "tokens_seen": 3235708928
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.97592778335005e-05,
      "loss": 1.2417,
      "theoretical_loss": 3.3301171797034947,
      "tokens_seen": 3235774464
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9739217652958877e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3301121169721277,
      "tokens_seen": 3235840000
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9719157472417253e-05,
      "loss": 1.2357,
      "theoretical_loss": 3.3301070543720055,
      "tokens_seen": 3235905536
    },
    {
      "epoch": 0.96,
      "objective/train/docs_used": 1175644,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2555317878723145,
      "objective/train/theoretical_loss": 3.330101991903122,
      "objective/train/tokens_used": 1606430176,
      "theoretical_loss": 3.330101991903122,
      "tokens_seen": 3235971072
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.969909729187563e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.330101991903122,
      "tokens_seen": 3235971072
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9679037111334e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.330096929565471,
      "tokens_seen": 3236036608
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.965897693079238e-05,
      "loss": 1.1928,
      "theoretical_loss": 3.3300918673590467,
      "tokens_seen": 3236102144
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9638916750250752e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.3300868052838433,
      "tokens_seen": 3236167680
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9618856569709128e-05,
      "loss": 1.2383,
      "theoretical_loss": 3.330081743339854,
      "tokens_seen": 3236233216
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9598796389167504e-05,
      "loss": 1.281,
      "theoretical_loss": 3.330076681527073,
      "tokens_seen": 3236298752
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9578736208625876e-05,
      "loss": 1.1203,
      "theoretical_loss": 3.3300716198454947,
      "tokens_seen": 3236364288
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9558676028084256e-05,
      "loss": 1.2364,
      "theoretical_loss": 3.3300665582951128,
      "tokens_seen": 3236429824
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9538615847542628e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.3300614968759215,
      "tokens_seen": 3236495360
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9518555667001004e-05,
      "loss": 1.1865,
      "theoretical_loss": 3.3300564355879136,
      "tokens_seen": 3236560896
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9498495486459376e-05,
      "loss": 1.3036,
      "theoretical_loss": 3.3300513744310845,
      "tokens_seen": 3236626432
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9478435305917752e-05,
      "loss": 1.1612,
      "theoretical_loss": 3.3300463134054272,
      "tokens_seen": 3236691968
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.945837512537613e-05,
      "loss": 1.1979,
      "theoretical_loss": 3.330041252510936,
      "tokens_seen": 3236757504
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9438314944834504e-05,
      "loss": 1.2317,
      "theoretical_loss": 3.330036191747605,
      "tokens_seen": 3236823040
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.941825476429288e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.330031131115428,
      "tokens_seen": 3236888576
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9398194583751252e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.330026070614398,
      "tokens_seen": 3236954112
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.937813440320963e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.3300210102445105,
      "tokens_seen": 3237019648
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9358074222668007e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.330015950005759,
      "tokens_seen": 3237085184
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.933801404212638e-05,
      "loss": 1.1795,
      "theoretical_loss": 3.330010889898137,
      "tokens_seen": 3237150720
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9317953861584755e-05,
      "loss": 1.1467,
      "theoretical_loss": 3.330005829921639,
      "tokens_seen": 3237216256
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9297893681043127e-05,
      "loss": 1.2147,
      "theoretical_loss": 3.330000770076258,
      "tokens_seen": 3237281792
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9277833500501506e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.3299957103619886,
      "tokens_seen": 3237347328
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9257773319959882e-05,
      "loss": 1.1848,
      "theoretical_loss": 3.329990650778825,
      "tokens_seen": 3237412864
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9237713139418255e-05,
      "loss": 1.2578,
      "theoretical_loss": 3.329985591326761,
      "tokens_seen": 3237478400
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.921765295887663e-05,
      "loss": 1.3356,
      "theoretical_loss": 3.3299805320057905,
      "tokens_seen": 3237543936
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9197592778335003e-05,
      "loss": 1.3654,
      "theoretical_loss": 3.329975472815907,
      "tokens_seen": 3237609472
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9177532597793382e-05,
      "loss": 1.246,
      "theoretical_loss": 3.329970413757105,
      "tokens_seen": 3237675008
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9157472417251758e-05,
      "loss": 1.3285,
      "theoretical_loss": 3.329965354829378,
      "tokens_seen": 3237740544
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.913741223671013e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.3299602960327204,
      "tokens_seen": 3237806080
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9117352056168506e-05,
      "loss": 1.201,
      "theoretical_loss": 3.329955237367126,
      "tokens_seen": 3237871616
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.909729187562688e-05,
      "loss": 1.1522,
      "theoretical_loss": 3.329950178832589,
      "tokens_seen": 3237937152
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9077231695085258e-05,
      "loss": 1.2109,
      "theoretical_loss": 3.3299451204291026,
      "tokens_seen": 3238002688
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.905717151454363e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.3299400621566613,
      "tokens_seen": 3238068224
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.9037111334002006e-05,
      "loss": 1.3452,
      "theoretical_loss": 3.329935004015259,
      "tokens_seen": 3238133760
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.901705115346038e-05,
      "loss": 1.1127,
      "theoretical_loss": 3.32992994600489,
      "tokens_seen": 3238199296
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8996990972918757e-05,
      "loss": 1.195,
      "theoretical_loss": 3.3299248881255474,
      "tokens_seen": 3238264832
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8976930792377133e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.329919830377226,
      "tokens_seen": 3238330368
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8956870611835506e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.3299147727599188,
      "tokens_seen": 3238395904
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.893681043129388e-05,
      "loss": 1.2256,
      "theoretical_loss": 3.329909715273621,
      "tokens_seen": 3238461440
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8916750250752257e-05,
      "loss": 1.1572,
      "theoretical_loss": 3.3299046579183256,
      "tokens_seen": 3238526976
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8896690070210633e-05,
      "loss": 1.3873,
      "theoretical_loss": 3.329899600694027,
      "tokens_seen": 3238592512
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.887662988966901e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.329894543600719,
      "tokens_seen": 3238658048
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.885656970912738e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.3298894866383955,
      "tokens_seen": 3238723584
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8836509528585757e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.3298844298070507,
      "tokens_seen": 3238789120
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.881644934804413e-05,
      "loss": 1.272,
      "theoretical_loss": 3.329879373106678,
      "tokens_seen": 3238854656
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.879638916750251e-05,
      "loss": 1.1508,
      "theoretical_loss": 3.329874316537272,
      "tokens_seen": 3238920192
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8776328986960884e-05,
      "loss": 1.1829,
      "theoretical_loss": 3.3298692600988264,
      "tokens_seen": 3238985728
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8756268806419257e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.3298642037913355,
      "tokens_seen": 3239051264
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8736208625877633e-05,
      "loss": 1.2371,
      "theoretical_loss": 3.3298591476147923,
      "tokens_seen": 3239116800
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.871614844533601e-05,
      "loss": 1.2766,
      "theoretical_loss": 3.3298540915691914,
      "tokens_seen": 3239182336
    },
    {
      "epoch": 0.96,
      "objective/train/docs_used": 1176615,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3661737442016602,
      "objective/train/theoretical_loss": 3.3298490356545276,
      "objective/train/tokens_used": 1609706976,
      "theoretical_loss": 3.3298490356545276,
      "tokens_seen": 3239247872
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8696088264794384e-05,
      "loss": 1.2656,
      "theoretical_loss": 3.3298490356545276,
      "tokens_seen": 3239247872
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.867602808425276e-05,
      "loss": 1.295,
      "theoretical_loss": 3.3298439798707933,
      "tokens_seen": 3239313408
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8655967903711132e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.329838924217983,
      "tokens_seen": 3239378944
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8635907723169508e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.3298338686960913,
      "tokens_seen": 3239444480
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8615847542627884e-05,
      "loss": 1.3478,
      "theoretical_loss": 3.3298288133051113,
      "tokens_seen": 3239510016
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.859578736208626e-05,
      "loss": 1.1873,
      "theoretical_loss": 3.3298237580450376,
      "tokens_seen": 3239575552
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8575727181544636e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.3298187029158637,
      "tokens_seen": 3239641088
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8555667001003008e-05,
      "loss": 1.2362,
      "theoretical_loss": 3.329813647917584,
      "tokens_seen": 3239706624
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8535606820461384e-05,
      "loss": 1.1752,
      "theoretical_loss": 3.329808593050192,
      "tokens_seen": 3239772160
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.851554663991976e-05,
      "loss": 1.2351,
      "theoretical_loss": 3.329803538313682,
      "tokens_seen": 3239837696
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8495486459378135e-05,
      "loss": 1.2023,
      "theoretical_loss": 3.329798483708048,
      "tokens_seen": 3239903232
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.847542627883651e-05,
      "loss": 1.2257,
      "theoretical_loss": 3.3297934292332836,
      "tokens_seen": 3239968768
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8455366098294884e-05,
      "loss": 1.2543,
      "theoretical_loss": 3.3297883748893833,
      "tokens_seen": 3240034304
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8435305917753263e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.3297833206763405,
      "tokens_seen": 3240099840
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8415245737211635e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.3297782665941495,
      "tokens_seen": 3240165376
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.839518555667001e-05,
      "loss": 1.3339,
      "theoretical_loss": 3.329773212642804,
      "tokens_seen": 3240230912
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8375125376128383e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.329768158822298,
      "tokens_seen": 3240296448
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.835506519558676e-05,
      "loss": 1.3657,
      "theoretical_loss": 3.329763105132626,
      "tokens_seen": 3240361984
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8335005015045138e-05,
      "loss": 1.3197,
      "theoretical_loss": 3.3297580515737817,
      "tokens_seen": 3240427520
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.831494483450351e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.3297529981457585,
      "tokens_seen": 3240493056
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8294884653961886e-05,
      "loss": 1.2419,
      "theoretical_loss": 3.329747944848551,
      "tokens_seen": 3240558592
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.827482447342026e-05,
      "loss": 1.3488,
      "theoretical_loss": 3.3297428916821534,
      "tokens_seen": 3240624128
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8254764292878635e-05,
      "loss": 1.1648,
      "theoretical_loss": 3.3297378386465586,
      "tokens_seen": 3240689664
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8234704112337014e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.329732785741762,
      "tokens_seen": 3240755200
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8214643931795386e-05,
      "loss": 1.203,
      "theoretical_loss": 3.329727732967756,
      "tokens_seen": 3240820736
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8194583751253762e-05,
      "loss": 1.17,
      "theoretical_loss": 3.3297226803245357,
      "tokens_seen": 3240886272
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8174523570712134e-05,
      "loss": 1.1967,
      "theoretical_loss": 3.329717627812095,
      "tokens_seen": 3240951808
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8154463390170514e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.329712575430427,
      "tokens_seen": 3241017344
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.813440320962889e-05,
      "loss": 1.2725,
      "theoretical_loss": 3.3297075231795263,
      "tokens_seen": 3241082880
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8114343029087262e-05,
      "loss": 1.2369,
      "theoretical_loss": 3.329702471059387,
      "tokens_seen": 3241148416
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8094282848545638e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.3296974190700035,
      "tokens_seen": 3241213952
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.807422266800401e-05,
      "loss": 1.263,
      "theoretical_loss": 3.3296923672113685,
      "tokens_seen": 3241279488
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.805416248746239e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.329687315483477,
      "tokens_seen": 3241345024
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.803410230692076e-05,
      "loss": 1.1681,
      "theoretical_loss": 3.3296822638863226,
      "tokens_seen": 3241410560
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.8014042126379137e-05,
      "loss": 1.1589,
      "theoretical_loss": 3.3296772124198992,
      "tokens_seen": 3241476096
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.7993981945837513e-05,
      "loss": 1.1314,
      "theoretical_loss": 3.329672161084201,
      "tokens_seen": 3241541632
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.7973921765295886e-05,
      "loss": 1.1998,
      "theoretical_loss": 3.329667109879222,
      "tokens_seen": 3241607168
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.7953861584754265e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.3296620588049555,
      "tokens_seen": 3241672704
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.7933801404212637e-05,
      "loss": 1.236,
      "theoretical_loss": 3.3296570078613965,
      "tokens_seen": 3241738240
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.7913741223671013e-05,
      "loss": 1.2183,
      "theoretical_loss": 3.3296519570485383,
      "tokens_seen": 3241803776
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.789368104312939e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.329646906366375,
      "tokens_seen": 3241869312
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.7873620862587765e-05,
      "loss": 1.1441,
      "theoretical_loss": 3.329641855814901,
      "tokens_seen": 3241934848
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.785356068204614e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.3296368053941094,
      "tokens_seen": 3242000384
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.7833500501504513e-05,
      "loss": 1.1861,
      "theoretical_loss": 3.329631755103995,
      "tokens_seen": 3242065920
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.781344032096289e-05,
      "loss": 1.2324,
      "theoretical_loss": 3.3296267049445514,
      "tokens_seen": 3242131456
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.7793380140421264e-05,
      "loss": 1.1769,
      "theoretical_loss": 3.3296216549157727,
      "tokens_seen": 3242196992
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.777331995987964e-05,
      "loss": 1.1948,
      "theoretical_loss": 3.3296166050176526,
      "tokens_seen": 3242262528
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7753259779338016e-05,
      "loss": 1.236,
      "theoretical_loss": 3.3296115552501853,
      "tokens_seen": 3242328064
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.773319959879639e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.329606505613365,
      "tokens_seen": 3242393600
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7713139418254764e-05,
      "loss": 1.2025,
      "theoretical_loss": 3.3296014561071856,
      "tokens_seen": 3242459136
    },
    {
      "epoch": 0.97,
      "objective/train/docs_used": 1177712,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.196829080581665,
      "objective/train/theoretical_loss": 3.329596406731641,
      "objective/train/tokens_used": 1612983776,
      "theoretical_loss": 3.329596406731641,
      "tokens_seen": 3242524672
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7693079237713137e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.329596406731641,
      "tokens_seen": 3242524672
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7673019057171516e-05,
      "loss": 1.2134,
      "theoretical_loss": 3.3295913574867244,
      "tokens_seen": 3242590208
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.765295887662989e-05,
      "loss": 1.3408,
      "theoretical_loss": 3.329586308372431,
      "tokens_seen": 3242655744
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7632898696088264e-05,
      "loss": 1.3963,
      "theoretical_loss": 3.329581259388754,
      "tokens_seen": 3242721280
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.761283851554664e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.329576210535688,
      "tokens_seen": 3242786816
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7592778335005016e-05,
      "loss": 1.273,
      "theoretical_loss": 3.3295711618132264,
      "tokens_seen": 3242852352
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.757271815446339e-05,
      "loss": 1.1919,
      "theoretical_loss": 3.3295661132213636,
      "tokens_seen": 3242917888
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7552657973921767e-05,
      "loss": 1.2039,
      "theoretical_loss": 3.3295610647600933,
      "tokens_seen": 3242983424
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.753259779338014e-05,
      "loss": 1.2257,
      "theoretical_loss": 3.32955601642941,
      "tokens_seen": 3243048960
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7512537612838515e-05,
      "loss": 1.3464,
      "theoretical_loss": 3.3295509682293067,
      "tokens_seen": 3243114496
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.749247743229689e-05,
      "loss": 1.3377,
      "theoretical_loss": 3.329545920159778,
      "tokens_seen": 3243180032
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7472417251755267e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.329540872220818,
      "tokens_seen": 3243245568
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7452357071213643e-05,
      "loss": 1.2011,
      "theoretical_loss": 3.329535824412421,
      "tokens_seen": 3243311104
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7432296890672015e-05,
      "loss": 1.2284,
      "theoretical_loss": 3.32953077673458,
      "tokens_seen": 3243376640
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.741223671013039e-05,
      "loss": 1.2219,
      "theoretical_loss": 3.3295257291872895,
      "tokens_seen": 3243442176
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7392176529588767e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.3295206817705436,
      "tokens_seen": 3243507712
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7372116349047142e-05,
      "loss": 1.3705,
      "theoretical_loss": 3.329515634484336,
      "tokens_seen": 3243573248
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7352056168505518e-05,
      "loss": 1.2038,
      "theoretical_loss": 3.329510587328661,
      "tokens_seen": 3243638784
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.733199598796389e-05,
      "loss": 1.1965,
      "theoretical_loss": 3.3295055403035123,
      "tokens_seen": 3243704320
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.731193580742227e-05,
      "loss": 1.286,
      "theoretical_loss": 3.329500493408884,
      "tokens_seen": 3243769856
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7291875626880642e-05,
      "loss": 1.2237,
      "theoretical_loss": 3.3294954466447706,
      "tokens_seen": 3243835392
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7271815446339018e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.329490400011165,
      "tokens_seen": 3243900928
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.725175526579739e-05,
      "loss": 1.2082,
      "theoretical_loss": 3.329485353508062,
      "tokens_seen": 3243966464
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7231695085255766e-05,
      "loss": 1.2235,
      "theoretical_loss": 3.3294803071354555,
      "tokens_seen": 3244032000
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7211634904714145e-05,
      "loss": 1.1693,
      "theoretical_loss": 3.3294752608933393,
      "tokens_seen": 3244097536
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7191574724172518e-05,
      "loss": 1.2504,
      "theoretical_loss": 3.3294702147817077,
      "tokens_seen": 3244163072
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7171514543630894e-05,
      "loss": 1.2152,
      "theoretical_loss": 3.329465168800554,
      "tokens_seen": 3244228608
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7151454363089266e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.3294601229498726,
      "tokens_seen": 3244294144
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7131394182547642e-05,
      "loss": 1.2012,
      "theoretical_loss": 3.329455077229658,
      "tokens_seen": 3244359680
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.711133400200602e-05,
      "loss": 1.2005,
      "theoretical_loss": 3.3294500316399036,
      "tokens_seen": 3244425216
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7091273821464393e-05,
      "loss": 1.2243,
      "theoretical_loss": 3.3294449861806035,
      "tokens_seen": 3244490752
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.707121364092277e-05,
      "loss": 1.2321,
      "theoretical_loss": 3.3294399408517514,
      "tokens_seen": 3244556288
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.705115346038114e-05,
      "loss": 1.2133,
      "theoretical_loss": 3.329434895653342,
      "tokens_seen": 3244621824
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.703109327983952e-05,
      "loss": 1.2198,
      "theoretical_loss": 3.3294298505853686,
      "tokens_seen": 3244687360
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.7011033099297897e-05,
      "loss": 1.3385,
      "theoretical_loss": 3.3294248056478257,
      "tokens_seen": 3244752896
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.699097291875627e-05,
      "loss": 1.189,
      "theoretical_loss": 3.329419760840707,
      "tokens_seen": 3244818432
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6970912738214645e-05,
      "loss": 1.2857,
      "theoretical_loss": 3.3294147161640066,
      "tokens_seen": 3244883968
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6950852557673017e-05,
      "loss": 1.394,
      "theoretical_loss": 3.3294096716177184,
      "tokens_seen": 3244949504
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6930792377131396e-05,
      "loss": 1.1718,
      "theoretical_loss": 3.3294046272018365,
      "tokens_seen": 3245015040
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.691073219658977e-05,
      "loss": 1.3417,
      "theoretical_loss": 3.3293995829163547,
      "tokens_seen": 3245080576
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6890672016048145e-05,
      "loss": 1.2176,
      "theoretical_loss": 3.3293945387612673,
      "tokens_seen": 3245146112
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.687061183550652e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.3293894947365685,
      "tokens_seen": 3245211648
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6850551654964893e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.3293844508422517,
      "tokens_seen": 3245277184
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6830491474423272e-05,
      "loss": 1.3385,
      "theoretical_loss": 3.329379407078311,
      "tokens_seen": 3245342720
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6810431293881644e-05,
      "loss": 1.2323,
      "theoretical_loss": 3.3293743634447406,
      "tokens_seen": 3245408256
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.679037111334002e-05,
      "loss": 1.249,
      "theoretical_loss": 3.3293693199415344,
      "tokens_seen": 3245473792
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6770310932798396e-05,
      "loss": 1.2046,
      "theoretical_loss": 3.3293642765686866,
      "tokens_seen": 3245539328
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6750250752256772e-05,
      "loss": 1.1242,
      "theoretical_loss": 3.329359233326191,
      "tokens_seen": 3245604864
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6730190571715148e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.3293541902140413,
      "tokens_seen": 3245670400
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.671013039117352e-05,
      "loss": 1.2214,
      "theoretical_loss": 3.3293491472322323,
      "tokens_seen": 3245735936
    },
    {
      "epoch": 0.97,
      "objective/train/docs_used": 1178984,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3602168560028076,
      "objective/train/theoretical_loss": 3.3293441043807572,
      "objective/train/tokens_used": 1616260576,
      "theoretical_loss": 3.3293441043807572,
      "tokens_seen": 3245801472
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6690070210631896e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.3293441043807572,
      "tokens_seen": 3245801472
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.667001003009027e-05,
      "loss": 1.1366,
      "theoretical_loss": 3.329339061659611,
      "tokens_seen": 3245867008
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6649949849548647e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.329334019068786,
      "tokens_seen": 3245932544
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6629889669007023e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.329328976608278,
      "tokens_seen": 3245998080
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6609829488465395e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.32932393427808,
      "tokens_seen": 3246063616
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.658976930792377e-05,
      "loss": 1.1739,
      "theoretical_loss": 3.329318892078186,
      "tokens_seen": 3246129152
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6569709127382144e-05,
      "loss": 1.2049,
      "theoretical_loss": 3.3293138500085906,
      "tokens_seen": 3246194688
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6549648946840523e-05,
      "loss": 1.1162,
      "theoretical_loss": 3.329308808069287,
      "tokens_seen": 3246260224
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.65295887662989e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.3293037662602702,
      "tokens_seen": 3246325760
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.650952858575727e-05,
      "loss": 1.2555,
      "theoretical_loss": 3.3292987245815335,
      "tokens_seen": 3246391296
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6489468405215647e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.329293683033071,
      "tokens_seen": 3246456832
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6469408224674023e-05,
      "loss": 1.265,
      "theoretical_loss": 3.329288641614877,
      "tokens_seen": 3246522368
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.64493480441324e-05,
      "loss": 1.1994,
      "theoretical_loss": 3.3292836003269444,
      "tokens_seen": 3246587904
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6429287863590774e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.3292785591692686,
      "tokens_seen": 3246653440
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6409227683049147e-05,
      "loss": 1.1224,
      "theoretical_loss": 3.3292735181418434,
      "tokens_seen": 3246718976
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6389167502507522e-05,
      "loss": 1.3525,
      "theoretical_loss": 3.329268477244662,
      "tokens_seen": 3246784512
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6369107321965898e-05,
      "loss": 1.1878,
      "theoretical_loss": 3.329263436477719,
      "tokens_seen": 3246850048
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6349047141424274e-05,
      "loss": 1.3561,
      "theoretical_loss": 3.3292583958410082,
      "tokens_seen": 3246915584
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.632898696088265e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.329253355334524,
      "tokens_seen": 3246981120
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6308926780341022e-05,
      "loss": 1.1972,
      "theoretical_loss": 3.32924831495826,
      "tokens_seen": 3247046656
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6288866599799398e-05,
      "loss": 1.276,
      "theoretical_loss": 3.3292432747122103,
      "tokens_seen": 3247112192
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6268806419257774e-05,
      "loss": 1.3757,
      "theoretical_loss": 3.329238234596369,
      "tokens_seen": 3247177728
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.624874623871615e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.3292331946107296,
      "tokens_seen": 3247243264
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6228686058174522e-05,
      "loss": 1.1982,
      "theoretical_loss": 3.3292281547552873,
      "tokens_seen": 3247308800
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6208625877632898e-05,
      "loss": 1.3658,
      "theoretical_loss": 3.3292231150300347,
      "tokens_seen": 3247374336
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6188565697091274e-05,
      "loss": 1.2469,
      "theoretical_loss": 3.3292180754349667,
      "tokens_seen": 3247439872
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.616850551654965e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.329213035970077,
      "tokens_seen": 3247505408
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6148445336008025e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.32920799663536,
      "tokens_seen": 3247570944
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6128385155466398e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.329202957430809,
      "tokens_seen": 3247636480
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6108324974924773e-05,
      "loss": 1.3081,
      "theoretical_loss": 3.329197918356418,
      "tokens_seen": 3247702016
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6088264794383153e-05,
      "loss": 1.3466,
      "theoretical_loss": 3.329192879412182,
      "tokens_seen": 3247767552
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6068204613841525e-05,
      "loss": 1.1757,
      "theoretical_loss": 3.3291878405980944,
      "tokens_seen": 3247833088
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.60481444332999e-05,
      "loss": 1.2516,
      "theoretical_loss": 3.3291828019141496,
      "tokens_seen": 3247898624
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.6028084252758273e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.3291777633603408,
      "tokens_seen": 3247964160
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.600802407221665e-05,
      "loss": 1.2573,
      "theoretical_loss": 3.329172724936662,
      "tokens_seen": 3248029696
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5987963891675028e-05,
      "loss": 1.2399,
      "theoretical_loss": 3.3291676866431086,
      "tokens_seen": 3248095232
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.59679037111334e-05,
      "loss": 1.2971,
      "theoretical_loss": 3.329162648479673,
      "tokens_seen": 3248160768
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5947843530591776e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.32915761044635,
      "tokens_seen": 3248226304
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.592778335005015e-05,
      "loss": 1.284,
      "theoretical_loss": 3.3291525725431335,
      "tokens_seen": 3248291840
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5907723169508525e-05,
      "loss": 1.2427,
      "theoretical_loss": 3.329147534770018,
      "tokens_seen": 3248357376
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5887662988966904e-05,
      "loss": 1.3289,
      "theoretical_loss": 3.3291424971269965,
      "tokens_seen": 3248422912
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5867602808425276e-05,
      "loss": 1.248,
      "theoretical_loss": 3.329137459614064,
      "tokens_seen": 3248488448
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5847542627883652e-05,
      "loss": 1.2313,
      "theoretical_loss": 3.329132422231214,
      "tokens_seen": 3248553984
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5827482447342024e-05,
      "loss": 1.2108,
      "theoretical_loss": 3.32912738497844,
      "tokens_seen": 3248619520
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5807422266800404e-05,
      "loss": 1.3355,
      "theoretical_loss": 3.3291223478557375,
      "tokens_seen": 3248685056
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5787362086258776e-05,
      "loss": 1.269,
      "theoretical_loss": 3.329117310863099,
      "tokens_seen": 3248750592
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5767301905717152e-05,
      "loss": 1.1721,
      "theoretical_loss": 3.3291122740005195,
      "tokens_seen": 3248816128
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5747241725175527e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.3291072372679924,
      "tokens_seen": 3248881664
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.57271815446339e-05,
      "loss": 1.2449,
      "theoretical_loss": 3.329102200665512,
      "tokens_seen": 3248947200
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.570712136409228e-05,
      "loss": 1.355,
      "theoretical_loss": 3.3290971641930724,
      "tokens_seen": 3249012736
    },
    {
      "epoch": 0.97,
      "objective/train/docs_used": 1180177,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2066723108291626,
      "objective/train/theoretical_loss": 3.3290921278506675,
      "objective/train/tokens_used": 1619537376,
      "theoretical_loss": 3.3290921278506675,
      "tokens_seen": 3249078272
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.568706118355065e-05,
      "loss": 1.214,
      "theoretical_loss": 3.3290921278506675,
      "tokens_seen": 3249078272
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5667001003009027e-05,
      "loss": 1.2547,
      "theoretical_loss": 3.3290870916382915,
      "tokens_seen": 3249143808
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5646940822467403e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.329082055555938,
      "tokens_seen": 3249209344
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5626880641925775e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.3290770196036013,
      "tokens_seen": 3249274880
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.560682046138415e-05,
      "loss": 1.2139,
      "theoretical_loss": 3.329071983781276,
      "tokens_seen": 3249340416
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5586760280842527e-05,
      "loss": 1.3525,
      "theoretical_loss": 3.329066948088955,
      "tokens_seen": 3249405952
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5566700100300903e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3290619125266327,
      "tokens_seen": 3249471488
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.554663991975928e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.3290568770943034,
      "tokens_seen": 3249537024
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5526579739217654e-05,
      "loss": 1.1762,
      "theoretical_loss": 3.3290518417919612,
      "tokens_seen": 3249602560
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.550651955867603e-05,
      "loss": 1.3426,
      "theoretical_loss": 3.3290468066195995,
      "tokens_seen": 3249668096
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5486459378134403e-05,
      "loss": 1.2347,
      "theoretical_loss": 3.3290417715772134,
      "tokens_seen": 3249733632
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.546639919759278e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.3290367366647957,
      "tokens_seen": 3249799168
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5446339017051154e-05,
      "loss": 1.1283,
      "theoretical_loss": 3.329031701882341,
      "tokens_seen": 3249864704
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.542627883650953e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.3290266672298436,
      "tokens_seen": 3249930240
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5406218655967906e-05,
      "loss": 1.3465,
      "theoretical_loss": 3.3290216327072972,
      "tokens_seen": 3249995776
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5386158475426278e-05,
      "loss": 1.1933,
      "theoretical_loss": 3.3290165983146958,
      "tokens_seen": 3250061312
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5366098294884654e-05,
      "loss": 1.3184,
      "theoretical_loss": 3.329011564052034,
      "tokens_seen": 3250126848
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.534603811434303e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.3290065299193046,
      "tokens_seen": 3250192384
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5325977933801402e-05,
      "loss": 1.2269,
      "theoretical_loss": 3.3290014959165024,
      "tokens_seen": 3250257920
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.530591775325978e-05,
      "loss": 1.2263,
      "theoretical_loss": 3.3289964620436217,
      "tokens_seen": 3250323456
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5285857572718154e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.3289914283006556,
      "tokens_seen": 3250388992
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.526579739217653e-05,
      "loss": 1.1153,
      "theoretical_loss": 3.3289863946875995,
      "tokens_seen": 3250454528
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5245737211634905e-05,
      "loss": 1.1645,
      "theoretical_loss": 3.3289813612044465,
      "tokens_seen": 3250520064
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5225677031093281e-05,
      "loss": 1.252,
      "theoretical_loss": 3.3289763278511906,
      "tokens_seen": 3250585600
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5205616850551655e-05,
      "loss": 1.2274,
      "theoretical_loss": 3.328971294627826,
      "tokens_seen": 3250651136
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.518555667001003e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.3289662615343465,
      "tokens_seen": 3250716672
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5165496489468405e-05,
      "loss": 1.3312,
      "theoretical_loss": 3.328961228570747,
      "tokens_seen": 3250782208
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5145436308926781e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.32895619573702,
      "tokens_seen": 3250847744
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5125376128385157e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.328951163033161,
      "tokens_seen": 3250913280
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5105315947843531e-05,
      "loss": 1.3633,
      "theoretical_loss": 3.3289461304591637,
      "tokens_seen": 3250978816
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5085255767301907e-05,
      "loss": 1.2412,
      "theoretical_loss": 3.3289410980150214,
      "tokens_seen": 3251044352
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.506519558676028e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.328936065700729,
      "tokens_seen": 3251109888
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5045135406218655e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.32893103351628,
      "tokens_seen": 3251175424
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5025075225677032e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.328926001461668,
      "tokens_seen": 3251240960
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.5005015045135406e-05,
      "loss": 1.2235,
      "theoretical_loss": 3.3289209695368887,
      "tokens_seen": 3251306496
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4984954864593782e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.328915937741934,
      "tokens_seen": 3251372032
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4964894684052156e-05,
      "loss": 1.1961,
      "theoretical_loss": 3.3289109060768,
      "tokens_seen": 3251437568
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.494483450351053e-05,
      "loss": 1.213,
      "theoretical_loss": 3.328905874541479,
      "tokens_seen": 3251503104
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4924774322968908e-05,
      "loss": 1.2025,
      "theoretical_loss": 3.3289008431359663,
      "tokens_seen": 3251568640
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4904714142427282e-05,
      "loss": 1.222,
      "theoretical_loss": 3.328895811860255,
      "tokens_seen": 3251634176
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4884653961885658e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.3288907807143397,
      "tokens_seen": 3251699712
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4864593781344032e-05,
      "loss": 1.2089,
      "theoretical_loss": 3.3288857496982143,
      "tokens_seen": 3251765248
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4844533600802408e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3288807188118725,
      "tokens_seen": 3251830784
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4824473420260782e-05,
      "loss": 1.3461,
      "theoretical_loss": 3.3288756880553088,
      "tokens_seen": 3251896320
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4804413239719158e-05,
      "loss": 1.281,
      "theoretical_loss": 3.328870657428517,
      "tokens_seen": 3251961856
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4784353059177533e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.3288656269314916,
      "tokens_seen": 3252027392
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4764292878635907e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.328860596564226,
      "tokens_seen": 3252092928
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4744232698094283e-05,
      "loss": 1.1586,
      "theoretical_loss": 3.3288555663267143,
      "tokens_seen": 3252158464
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4724172517552657e-05,
      "loss": 1.2002,
      "theoretical_loss": 3.328850536218951,
      "tokens_seen": 3252224000
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4704112337011035e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.32884550624093,
      "tokens_seen": 3252289536
    },
    {
      "epoch": 0.97,
      "objective/train/docs_used": 1181412,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1936290264129639,
      "objective/train/theoretical_loss": 3.328840476392645,
      "objective/train/tokens_used": 1622814176,
      "theoretical_loss": 3.328840476392645,
      "tokens_seen": 3252355072
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4684052156469409e-05,
      "loss": 1.2272,
      "theoretical_loss": 3.328840476392645,
      "tokens_seen": 3252355072
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4663991975927783e-05,
      "loss": 1.3373,
      "theoretical_loss": 3.32883544667409,
      "tokens_seen": 3252420608
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4643931795386159e-05,
      "loss": 1.326,
      "theoretical_loss": 3.3288304170852596,
      "tokens_seen": 3252486144
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4623871614844533e-05,
      "loss": 1.25,
      "theoretical_loss": 3.3288253876261473,
      "tokens_seen": 3252551680
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4603811434302909e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.3288203582967473,
      "tokens_seen": 3252617216
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4583751253761285e-05,
      "loss": 1.345,
      "theoretical_loss": 3.328815329097054,
      "tokens_seen": 3252682752
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.456369107321966e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.328810300027061,
      "tokens_seen": 3252748288
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4543630892678034e-05,
      "loss": 1.1658,
      "theoretical_loss": 3.328805271086763,
      "tokens_seen": 3252813824
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4523570712136409e-05,
      "loss": 1.1806,
      "theoretical_loss": 3.3288002422761527,
      "tokens_seen": 3252879360
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4503510531594784e-05,
      "loss": 1.2601,
      "theoretical_loss": 3.3287952135952255,
      "tokens_seen": 3252944896
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4483450351053158e-05,
      "loss": 1.2152,
      "theoretical_loss": 3.3287901850439745,
      "tokens_seen": 3253010432
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4463390170511536e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.3287851566223945,
      "tokens_seen": 3253075968
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.444332998996991e-05,
      "loss": 1.3417,
      "theoretical_loss": 3.3287801283304788,
      "tokens_seen": 3253141504
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4423269809428286e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.3287751001682224,
      "tokens_seen": 3253207040
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.440320962888666e-05,
      "loss": 1.2577,
      "theoretical_loss": 3.3287700721356184,
      "tokens_seen": 3253272576
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4383149448345034e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.3287650442326617,
      "tokens_seen": 3253338112
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4363089267803411e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.3287600164593454,
      "tokens_seen": 3253403648
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4343029087261786e-05,
      "loss": 1.2273,
      "theoretical_loss": 3.328754988815664,
      "tokens_seen": 3253469184
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4322968906720161e-05,
      "loss": 1.212,
      "theoretical_loss": 3.328749961301612,
      "tokens_seen": 3253534720
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4302908726178535e-05,
      "loss": 1.288,
      "theoretical_loss": 3.3287449339171826,
      "tokens_seen": 3253600256
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4282848545636911e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.3287399066623706,
      "tokens_seen": 3253665792
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4262788365095285e-05,
      "loss": 1.2018,
      "theoretical_loss": 3.3287348795371696,
      "tokens_seen": 3253731328
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4242728184553661e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.328729852541574,
      "tokens_seen": 3253796864
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4222668004012037e-05,
      "loss": 1.2017,
      "theoretical_loss": 3.328724825675577,
      "tokens_seen": 3253862400
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4202607823470411e-05,
      "loss": 1.2094,
      "theoretical_loss": 3.328719798939174,
      "tokens_seen": 3253927936
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4182547642928787e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.328714772332358,
      "tokens_seen": 3253993472
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4162487462387161e-05,
      "loss": 1.1728,
      "theoretical_loss": 3.328709745855123,
      "tokens_seen": 3254059008
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4142427281845538e-05,
      "loss": 1.385,
      "theoretical_loss": 3.328704719507464,
      "tokens_seen": 3254124544
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4122367101303913e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.3286996932893738,
      "tokens_seen": 3254190080
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4102306920762287e-05,
      "loss": 1.3689,
      "theoretical_loss": 3.328694667200848,
      "tokens_seen": 3254255616
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4082246740220662e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.328689641241879,
      "tokens_seen": 3254321152
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4062186559679037e-05,
      "loss": 1.2017,
      "theoretical_loss": 3.328684615412462,
      "tokens_seen": 3254386688
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4042126379137412e-05,
      "loss": 1.3289,
      "theoretical_loss": 3.3286795897125905,
      "tokens_seen": 3254452224
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4022066198595788e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.3286745641422586,
      "tokens_seen": 3254517760
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.4002006018054164e-05,
      "loss": 1.315,
      "theoretical_loss": 3.328669538701461,
      "tokens_seen": 3254583296
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3981945837512538e-05,
      "loss": 1.3348,
      "theoretical_loss": 3.3286645133901906,
      "tokens_seen": 3254648832
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3961885656970912e-05,
      "loss": 1.1576,
      "theoretical_loss": 3.3286594882084426,
      "tokens_seen": 3254714368
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3941825476429288e-05,
      "loss": 1.2104,
      "theoretical_loss": 3.32865446315621,
      "tokens_seen": 3254779904
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3921765295887662e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.3286494382334877,
      "tokens_seen": 3254845440
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.390170511534604e-05,
      "loss": 1.3548,
      "theoretical_loss": 3.3286444134402693,
      "tokens_seen": 3254910976
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3881644934804414e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.3286393887765495,
      "tokens_seen": 3254976512
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.386158475426279e-05,
      "loss": 1.2072,
      "theoretical_loss": 3.3286343642423213,
      "tokens_seen": 3255042048
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3841524573721163e-05,
      "loss": 1.2237,
      "theoretical_loss": 3.3286293398375792,
      "tokens_seen": 3255107584
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3821464393179538e-05,
      "loss": 1.3324,
      "theoretical_loss": 3.328624315562318,
      "tokens_seen": 3255173120
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3801404212637915e-05,
      "loss": 1.2528,
      "theoretical_loss": 3.3286192914165307,
      "tokens_seen": 3255238656
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.378134403209629e-05,
      "loss": 1.1666,
      "theoretical_loss": 3.3286142674002117,
      "tokens_seen": 3255304192
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3761283851554665e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.3286092435133554,
      "tokens_seen": 3255369728
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3741223671013039e-05,
      "loss": 1.2528,
      "theoretical_loss": 3.328604219755955,
      "tokens_seen": 3255435264
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3721163490471415e-05,
      "loss": 1.1949,
      "theoretical_loss": 3.328599196128006,
      "tokens_seen": 3255500800
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3701103309929789e-05,
      "loss": 1.2392,
      "theoretical_loss": 3.328594172629501,
      "tokens_seen": 3255566336
    },
    {
      "epoch": 0.97,
      "objective/train/docs_used": 1182603,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2460951805114746,
      "objective/train/theoretical_loss": 3.328589149260435,
      "objective/train/tokens_used": 1626090976,
      "theoretical_loss": 3.328589149260435,
      "tokens_seen": 3255631872
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3681043129388165e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.328589149260435,
      "tokens_seen": 3255631872
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.366098294884654e-05,
      "loss": 1.1836,
      "theoretical_loss": 3.3285841260208016,
      "tokens_seen": 3255697408
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3640922768304915e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.328579102910595,
      "tokens_seen": 3255762944
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.362086258776329e-05,
      "loss": 1.0648,
      "theoretical_loss": 3.328574079929809,
      "tokens_seen": 3255828480
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3600802407221665e-05,
      "loss": 1.2075,
      "theoretical_loss": 3.3285690570784383,
      "tokens_seen": 3255894016
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.358074222668004e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.3285640343564764,
      "tokens_seen": 3255959552
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3560682046138416e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.3285590117639177,
      "tokens_seen": 3256025088
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.354062186559679e-05,
      "loss": 1.293,
      "theoretical_loss": 3.328553989300756,
      "tokens_seen": 3256090624
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3520561685055166e-05,
      "loss": 1.3946,
      "theoretical_loss": 3.328548966966985,
      "tokens_seen": 3256156160
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.350050150451354e-05,
      "loss": 1.2306,
      "theoretical_loss": 3.3285439447626,
      "tokens_seen": 3256221696
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3480441323971916e-05,
      "loss": 1.1951,
      "theoretical_loss": 3.328538922687594,
      "tokens_seen": 3256287232
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3460381143430292e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.3285339007419608,
      "tokens_seen": 3256352768
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3440320962888667e-05,
      "loss": 1.2091,
      "theoretical_loss": 3.3285288789256957,
      "tokens_seen": 3256418304
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3420260782347042e-05,
      "loss": 1.279,
      "theoretical_loss": 3.3285238572387916,
      "tokens_seen": 3256483840
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3400200601805416e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.328518835681243,
      "tokens_seen": 3256549376
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3380140421263791e-05,
      "loss": 1.3645,
      "theoretical_loss": 3.3285138142530446,
      "tokens_seen": 3256614912
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3360080240722166e-05,
      "loss": 1.2286,
      "theoretical_loss": 3.328508792954189,
      "tokens_seen": 3256680448
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3340020060180543e-05,
      "loss": 1.3713,
      "theoretical_loss": 3.328503771784672,
      "tokens_seen": 3256745984
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3319959879638917e-05,
      "loss": 1.3198,
      "theoretical_loss": 3.3284987507444863,
      "tokens_seen": 3256811520
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3299899699097293e-05,
      "loss": 1.287,
      "theoretical_loss": 3.3284937298336263,
      "tokens_seen": 3256877056
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3279839518555667e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.328488709052087,
      "tokens_seen": 3256942592
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3259779338014041e-05,
      "loss": 1.1663,
      "theoretical_loss": 3.328483688399861,
      "tokens_seen": 3257008128
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3239719157472419e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.328478667876943,
      "tokens_seen": 3257073664
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3219658976930793e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.3284736474833276,
      "tokens_seen": 3257139200
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3199598796389169e-05,
      "loss": 1.1923,
      "theoretical_loss": 3.3284686272190083,
      "tokens_seen": 3257204736
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3179538615847543e-05,
      "loss": 1.2388,
      "theoretical_loss": 3.328463607083979,
      "tokens_seen": 3257270272
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3159478435305918e-05,
      "loss": 1.315,
      "theoretical_loss": 3.3284585870782344,
      "tokens_seen": 3257335808
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3139418254764293e-05,
      "loss": 1.1677,
      "theoretical_loss": 3.328453567201768,
      "tokens_seen": 3257401344
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3119358074222668e-05,
      "loss": 1.3016,
      "theoretical_loss": 3.328448547454574,
      "tokens_seen": 3257466880
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3099297893681044e-05,
      "loss": 1.2404,
      "theoretical_loss": 3.3284435278366464,
      "tokens_seen": 3257532416
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3079237713139418e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.32843850834798,
      "tokens_seen": 3257597952
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3059177532597794e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.328433488988568,
      "tokens_seen": 3257663488
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3039117352056168e-05,
      "loss": 1.2481,
      "theoretical_loss": 3.3284284697584043,
      "tokens_seen": 3257729024
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3019057171514544e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.3284234506574837,
      "tokens_seen": 3257794560
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.299899699097292e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.3284184316858,
      "tokens_seen": 3257860096
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.2978936810431294e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.328413412843348,
      "tokens_seen": 3257925632
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.295887662988967e-05,
      "loss": 1.2656,
      "theoretical_loss": 3.32840839413012,
      "tokens_seen": 3257991168
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.2938816449348044e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.3284033755461118,
      "tokens_seen": 3258056704
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.291875626880642e-05,
      "loss": 1.2501,
      "theoretical_loss": 3.3283983570913165,
      "tokens_seen": 3258122240
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.2898696088264795e-05,
      "loss": 1.2104,
      "theoretical_loss": 3.3283933387657285,
      "tokens_seen": 3258187776
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.2878635907723171e-05,
      "loss": 1.2428,
      "theoretical_loss": 3.328388320569342,
      "tokens_seen": 3258253312
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.2858575727181545e-05,
      "loss": 1.2353,
      "theoretical_loss": 3.3283833025021505,
      "tokens_seen": 3258318848
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.283851554663992e-05,
      "loss": 1.2346,
      "theoretical_loss": 3.328378284564149,
      "tokens_seen": 3258384384
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.2818455366098295e-05,
      "loss": 1.2238,
      "theoretical_loss": 3.3283732667553307,
      "tokens_seen": 3258449920
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.2798395185556669e-05,
      "loss": 1.2167,
      "theoretical_loss": 3.32836824907569,
      "tokens_seen": 3258515456
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.2778335005015047e-05,
      "loss": 1.2314,
      "theoretical_loss": 3.3283632315252216,
      "tokens_seen": 3258580992
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.275827482447342e-05,
      "loss": 1.1805,
      "theoretical_loss": 3.3283582141039187,
      "tokens_seen": 3258646528
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.2738214643931797e-05,
      "loss": 1.1679,
      "theoretical_loss": 3.3283531968117757,
      "tokens_seen": 3258712064
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.271815446339017e-05,
      "loss": 1.1496,
      "theoretical_loss": 3.3283481796487866,
      "tokens_seen": 3258777600
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2698094282848545e-05,
      "loss": 1.1899,
      "theoretical_loss": 3.3283431626149453,
      "tokens_seen": 3258843136
    },
    {
      "epoch": 0.98,
      "objective/train/docs_used": 1183737,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.291218638420105,
      "objective/train/theoretical_loss": 3.328338145710246,
      "objective/train/tokens_used": 1629367776,
      "theoretical_loss": 3.328338145710246,
      "tokens_seen": 3258908672
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.267803410230692e-05,
      "loss": 1.2191,
      "theoretical_loss": 3.328338145710246,
      "tokens_seen": 3258908672
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2657973921765296e-05,
      "loss": 1.1466,
      "theoretical_loss": 3.3283331289346836,
      "tokens_seen": 3258974208
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2637913741223672e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.328328112288251,
      "tokens_seen": 3259039744
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2617853560682046e-05,
      "loss": 1.2066,
      "theoretical_loss": 3.328323095770943,
      "tokens_seen": 3259105280
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2597793380140422e-05,
      "loss": 1.2177,
      "theoretical_loss": 3.328318079382753,
      "tokens_seen": 3259170816
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2577733199598796e-05,
      "loss": 1.2354,
      "theoretical_loss": 3.3283130631236757,
      "tokens_seen": 3259236352
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2557673019057172e-05,
      "loss": 1.2309,
      "theoretical_loss": 3.328308046993705,
      "tokens_seen": 3259301888
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2537612838515548e-05,
      "loss": 1.2423,
      "theoretical_loss": 3.328303030992835,
      "tokens_seen": 3259367424
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2517552657973922e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.32829801512106,
      "tokens_seen": 3259432960
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2497492477432298e-05,
      "loss": 1.2153,
      "theoretical_loss": 3.3282929993783736,
      "tokens_seen": 3259498496
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2477432296890672e-05,
      "loss": 1.1429,
      "theoretical_loss": 3.32828798376477,
      "tokens_seen": 3259564032
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2457372116349047e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.3282829682802437,
      "tokens_seen": 3259629568
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2437311935807423e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.3282779529247883,
      "tokens_seen": 3259695104
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2417251755265797e-05,
      "loss": 1.3375,
      "theoretical_loss": 3.328272937698398,
      "tokens_seen": 3259760640
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2397191574724173e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.328267922601067,
      "tokens_seen": 3259826176
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2377131394182547e-05,
      "loss": 1.1663,
      "theoretical_loss": 3.3282629076327894,
      "tokens_seen": 3259891712
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2357071213640923e-05,
      "loss": 1.2167,
      "theoretical_loss": 3.3282578927935593,
      "tokens_seen": 3259957248
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2337011033099299e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.3282528780833704,
      "tokens_seen": 3260022784
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2316950852557675e-05,
      "loss": 1.1532,
      "theoretical_loss": 3.3282478635022175,
      "tokens_seen": 3260088320
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2296890672016049e-05,
      "loss": 1.148,
      "theoretical_loss": 3.328242849050094,
      "tokens_seen": 3260153856
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2276830491474423e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.3282378347269943,
      "tokens_seen": 3260219392
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2256770310932799e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.3282328205329126,
      "tokens_seen": 3260284928
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2236710130391173e-05,
      "loss": 1.1476,
      "theoretical_loss": 3.3282278064678428,
      "tokens_seen": 3260350464
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.221664994984955e-05,
      "loss": 1.1461,
      "theoretical_loss": 3.3282227925317787,
      "tokens_seen": 3260416000
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2196589769307924e-05,
      "loss": 1.1231,
      "theoretical_loss": 3.328217778724715,
      "tokens_seen": 3260481536
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2176529588766298e-05,
      "loss": 1.2041,
      "theoretical_loss": 3.3282127650466458,
      "tokens_seen": 3260547072
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2156469408224674e-05,
      "loss": 1.1912,
      "theoretical_loss": 3.3282077514975645,
      "tokens_seen": 3260612608
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2136409227683048e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.3282027380774655,
      "tokens_seen": 3260678144
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2116349047141424e-05,
      "loss": 1.1813,
      "theoretical_loss": 3.3281977247863432,
      "tokens_seen": 3260743680
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.20962888665998e-05,
      "loss": 1.2307,
      "theoretical_loss": 3.328192711624191,
      "tokens_seen": 3260809216
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2076228686058176e-05,
      "loss": 1.15,
      "theoretical_loss": 3.328187698591004,
      "tokens_seen": 3260874752
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.205616850551655e-05,
      "loss": 1.1544,
      "theoretical_loss": 3.3281826856867758,
      "tokens_seen": 3260940288
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2036108324974924e-05,
      "loss": 1.2035,
      "theoretical_loss": 3.3281776729115,
      "tokens_seen": 3261005824
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.20160481444333e-05,
      "loss": 1.1989,
      "theoretical_loss": 3.3281726602651713,
      "tokens_seen": 3261071360
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1995987963891675e-05,
      "loss": 1.3514,
      "theoretical_loss": 3.3281676477477835,
      "tokens_seen": 3261136896
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1975927783350051e-05,
      "loss": 1.1314,
      "theoretical_loss": 3.3281626353593308,
      "tokens_seen": 3261202432
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1955867602808425e-05,
      "loss": 1.2421,
      "theoretical_loss": 3.3281576230998073,
      "tokens_seen": 3261267968
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1935807422266801e-05,
      "loss": 1.1828,
      "theoretical_loss": 3.3281526109692074,
      "tokens_seen": 3261333504
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1915747241725175e-05,
      "loss": 1.1907,
      "theoretical_loss": 3.328147598967525,
      "tokens_seen": 3261399040
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.189568706118355e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.3281425870947534,
      "tokens_seen": 3261464576
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1875626880641927e-05,
      "loss": 1.1705,
      "theoretical_loss": 3.3281375753508877,
      "tokens_seen": 3261530112
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1855566700100301e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.3281325637359216,
      "tokens_seen": 3261595648
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1835506519558677e-05,
      "loss": 1.188,
      "theoretical_loss": 3.3281275522498497,
      "tokens_seen": 3261661184
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.181544633901705e-05,
      "loss": 1.1853,
      "theoretical_loss": 3.3281225408926653,
      "tokens_seen": 3261726720
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1795386158475427e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.328117529664363,
      "tokens_seen": 3261792256
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1775325977933802e-05,
      "loss": 1.2031,
      "theoretical_loss": 3.3281125185649363,
      "tokens_seen": 3261857792
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1755265797392177e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.32810750759438,
      "tokens_seen": 3261923328
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1735205616850552e-05,
      "loss": 1.1188,
      "theoretical_loss": 3.328102496752688,
      "tokens_seen": 3261988864
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1715145436308926e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.3280974860398542,
      "tokens_seen": 3262054400
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1695085255767302e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.328092475455873,
      "tokens_seen": 3262119936
    },
    {
      "epoch": 0.98,
      "objective/train/docs_used": 1184948,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.343144178390503,
      "objective/train/theoretical_loss": 3.3280874650007384,
      "objective/train/tokens_used": 1632644576,
      "theoretical_loss": 3.3280874650007384,
      "tokens_seen": 3262185472
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1675025075225676e-05,
      "loss": 1.3316,
      "theoretical_loss": 3.3280874650007384,
      "tokens_seen": 3262185472
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1654964894684054e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.3280824546744445,
      "tokens_seen": 3262251008
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1634904714142428e-05,
      "loss": 1.3283,
      "theoretical_loss": 3.328077444476985,
      "tokens_seen": 3262316544
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1614844533600802e-05,
      "loss": 1.3862,
      "theoretical_loss": 3.3280724344083543,
      "tokens_seen": 3262382080
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1594784353059178e-05,
      "loss": 1.1528,
      "theoretical_loss": 3.328067424468547,
      "tokens_seen": 3262447616
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1574724172517552e-05,
      "loss": 1.221,
      "theoretical_loss": 3.3280624146575564,
      "tokens_seen": 3262513152
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1554663991975928e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.328057404975377,
      "tokens_seen": 3262578688
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1534603811434303e-05,
      "loss": 1.1985,
      "theoretical_loss": 3.3280523954220027,
      "tokens_seen": 3262644224
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.151454363089268e-05,
      "loss": 1.2146,
      "theoretical_loss": 3.3280473859974276,
      "tokens_seen": 3262709760
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1494483450351053e-05,
      "loss": 1.4185,
      "theoretical_loss": 3.3280423767016467,
      "tokens_seen": 3262775296
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1474423269809427e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.3280373675346526,
      "tokens_seen": 3262840832
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1454363089267803e-05,
      "loss": 1.3503,
      "theoretical_loss": 3.3280323584964404,
      "tokens_seen": 3262906368
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1434302908726179e-05,
      "loss": 1.249,
      "theoretical_loss": 3.328027349587004,
      "tokens_seen": 3262971904
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1414242728184555e-05,
      "loss": 1.0516,
      "theoretical_loss": 3.328022340806337,
      "tokens_seen": 3263037440
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1394182547642929e-05,
      "loss": 1.2017,
      "theoretical_loss": 3.328017332154434,
      "tokens_seen": 3263102976
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1374122367101305e-05,
      "loss": 1.1819,
      "theoretical_loss": 3.3280123236312895,
      "tokens_seen": 3263168512
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1354062186559679e-05,
      "loss": 1.2312,
      "theoretical_loss": 3.328007315236897,
      "tokens_seen": 3263234048
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1334002006018053e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.3280023069712508,
      "tokens_seen": 3263299584
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.131394182547643e-05,
      "loss": 1.1933,
      "theoretical_loss": 3.3279972988343447,
      "tokens_seen": 3263365120
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1293881644934805e-05,
      "loss": 1.0615,
      "theoretical_loss": 3.3279922908261734,
      "tokens_seen": 3263430656
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.127382146439318e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.3279872829467307,
      "tokens_seen": 3263496192
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1253761283851554e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.3279822751960104,
      "tokens_seen": 3263561728
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.123370110330993e-05,
      "loss": 1.1968,
      "theoretical_loss": 3.327977267574007,
      "tokens_seen": 3263627264
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1213640922768304e-05,
      "loss": 1.2229,
      "theoretical_loss": 3.3279722600807142,
      "tokens_seen": 3263692800
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.119358074222668e-05,
      "loss": 1.2492,
      "theoretical_loss": 3.327967252716127,
      "tokens_seen": 3263758336
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1173520561685056e-05,
      "loss": 1.2826,
      "theoretical_loss": 3.3279622454802382,
      "tokens_seen": 3263823872
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.115346038114343e-05,
      "loss": 1.2047,
      "theoretical_loss": 3.3279572383730436,
      "tokens_seen": 3263889408
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1133400200601806e-05,
      "loss": 1.2129,
      "theoretical_loss": 3.3279522313945353,
      "tokens_seen": 3263954944
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.111334002006018e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.327947224544709,
      "tokens_seen": 3264020480
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1093279839518557e-05,
      "loss": 1.2205,
      "theoretical_loss": 3.3279422178235585,
      "tokens_seen": 3264086016
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1073219658976931e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.327937211231077,
      "tokens_seen": 3264151552
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1053159478435306e-05,
      "loss": 1.2399,
      "theoretical_loss": 3.3279322047672597,
      "tokens_seen": 3264217088
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1033099297893681e-05,
      "loss": 1.2497,
      "theoretical_loss": 3.3279271984321,
      "tokens_seen": 3264282624
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1013039117352055e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.3279221922255924,
      "tokens_seen": 3264348160
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0992978936810431e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.327917186147731,
      "tokens_seen": 3264413696
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0972918756268807e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.3279121801985094,
      "tokens_seen": 3264479232
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0952858575727183e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.3279071743779225,
      "tokens_seen": 3264544768
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0932798395185557e-05,
      "loss": 1.2426,
      "theoretical_loss": 3.327902168685964,
      "tokens_seen": 3264610304
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0912738214643931e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.327897163122628,
      "tokens_seen": 3264675840
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0892678034102307e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.3278921576879084,
      "tokens_seen": 3264741376
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0872617853560683e-05,
      "loss": 1.245,
      "theoretical_loss": 3.3278871523818,
      "tokens_seen": 3264806912
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0852557673019058e-05,
      "loss": 1.3083,
      "theoretical_loss": 3.3278821472042965,
      "tokens_seen": 3264872448
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0832497492477433e-05,
      "loss": 1.239,
      "theoretical_loss": 3.3278771421553914,
      "tokens_seen": 3264937984
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0812437311935808e-05,
      "loss": 1.2552,
      "theoretical_loss": 3.32787213723508,
      "tokens_seen": 3265003520
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0792377131394182e-05,
      "loss": 1.1537,
      "theoretical_loss": 3.3278671324433557,
      "tokens_seen": 3265069056
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0772316950852556e-05,
      "loss": 1.303,
      "theoretical_loss": 3.3278621277802127,
      "tokens_seen": 3265134592
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0752256770310934e-05,
      "loss": 1.1228,
      "theoretical_loss": 3.327857123245645,
      "tokens_seen": 3265200128
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0732196589769308e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.327852118839647,
      "tokens_seen": 3265265664
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0712136409227684e-05,
      "loss": 1.2225,
      "theoretical_loss": 3.327847114562213,
      "tokens_seen": 3265331200
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0692076228686058e-05,
      "loss": 1.3288,
      "theoretical_loss": 3.327842110413336,
      "tokens_seen": 3265396736
    },
    {
      "epoch": 0.98,
      "objective/train/docs_used": 1186260,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2703136205673218,
      "objective/train/theoretical_loss": 3.3278371063930114,
      "objective/train/tokens_used": 1635921376,
      "theoretical_loss": 3.3278371063930114,
      "tokens_seen": 3265462272
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0672016048144434e-05,
      "loss": 1.2499,
      "theoretical_loss": 3.3278371063930114,
      "tokens_seen": 3265462272
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0651955867602808e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.327832102501233,
      "tokens_seen": 3265527808
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0631895687061184e-05,
      "loss": 1.3546,
      "theoretical_loss": 3.3278270987379948,
      "tokens_seen": 3265593344
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.061183550651956e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.327822095103291,
      "tokens_seen": 3265658880
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0591775325977934e-05,
      "loss": 1.1893,
      "theoretical_loss": 3.3278170915971153,
      "tokens_seen": 3265724416
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.057171514543631e-05,
      "loss": 1.2368,
      "theoretical_loss": 3.3278120882194617,
      "tokens_seen": 3265789952
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0551654964894683e-05,
      "loss": 1.1226,
      "theoretical_loss": 3.3278070849703254,
      "tokens_seen": 3265855488
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0531594784353061e-05,
      "loss": 1.1854,
      "theoretical_loss": 3.3278020818496996,
      "tokens_seen": 3265921024
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0511534603811435e-05,
      "loss": 1.2571,
      "theoretical_loss": 3.3277970788575786,
      "tokens_seen": 3265986560
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0491474423269809e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.327792075993957,
      "tokens_seen": 3266052096
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0471414242728185e-05,
      "loss": 1.3108,
      "theoretical_loss": 3.327787073258828,
      "tokens_seen": 3266117632
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0451354062186559e-05,
      "loss": 1.231,
      "theoretical_loss": 3.327782070652187,
      "tokens_seen": 3266183168
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0431293881644935e-05,
      "loss": 1.1006,
      "theoretical_loss": 3.3277770681740266,
      "tokens_seen": 3266248704
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.041123370110331e-05,
      "loss": 1.282,
      "theoretical_loss": 3.327772065824342,
      "tokens_seen": 3266314240
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0391173520561686e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.327767063603127,
      "tokens_seen": 3266379776
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.037111334002006e-05,
      "loss": 1.274,
      "theoretical_loss": 3.3277620615103753,
      "tokens_seen": 3266445312
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0351053159478435e-05,
      "loss": 1.2055,
      "theoretical_loss": 3.327757059546082,
      "tokens_seen": 3266510848
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.033099297893681e-05,
      "loss": 1.207,
      "theoretical_loss": 3.3277520577102404,
      "tokens_seen": 3266576384
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0310932798395184e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.327747056002845,
      "tokens_seen": 3266641920
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0290872617853562e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.32774205442389,
      "tokens_seen": 3266707456
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0270812437311936e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.327737052973369,
      "tokens_seen": 3266772992
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0250752256770312e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.327732051651277,
      "tokens_seen": 3266838528
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0230692076228686e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.3277270504576073,
      "tokens_seen": 3266904064
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.021063189568706e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.3277220493923543,
      "tokens_seen": 3266969600
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0190571715145438e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.3277170484555123,
      "tokens_seen": 3267035136
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0170511534603812e-05,
      "loss": 1.2062,
      "theoretical_loss": 3.327712047647075,
      "tokens_seen": 3267100672
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0150451354062187e-05,
      "loss": 1.2556,
      "theoretical_loss": 3.327707046967037,
      "tokens_seen": 3267166208
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0130391173520562e-05,
      "loss": 1.1729,
      "theoretical_loss": 3.3277020464153924,
      "tokens_seen": 3267231744
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0110330992978937e-05,
      "loss": 1.1693,
      "theoretical_loss": 3.327697045992135,
      "tokens_seen": 3267297280
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0090270812437311e-05,
      "loss": 1.2212,
      "theoretical_loss": 3.327692045697259,
      "tokens_seen": 3267362816
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0070210631895687e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.327687045530759,
      "tokens_seen": 3267428352
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0050150451354063e-05,
      "loss": 1.2405,
      "theoretical_loss": 3.327682045492628,
      "tokens_seen": 3267493888
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0030090270812437e-05,
      "loss": 1.2056,
      "theoretical_loss": 3.3276770455828615,
      "tokens_seen": 3267559424
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0010030090270813e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.327672045801453,
      "tokens_seen": 3267624960
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.989969909729187e-06,
      "loss": 1.215,
      "theoretical_loss": 3.3276670461483966,
      "tokens_seen": 3267690496
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.969909729187565e-06,
      "loss": 1.3275,
      "theoretical_loss": 3.3276620466236864,
      "tokens_seen": 3267756032
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.949849548645939e-06,
      "loss": 1.1964,
      "theoretical_loss": 3.3276570472273166,
      "tokens_seen": 3267821568
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.929789368104313e-06,
      "loss": 1.3369,
      "theoretical_loss": 3.3276520479592815,
      "tokens_seen": 3267887104
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.909729187562688e-06,
      "loss": 1.2774,
      "theoretical_loss": 3.327647048819575,
      "tokens_seen": 3267952640
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.889669007021063e-06,
      "loss": 1.2673,
      "theoretical_loss": 3.327642049808191,
      "tokens_seen": 3268018176
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.869608826479438e-06,
      "loss": 1.1772,
      "theoretical_loss": 3.3276370509251243,
      "tokens_seen": 3268083712
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.849548645937814e-06,
      "loss": 1.312,
      "theoretical_loss": 3.3276320521703684,
      "tokens_seen": 3268149248
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.82948846539619e-06,
      "loss": 1.1865,
      "theoretical_loss": 3.327627053543918,
      "tokens_seen": 3268214784
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.809428284854564e-06,
      "loss": 1.2231,
      "theoretical_loss": 3.327622055045767,
      "tokens_seen": 3268280320
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.789368104312938e-06,
      "loss": 1.3034,
      "theoretical_loss": 3.3276170566759093,
      "tokens_seen": 3268345856
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.769307923771314e-06,
      "loss": 1.3538,
      "theoretical_loss": 3.327612058434339,
      "tokens_seen": 3268411392
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.749247743229688e-06,
      "loss": 1.16,
      "theoretical_loss": 3.327607060321051,
      "tokens_seen": 3268476928
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.729187562688066e-06,
      "loss": 1.1521,
      "theoretical_loss": 3.3276020623360383,
      "tokens_seen": 3268542464
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.70912738214644e-06,
      "loss": 1.1405,
      "theoretical_loss": 3.3275970644792956,
      "tokens_seen": 3268608000
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.689067201604815e-06,
      "loss": 1.2841,
      "theoretical_loss": 3.3275920667508174,
      "tokens_seen": 3268673536
    },
    {
      "epoch": 0.98,
      "objective/train/docs_used": 1187457,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2629066705703735,
      "objective/train/theoretical_loss": 3.327587069150597,
      "objective/train/tokens_used": 1639198176,
      "theoretical_loss": 3.327587069150597,
      "tokens_seen": 3268739072
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.66900702106319e-06,
      "loss": 1.2422,
      "theoretical_loss": 3.327587069150597,
      "tokens_seen": 3268739072
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.648946840521564e-06,
      "loss": 1.2651,
      "theoretical_loss": 3.3275820716786297,
      "tokens_seen": 3268804608
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.628886659979941e-06,
      "loss": 1.2401,
      "theoretical_loss": 3.3275770743349087,
      "tokens_seen": 3268870144
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.608826479438315e-06,
      "loss": 1.2943,
      "theoretical_loss": 3.327572077119428,
      "tokens_seen": 3268935680
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.588766298896691e-06,
      "loss": 1.2836,
      "theoretical_loss": 3.327567080032183,
      "tokens_seen": 3269001216
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.568706118355065e-06,
      "loss": 1.1451,
      "theoretical_loss": 3.327562083073166,
      "tokens_seen": 3269066752
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.54864593781344e-06,
      "loss": 1.1994,
      "theoretical_loss": 3.327557086242373,
      "tokens_seen": 3269132288
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.528585757271815e-06,
      "loss": 1.4072,
      "theoretical_loss": 3.3275520895397968,
      "tokens_seen": 3269197824
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.50852557673019e-06,
      "loss": 1.3884,
      "theoretical_loss": 3.327547092965432,
      "tokens_seen": 3269263360
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.488465396188567e-06,
      "loss": 1.3081,
      "theoretical_loss": 3.3275420965192724,
      "tokens_seen": 3269328896
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.46840521564694e-06,
      "loss": 1.1851,
      "theoretical_loss": 3.327537100201313,
      "tokens_seen": 3269394432
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.448345035105316e-06,
      "loss": 1.2105,
      "theoretical_loss": 3.3275321040115475,
      "tokens_seen": 3269459968
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.42828485456369e-06,
      "loss": 1.1922,
      "theoretical_loss": 3.3275271079499693,
      "tokens_seen": 3269525504
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.408224674022065e-06,
      "loss": 1.227,
      "theoretical_loss": 3.327522112016574,
      "tokens_seen": 3269591040
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.388164493480442e-06,
      "loss": 1.2702,
      "theoretical_loss": 3.3275171162113546,
      "tokens_seen": 3269656576
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.368104312938816e-06,
      "loss": 1.3654,
      "theoretical_loss": 3.3275121205343052,
      "tokens_seen": 3269722112
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.348044132397192e-06,
      "loss": 1.2376,
      "theoretical_loss": 3.327507124985421,
      "tokens_seen": 3269787648
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.327983951855566e-06,
      "loss": 1.2938,
      "theoretical_loss": 3.327502129564695,
      "tokens_seen": 3269853184
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.307923771313942e-06,
      "loss": 1.1103,
      "theoretical_loss": 3.327497134272122,
      "tokens_seen": 3269918720
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.287863590772318e-06,
      "loss": 1.2271,
      "theoretical_loss": 3.3274921391076955,
      "tokens_seen": 3269984256
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.267803410230692e-06,
      "loss": 1.2453,
      "theoretical_loss": 3.327487144071411,
      "tokens_seen": 3270049792
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.247743229689068e-06,
      "loss": 1.1314,
      "theoretical_loss": 3.327482149163261,
      "tokens_seen": 3270115328
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.227683049147442e-06,
      "loss": 1.2628,
      "theoretical_loss": 3.3274771543832404,
      "tokens_seen": 3270180864
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.207622868605818e-06,
      "loss": 1.225,
      "theoretical_loss": 3.327472159731344,
      "tokens_seen": 3270246400
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.187562688064192e-06,
      "loss": 1.1356,
      "theoretical_loss": 3.327467165207565,
      "tokens_seen": 3270311936
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.167502507522569e-06,
      "loss": 1.2122,
      "theoretical_loss": 3.3274621708118977,
      "tokens_seen": 3270377472
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.147442326980943e-06,
      "loss": 1.1149,
      "theoretical_loss": 3.3274571765443364,
      "tokens_seen": 3270443008
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.127382146439317e-06,
      "loss": 1.1836,
      "theoretical_loss": 3.327452182404875,
      "tokens_seen": 3270508544
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.107321965897693e-06,
      "loss": 1.2738,
      "theoretical_loss": 3.327447188393508,
      "tokens_seen": 3270574080
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.087261785356067e-06,
      "loss": 1.2059,
      "theoretical_loss": 3.3274421945102297,
      "tokens_seen": 3270639616
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.067201604814445e-06,
      "loss": 1.1482,
      "theoretical_loss": 3.3274372007550337,
      "tokens_seen": 3270705152
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.047141424272819e-06,
      "loss": 1.1733,
      "theoretical_loss": 3.3274322071279143,
      "tokens_seen": 3270770688
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.027081243731195e-06,
      "loss": 1.298,
      "theoretical_loss": 3.327427213628866,
      "tokens_seen": 3270836224
    },
    {
      "epoch": 0.98,
      "learning_rate": 9.007021063189569e-06,
      "loss": 1.3288,
      "theoretical_loss": 3.3274222202578825,
      "tokens_seen": 3270901760
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.986960882647943e-06,
      "loss": 1.2179,
      "theoretical_loss": 3.3274172270149585,
      "tokens_seen": 3270967296
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.966900702106319e-06,
      "loss": 1.113,
      "theoretical_loss": 3.3274122339000876,
      "tokens_seen": 3271032832
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.946840521564694e-06,
      "loss": 1.3052,
      "theoretical_loss": 3.3274072409132645,
      "tokens_seen": 3271098368
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.92678034102307e-06,
      "loss": 1.1318,
      "theoretical_loss": 3.3274022480544825,
      "tokens_seen": 3271163904
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.906720160481444e-06,
      "loss": 1.2751,
      "theoretical_loss": 3.3273972553237363,
      "tokens_seen": 3271229440
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.88665997993982e-06,
      "loss": 1.2206,
      "theoretical_loss": 3.3273922627210206,
      "tokens_seen": 3271294976
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.866599799398194e-06,
      "loss": 1.2586,
      "theoretical_loss": 3.3273872702463283,
      "tokens_seen": 3271360512
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.846539618856568e-06,
      "loss": 1.2474,
      "theoretical_loss": 3.3273822778996545,
      "tokens_seen": 3271426048
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.826479438314946e-06,
      "loss": 1.2413,
      "theoretical_loss": 3.327377285680993,
      "tokens_seen": 3271491584
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.80641925777332e-06,
      "loss": 1.1794,
      "theoretical_loss": 3.3273722935903383,
      "tokens_seen": 3271557120
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.786359077231696e-06,
      "loss": 1.2847,
      "theoretical_loss": 3.327367301627684,
      "tokens_seen": 3271622656
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.76629889669007e-06,
      "loss": 1.2262,
      "theoretical_loss": 3.327362309793025,
      "tokens_seen": 3271688192
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.746238716148446e-06,
      "loss": 1.3228,
      "theoretical_loss": 3.3273573180863543,
      "tokens_seen": 3271753728
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.726178535606821e-06,
      "loss": 1.1662,
      "theoretical_loss": 3.327352326507667,
      "tokens_seen": 3271819264
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.706118355065195e-06,
      "loss": 1.2003,
      "theoretical_loss": 3.327347335056957,
      "tokens_seen": 3271884800
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.686058174523571e-06,
      "loss": 1.0696,
      "theoretical_loss": 3.3273423437342187,
      "tokens_seen": 3271950336
    },
    {
      "epoch": 0.98,
      "objective/train/docs_used": 1188646,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.1247133016586304,
      "objective/train/theoretical_loss": 3.3273373525394456,
      "objective/train/tokens_used": 1642474976,
      "theoretical_loss": 3.3273373525394456,
      "tokens_seen": 3272015872
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.665997993981945e-06,
      "loss": 1.1487,
      "theoretical_loss": 3.3273373525394456,
      "tokens_seen": 3272015872
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.645937813440321e-06,
      "loss": 1.2024,
      "theoretical_loss": 3.327332361472633,
      "tokens_seen": 3272081408
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.625877632898695e-06,
      "loss": 1.2993,
      "theoretical_loss": 3.327327370533774,
      "tokens_seen": 3272146944
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.605817452357073e-06,
      "loss": 1.1568,
      "theoretical_loss": 3.3273223797228626,
      "tokens_seen": 3272212480
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.585757271815447e-06,
      "loss": 1.2633,
      "theoretical_loss": 3.3273173890398935,
      "tokens_seen": 3272278016
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.565697091273821e-06,
      "loss": 1.2707,
      "theoretical_loss": 3.327312398484861,
      "tokens_seen": 3272343552
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.545636910732197e-06,
      "loss": 1.3644,
      "theoretical_loss": 3.3273074080577594,
      "tokens_seen": 3272409088
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.52557673019057e-06,
      "loss": 1.347,
      "theoretical_loss": 3.3273024177585824,
      "tokens_seen": 3272474624
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.505516549648948e-06,
      "loss": 1.2442,
      "theoretical_loss": 3.327297427587324,
      "tokens_seen": 3272540160
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.485456369107322e-06,
      "loss": 1.1388,
      "theoretical_loss": 3.3272924375439787,
      "tokens_seen": 3272605696
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.465396188565698e-06,
      "loss": 1.3811,
      "theoretical_loss": 3.3272874476285406,
      "tokens_seen": 3272671232
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.445336008024072e-06,
      "loss": 1.2336,
      "theoretical_loss": 3.3272824578410036,
      "tokens_seen": 3272736768
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.425275827482446e-06,
      "loss": 1.3284,
      "theoretical_loss": 3.3272774681813626,
      "tokens_seen": 3272802304
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.405215646940822e-06,
      "loss": 1.0991,
      "theoretical_loss": 3.3272724786496113,
      "tokens_seen": 3272867840
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.385155466399198e-06,
      "loss": 1.3947,
      "theoretical_loss": 3.3272674892457434,
      "tokens_seen": 3272933376
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.365095285857574e-06,
      "loss": 1.1369,
      "theoretical_loss": 3.3272624999697538,
      "tokens_seen": 3272998912
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.345035105315948e-06,
      "loss": 1.2269,
      "theoretical_loss": 3.3272575108216365,
      "tokens_seen": 3273064448
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.324974924774324e-06,
      "loss": 1.1551,
      "theoretical_loss": 3.327252521801385,
      "tokens_seen": 3273129984
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.304914744232698e-06,
      "loss": 1.2579,
      "theoretical_loss": 3.3272475329089946,
      "tokens_seen": 3273195520
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.284854563691072e-06,
      "loss": 1.2495,
      "theoretical_loss": 3.3272425441444584,
      "tokens_seen": 3273261056
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.26479438314945e-06,
      "loss": 1.2456,
      "theoretical_loss": 3.327237555507771,
      "tokens_seen": 3273326592
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.244734202607823e-06,
      "loss": 1.3888,
      "theoretical_loss": 3.327232566998927,
      "tokens_seen": 3273392128
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.2246740220662e-06,
      "loss": 1.1258,
      "theoretical_loss": 3.3272275786179195,
      "tokens_seen": 3273457664
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.204613841524573e-06,
      "loss": 1.3112,
      "theoretical_loss": 3.3272225903647437,
      "tokens_seen": 3273523200
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.184553660982949e-06,
      "loss": 1.2697,
      "theoretical_loss": 3.3272176022393936,
      "tokens_seen": 3273588736
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.164493480441325e-06,
      "loss": 1.2033,
      "theoretical_loss": 3.327212614241863,
      "tokens_seen": 3273654272
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.144433299899699e-06,
      "loss": 1.2083,
      "theoretical_loss": 3.3272076263721457,
      "tokens_seen": 3273719808
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.124373119358075e-06,
      "loss": 1.3542,
      "theoretical_loss": 3.327202638630237,
      "tokens_seen": 3273785344
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.104312938816449e-06,
      "loss": 1.3098,
      "theoretical_loss": 3.32719765101613,
      "tokens_seen": 3273850880
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.084252758274825e-06,
      "loss": 1.2142,
      "theoretical_loss": 3.3271926635298197,
      "tokens_seen": 3273916416
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.064192577733199e-06,
      "loss": 1.3258,
      "theoretical_loss": 3.3271876761712993,
      "tokens_seen": 3273981952
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.044132397191576e-06,
      "loss": 1.2625,
      "theoretical_loss": 3.3271826889405642,
      "tokens_seen": 3274047488
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.02407221664995e-06,
      "loss": 1.1777,
      "theoretical_loss": 3.3271777018376074,
      "tokens_seen": 3274113024
    },
    {
      "epoch": 0.98,
      "learning_rate": 8.004012036108324e-06,
      "loss": 1.1308,
      "theoretical_loss": 3.327172714862424,
      "tokens_seen": 3274178560
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.9839518555667e-06,
      "loss": 1.221,
      "theoretical_loss": 3.3271677280150076,
      "tokens_seen": 3274244096
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.963891675025074e-06,
      "loss": 1.2728,
      "theoretical_loss": 3.327162741295352,
      "tokens_seen": 3274309632
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.943831494483452e-06,
      "loss": 1.346,
      "theoretical_loss": 3.3271577547034523,
      "tokens_seen": 3274375168
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.923771313941826e-06,
      "loss": 1.286,
      "theoretical_loss": 3.3271527682393023,
      "tokens_seen": 3274440704
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.903711133400202e-06,
      "loss": 1.1868,
      "theoretical_loss": 3.3271477819028963,
      "tokens_seen": 3274506240
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.883650952858576e-06,
      "loss": 1.2461,
      "theoretical_loss": 3.327142795694228,
      "tokens_seen": 3274571776
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.86359077231695e-06,
      "loss": 1.2521,
      "theoretical_loss": 3.327137809613292,
      "tokens_seen": 3274637312
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.843530591775326e-06,
      "loss": 1.2259,
      "theoretical_loss": 3.327132823660082,
      "tokens_seen": 3274702848
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.823470411233702e-06,
      "loss": 1.1532,
      "theoretical_loss": 3.3271278378345928,
      "tokens_seen": 3274768384
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.803410230692076e-06,
      "loss": 1.2058,
      "theoretical_loss": 3.3271228521368186,
      "tokens_seen": 3274833920
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.783350050150451e-06,
      "loss": 1.3265,
      "theoretical_loss": 3.327117866566753,
      "tokens_seen": 3274899456
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.763289869608827e-06,
      "loss": 1.2839,
      "theoretical_loss": 3.32711288112439,
      "tokens_seen": 3274964992
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.743229689067201e-06,
      "loss": 1.2782,
      "theoretical_loss": 3.327107895809725,
      "tokens_seen": 3275030528
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.723169508525577e-06,
      "loss": 1.3296,
      "theoretical_loss": 3.3271029106227505,
      "tokens_seen": 3275096064
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.703109327983953e-06,
      "loss": 1.2955,
      "theoretical_loss": 3.327097925563462,
      "tokens_seen": 3275161600
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.683049147442327e-06,
      "loss": 1.1839,
      "theoretical_loss": 3.3270929406318532,
      "tokens_seen": 3275227136
    },
    {
      "epoch": 0.98,
      "objective/train/docs_used": 1189805,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2356234788894653,
      "objective/train/theoretical_loss": 3.3270879558279183,
      "objective/train/tokens_used": 1645751776,
      "theoretical_loss": 3.3270879558279183,
      "tokens_seen": 3275292672
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.662988966900701e-06,
      "loss": 1.2169,
      "theoretical_loss": 3.3270879558279183,
      "tokens_seen": 3275292672
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.642928786359077e-06,
      "loss": 1.1747,
      "theoretical_loss": 3.3270829711516514,
      "tokens_seen": 3275358208
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.622868605817453e-06,
      "loss": 1.3522,
      "theoretical_loss": 3.327077986603047,
      "tokens_seen": 3275423744
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.602808425275828e-06,
      "loss": 1.1514,
      "theoretical_loss": 3.3270730021820984,
      "tokens_seen": 3275489280
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.582748244734203e-06,
      "loss": 1.3263,
      "theoretical_loss": 3.3270680178888012,
      "tokens_seen": 3275554816
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.562688064192578e-06,
      "loss": 1.2838,
      "theoretical_loss": 3.3270630337231486,
      "tokens_seen": 3275620352
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.542627883650953e-06,
      "loss": 1.1717,
      "theoretical_loss": 3.3270580496851347,
      "tokens_seen": 3275685888
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.522567703109327e-06,
      "loss": 1.2191,
      "theoretical_loss": 3.327053065774754,
      "tokens_seen": 3275751424
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.502507522567703e-06,
      "loss": 1.2855,
      "theoretical_loss": 3.3270480819920003,
      "tokens_seen": 3275816960
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.482447342026078e-06,
      "loss": 1.2846,
      "theoretical_loss": 3.3270430983368686,
      "tokens_seen": 3275882496
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.462387161484454e-06,
      "loss": 1.2448,
      "theoretical_loss": 3.327038114809352,
      "tokens_seen": 3275948032
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.442326980942829e-06,
      "loss": 1.2377,
      "theoretical_loss": 3.327033131409446,
      "tokens_seen": 3276013568
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.422266800401204e-06,
      "loss": 1.3179,
      "theoretical_loss": 3.3270281481371438,
      "tokens_seen": 3276079104
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.402206619859579e-06,
      "loss": 1.1789,
      "theoretical_loss": 3.3270231649924398,
      "tokens_seen": 3276144640
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.382146439317954e-06,
      "loss": 1.2516,
      "theoretical_loss": 3.327018181975328,
      "tokens_seen": 3276210176
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.362086258776329e-06,
      "loss": 1.4164,
      "theoretical_loss": 3.3270131990858025,
      "tokens_seen": 3276275712
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.3420260782347045e-06,
      "loss": 1.2033,
      "theoretical_loss": 3.3270082163238586,
      "tokens_seen": 3276341248
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.3219658976930794e-06,
      "loss": 1.2229,
      "theoretical_loss": 3.327003233689489,
      "tokens_seen": 3276406784
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.301905717151454e-06,
      "loss": 0.9995,
      "theoretical_loss": 3.3269982511826885,
      "tokens_seen": 3276472320
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.28184553660983e-06,
      "loss": 1.1979,
      "theoretical_loss": 3.3269932688034514,
      "tokens_seen": 3276537856
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.261785356068204e-06,
      "loss": 1.3219,
      "theoretical_loss": 3.3269882865517717,
      "tokens_seen": 3276603392
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.241725175526579e-06,
      "loss": 1.1667,
      "theoretical_loss": 3.326983304427644,
      "tokens_seen": 3276668928
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.221664994984955e-06,
      "loss": 1.2598,
      "theoretical_loss": 3.326978322431062,
      "tokens_seen": 3276734464
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.20160481444333e-06,
      "loss": 1.1848,
      "theoretical_loss": 3.3269733405620197,
      "tokens_seen": 3276800000
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.181544633901706e-06,
      "loss": 1.166,
      "theoretical_loss": 3.326968358820512,
      "tokens_seen": 3276865536
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.161484453360081e-06,
      "loss": 1.1907,
      "theoretical_loss": 3.326963377206533,
      "tokens_seen": 3276931072
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.141424272818456e-06,
      "loss": 1.2616,
      "theoretical_loss": 3.3269583957200757,
      "tokens_seen": 3276996608
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.121364092276831e-06,
      "loss": 1.2074,
      "theoretical_loss": 3.3269534143611357,
      "tokens_seen": 3277062144
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.1013039117352055e-06,
      "loss": 1.2233,
      "theoretical_loss": 3.3269484331297066,
      "tokens_seen": 3277127680
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.0812437311935805e-06,
      "loss": 1.2435,
      "theoretical_loss": 3.3269434520257826,
      "tokens_seen": 3277193216
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.061183550651956e-06,
      "loss": 1.2749,
      "theoretical_loss": 3.326938471049358,
      "tokens_seen": 3277258752
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.041123370110331e-06,
      "loss": 1.2484,
      "theoretical_loss": 3.326933490200427,
      "tokens_seen": 3277324288
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.021063189568706e-06,
      "loss": 1.1987,
      "theoretical_loss": 3.326928509478983,
      "tokens_seen": 3277389824
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.001003009027082e-06,
      "loss": 1.1654,
      "theoretical_loss": 3.326923528885022,
      "tokens_seen": 3277455360
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.980942828485456e-06,
      "loss": 1.3083,
      "theoretical_loss": 3.3269185484185364,
      "tokens_seen": 3277520896
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.960882647943831e-06,
      "loss": 1.1691,
      "theoretical_loss": 3.326913568079521,
      "tokens_seen": 3277586432
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.940822467402207e-06,
      "loss": 1.2001,
      "theoretical_loss": 3.3269085878679703,
      "tokens_seen": 3277651968
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.920762286860582e-06,
      "loss": 1.1655,
      "theoretical_loss": 3.326903607783878,
      "tokens_seen": 3277717504
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.9007021063189575e-06,
      "loss": 1.2866,
      "theoretical_loss": 3.326898627827239,
      "tokens_seen": 3277783040
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.8806419257773325e-06,
      "loss": 1.2919,
      "theoretical_loss": 3.326893647998047,
      "tokens_seen": 3277848576
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.860581745235707e-06,
      "loss": 1.2182,
      "theoretical_loss": 3.326888668296296,
      "tokens_seen": 3277914112
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.840521564694082e-06,
      "loss": 1.2619,
      "theoretical_loss": 3.3268836887219804,
      "tokens_seen": 3277979648
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.820461384152457e-06,
      "loss": 1.298,
      "theoretical_loss": 3.326878709275094,
      "tokens_seen": 3278045184
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.800401203610832e-06,
      "loss": 1.25,
      "theoretical_loss": 3.326873729955632,
      "tokens_seen": 3278110720
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.780341023069208e-06,
      "loss": 1.2109,
      "theoretical_loss": 3.3268687507635875,
      "tokens_seen": 3278176256
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.760280842527583e-06,
      "loss": 1.2643,
      "theoretical_loss": 3.3268637716989558,
      "tokens_seen": 3278241792
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.740220661985958e-06,
      "loss": 1.2653,
      "theoretical_loss": 3.3268587927617297,
      "tokens_seen": 3278307328
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.720160481444334e-06,
      "loss": 1.2195,
      "theoretical_loss": 3.326853813951905,
      "tokens_seen": 3278372864
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.700100300902708e-06,
      "loss": 1.2569,
      "theoretical_loss": 3.326848835269474,
      "tokens_seen": 3278438400
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.680040120361083e-06,
      "loss": 1.2175,
      "theoretical_loss": 3.326843856714433,
      "tokens_seen": 3278503936
    },
    {
      "epoch": 0.99,
      "objective/train/docs_used": 1191058,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.281537413597107,
      "objective/train/theoretical_loss": 3.3268388782867744,
      "objective/train/tokens_used": 1649028576,
      "theoretical_loss": 3.3268388782867744,
      "tokens_seen": 3278569472
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.659979939819459e-06,
      "loss": 1.2865,
      "theoretical_loss": 3.3268388782867744,
      "tokens_seen": 3278569472
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.6399197592778335e-06,
      "loss": 1.3637,
      "theoretical_loss": 3.3268338999864935,
      "tokens_seen": 3278635008
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.619859578736209e-06,
      "loss": 1.1381,
      "theoretical_loss": 3.326828921813584,
      "tokens_seen": 3278700544
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.599799398194584e-06,
      "loss": 1.2147,
      "theoretical_loss": 3.3268239437680402,
      "tokens_seen": 3278766080
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.579739217652959e-06,
      "loss": 1.2252,
      "theoretical_loss": 3.3268189658498564,
      "tokens_seen": 3278831616
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.559679037111334e-06,
      "loss": 1.3043,
      "theoretical_loss": 3.326813988059027,
      "tokens_seen": 3278897152
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.539618856569709e-06,
      "loss": 1.1832,
      "theoretical_loss": 3.3268090103955457,
      "tokens_seen": 3278962688
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.519558676028084e-06,
      "loss": 1.232,
      "theoretical_loss": 3.326804032859407,
      "tokens_seen": 3279028224
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.49949849548646e-06,
      "loss": 1.2406,
      "theoretical_loss": 3.3267990554506044,
      "tokens_seen": 3279093760
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.479438314944835e-06,
      "loss": 1.3241,
      "theoretical_loss": 3.326794078169133,
      "tokens_seen": 3279159296
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.45937813440321e-06,
      "loss": 1.286,
      "theoretical_loss": 3.3267891010149873,
      "tokens_seen": 3279224832
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.4393179538615855e-06,
      "loss": 1.3377,
      "theoretical_loss": 3.32678412398816,
      "tokens_seen": 3279290368
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.41925777331996e-06,
      "loss": 1.2111,
      "theoretical_loss": 3.3267791470886467,
      "tokens_seen": 3279355904
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.3991975927783346e-06,
      "loss": 1.2186,
      "theoretical_loss": 3.3267741703164413,
      "tokens_seen": 3279421440
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.37913741223671e-06,
      "loss": 1.2705,
      "theoretical_loss": 3.326769193671537,
      "tokens_seen": 3279486976
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.359077231695085e-06,
      "loss": 1.153,
      "theoretical_loss": 3.32676421715393,
      "tokens_seen": 3279552512
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.33901705115346e-06,
      "loss": 1.1856,
      "theoretical_loss": 3.326759240763612,
      "tokens_seen": 3279618048
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.318956870611836e-06,
      "loss": 1.1462,
      "theoretical_loss": 3.3267542645005794,
      "tokens_seen": 3279683584
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.298896690070211e-06,
      "loss": 1.1243,
      "theoretical_loss": 3.3267492883648253,
      "tokens_seen": 3279749120
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.278836509528586e-06,
      "loss": 1.2597,
      "theoretical_loss": 3.326744312356344,
      "tokens_seen": 3279814656
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.258776328986961e-06,
      "loss": 1.4619,
      "theoretical_loss": 3.32673933647513,
      "tokens_seen": 3279880192
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.238716148445336e-06,
      "loss": 1.2687,
      "theoretical_loss": 3.326734360721177,
      "tokens_seen": 3279945728
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.218655967903712e-06,
      "loss": 1.1921,
      "theoretical_loss": 3.3267293850944797,
      "tokens_seen": 3280011264
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.1985957873620866e-06,
      "loss": 1.3959,
      "theoretical_loss": 3.326724409595032,
      "tokens_seen": 3280076800
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.1785356068204615e-06,
      "loss": 1.2609,
      "theoretical_loss": 3.326719434222828,
      "tokens_seen": 3280142336
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.158475426278837e-06,
      "loss": 1.3338,
      "theoretical_loss": 3.3267144589778628,
      "tokens_seen": 3280207872
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.138415245737211e-06,
      "loss": 1.2287,
      "theoretical_loss": 3.326709483860129,
      "tokens_seen": 3280273408
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.118355065195586e-06,
      "loss": 1.1483,
      "theoretical_loss": 3.3267045088696223,
      "tokens_seen": 3280338944
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.098294884653962e-06,
      "loss": 1.1787,
      "theoretical_loss": 3.3266995340063366,
      "tokens_seen": 3280404480
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.078234704112337e-06,
      "loss": 1.3678,
      "theoretical_loss": 3.3266945592702655,
      "tokens_seen": 3280470016
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.058174523570712e-06,
      "loss": 1.2575,
      "theoretical_loss": 3.3266895846614033,
      "tokens_seen": 3280535552
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.038114343029088e-06,
      "loss": 1.1839,
      "theoretical_loss": 3.3266846101797447,
      "tokens_seen": 3280601088
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.018054162487462e-06,
      "loss": 1.2354,
      "theoretical_loss": 3.3266796358252835,
      "tokens_seen": 3280666624
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.997993981945838e-06,
      "loss": 1.2435,
      "theoretical_loss": 3.3266746615980143,
      "tokens_seen": 3280732160
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.977933801404213e-06,
      "loss": 1.2373,
      "theoretical_loss": 3.3266696874979313,
      "tokens_seen": 3280797696
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.957873620862588e-06,
      "loss": 1.2502,
      "theoretical_loss": 3.326664713525028,
      "tokens_seen": 3280863232
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.937813440320963e-06,
      "loss": 1.0898,
      "theoretical_loss": 3.326659739679299,
      "tokens_seen": 3280928768
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.917753259779338e-06,
      "loss": 1.2984,
      "theoretical_loss": 3.326654765960739,
      "tokens_seen": 3280994304
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.897693079237713e-06,
      "loss": 1.2581,
      "theoretical_loss": 3.3266497923693414,
      "tokens_seen": 3281059840
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.877632898696088e-06,
      "loss": 1.2397,
      "theoretical_loss": 3.326644818905101,
      "tokens_seen": 3281125376
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.857572718154463e-06,
      "loss": 1.3176,
      "theoretical_loss": 3.326639845568012,
      "tokens_seen": 3281190912
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.837512537612838e-06,
      "loss": 1.1153,
      "theoretical_loss": 3.3266348723580683,
      "tokens_seen": 3281256448
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.817452357071214e-06,
      "loss": 1.2397,
      "theoretical_loss": 3.326629899275264,
      "tokens_seen": 3281321984
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.797392176529589e-06,
      "loss": 1.2683,
      "theoretical_loss": 3.326624926319594,
      "tokens_seen": 3281387520
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.777331995987964e-06,
      "loss": 1.2463,
      "theoretical_loss": 3.326619953491052,
      "tokens_seen": 3281453056
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.75727181544634e-06,
      "loss": 1.2091,
      "theoretical_loss": 3.326614980789632,
      "tokens_seen": 3281518592
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.737211634904714e-06,
      "loss": 1.245,
      "theoretical_loss": 3.3266100082153285,
      "tokens_seen": 3281584128
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.7171514543630895e-06,
      "loss": 1.283,
      "theoretical_loss": 3.326605035768136,
      "tokens_seen": 3281649664
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.6970912738214645e-06,
      "loss": 1.2049,
      "theoretical_loss": 3.3266000634480477,
      "tokens_seen": 3281715200
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.677031093279839e-06,
      "loss": 1.2097,
      "theoretical_loss": 3.326595091255059,
      "tokens_seen": 3281780736
    },
    {
      "epoch": 0.99,
      "objective/train/docs_used": 1192226,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.3532216548919678,
      "objective/train/theoretical_loss": 3.326590119189164,
      "objective/train/tokens_used": 1652305376,
      "theoretical_loss": 3.326590119189164,
      "tokens_seen": 3281846272
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.656970912738215e-06,
      "loss": 1.2546,
      "theoretical_loss": 3.326590119189164,
      "tokens_seen": 3281846272
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.63691073219659e-06,
      "loss": 1.2939,
      "theoretical_loss": 3.3265851472503556,
      "tokens_seen": 3281911808
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.616850551654965e-06,
      "loss": 1.3417,
      "theoretical_loss": 3.3265801754386297,
      "tokens_seen": 3281977344
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.59679037111334e-06,
      "loss": 1.2101,
      "theoretical_loss": 3.3265752037539795,
      "tokens_seen": 3282042880
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.576730190571715e-06,
      "loss": 1.2092,
      "theoretical_loss": 3.3265702321963997,
      "tokens_seen": 3282108416
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.55667001003009e-06,
      "loss": 1.2494,
      "theoretical_loss": 3.326565260765884,
      "tokens_seen": 3282173952
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.536609829488466e-06,
      "loss": 1.2048,
      "theoretical_loss": 3.326560289462427,
      "tokens_seen": 3282239488
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.516549648946841e-06,
      "loss": 1.2325,
      "theoretical_loss": 3.3265553182860232,
      "tokens_seen": 3282305024
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.496489468405216e-06,
      "loss": 1.245,
      "theoretical_loss": 3.326550347236666,
      "tokens_seen": 3282370560
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.476429287863591e-06,
      "loss": 1.1769,
      "theoretical_loss": 3.32654537631435,
      "tokens_seen": 3282436096
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.4563691073219655e-06,
      "loss": 1.3459,
      "theoretical_loss": 3.3265404055190704,
      "tokens_seen": 3282501632
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.436308926780341e-06,
      "loss": 1.1513,
      "theoretical_loss": 3.3265354348508196,
      "tokens_seen": 3282567168
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.416248746238716e-06,
      "loss": 1.2469,
      "theoretical_loss": 3.326530464309593,
      "tokens_seen": 3282632704
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.396188565697091e-06,
      "loss": 1.2873,
      "theoretical_loss": 3.3265254938953843,
      "tokens_seen": 3282698240
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.376128385155467e-06,
      "loss": 1.2867,
      "theoretical_loss": 3.3265205236081883,
      "tokens_seen": 3282763776
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.356068204613842e-06,
      "loss": 1.2819,
      "theoretical_loss": 3.3265155534479987,
      "tokens_seen": 3282829312
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.336008024072217e-06,
      "loss": 1.2396,
      "theoretical_loss": 3.3265105834148097,
      "tokens_seen": 3282894848
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.315947843530592e-06,
      "loss": 1.3211,
      "theoretical_loss": 3.326505613508616,
      "tokens_seen": 3282960384
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.295887662988967e-06,
      "loss": 1.2013,
      "theoretical_loss": 3.3265006437294113,
      "tokens_seen": 3283025920
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.275827482447342e-06,
      "loss": 1.2709,
      "theoretical_loss": 3.3264956740771905,
      "tokens_seen": 3283091456
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.2557673019057175e-06,
      "loss": 1.2177,
      "theoretical_loss": 3.326490704551947,
      "tokens_seen": 3283156992
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.2357071213640925e-06,
      "loss": 1.2369,
      "theoretical_loss": 3.3264857351536756,
      "tokens_seen": 3283222528
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.215646940822467e-06,
      "loss": 1.2898,
      "theoretical_loss": 3.32648076588237,
      "tokens_seen": 3283288064
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.195586760280843e-06,
      "loss": 1.1685,
      "theoretical_loss": 3.326475796738025,
      "tokens_seen": 3283353600
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.175526579739217e-06,
      "loss": 1.2923,
      "theoretical_loss": 3.3264708277206347,
      "tokens_seen": 3283419136
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.155466399197592e-06,
      "loss": 1.3577,
      "theoretical_loss": 3.326465858830193,
      "tokens_seen": 3283484672
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.135406218655968e-06,
      "loss": 1.1845,
      "theoretical_loss": 3.326460890066694,
      "tokens_seen": 3283550208
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.115346038114343e-06,
      "loss": 1.1306,
      "theoretical_loss": 3.3264559214301324,
      "tokens_seen": 3283615744
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.095285857572719e-06,
      "loss": 1.26,
      "theoretical_loss": 3.3264509529205024,
      "tokens_seen": 3283681280
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.075225677031094e-06,
      "loss": 1.2087,
      "theoretical_loss": 3.3264459845377985,
      "tokens_seen": 3283746816
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.055165496489469e-06,
      "loss": 1.2744,
      "theoretical_loss": 3.326441016282014,
      "tokens_seen": 3283812352
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.035105315947844e-06,
      "loss": 1.2052,
      "theoretical_loss": 3.3264360481531434,
      "tokens_seen": 3283877888
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.0150451354062186e-06,
      "loss": 1.2628,
      "theoretical_loss": 3.3264310801511816,
      "tokens_seen": 3283943424
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.9949849548645935e-06,
      "loss": 1.1248,
      "theoretical_loss": 3.3264261122761223,
      "tokens_seen": 3284008960
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.974924774322969e-06,
      "loss": 1.145,
      "theoretical_loss": 3.3264211445279592,
      "tokens_seen": 3284074496
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.954864593781344e-06,
      "loss": 1.1055,
      "theoretical_loss": 3.326416176906688,
      "tokens_seen": 3284140032
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.934804413239719e-06,
      "loss": 1.093,
      "theoretical_loss": 3.3264112094123015,
      "tokens_seen": 3284205568
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.914744232698095e-06,
      "loss": 1.197,
      "theoretical_loss": 3.326406242044795,
      "tokens_seen": 3284271104
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.894684052156469e-06,
      "loss": 1.161,
      "theoretical_loss": 3.326401274804162,
      "tokens_seen": 3284336640
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.874623871614844e-06,
      "loss": 1.1979,
      "theoretical_loss": 3.3263963076903966,
      "tokens_seen": 3284402176
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.85456369107322e-06,
      "loss": 1.1886,
      "theoretical_loss": 3.3263913407034935,
      "tokens_seen": 3284467712
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.834503510531595e-06,
      "loss": 1.2801,
      "theoretical_loss": 3.326386373843447,
      "tokens_seen": 3284533248
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.8144433299899706e-06,
      "loss": 1.238,
      "theoretical_loss": 3.326381407110251,
      "tokens_seen": 3284598784
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.7943831494483455e-06,
      "loss": 1.2238,
      "theoretical_loss": 3.3263764405039,
      "tokens_seen": 3284664320
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.77432296890672e-06,
      "loss": 1.3227,
      "theoretical_loss": 3.3263714740243877,
      "tokens_seen": 3284729856
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.754262788365095e-06,
      "loss": 1.1752,
      "theoretical_loss": 3.3263665076717093,
      "tokens_seen": 3284795392
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.73420260782347e-06,
      "loss": 1.3283,
      "theoretical_loss": 3.326361541445858,
      "tokens_seen": 3284860928
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.714142427281845e-06,
      "loss": 1.1776,
      "theoretical_loss": 3.3263565753468285,
      "tokens_seen": 3284926464
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.694082246740221e-06,
      "loss": 1.0673,
      "theoretical_loss": 3.3263516093746155,
      "tokens_seen": 3284992000
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.674022066198596e-06,
      "loss": 1.2555,
      "theoretical_loss": 3.326346643529212,
      "tokens_seen": 3285057536
    },
    {
      "epoch": 0.99,
      "objective/train/docs_used": 1193279,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.219916582107544,
      "objective/train/theoretical_loss": 3.3263416778106136,
      "objective/train/tokens_used": 1655582176,
      "theoretical_loss": 3.3263416778106136,
      "tokens_seen": 3285123072
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.653961885656971e-06,
      "loss": 1.2843,
      "theoretical_loss": 3.3263416778106136,
      "tokens_seen": 3285123072
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.633901705115346e-06,
      "loss": 1.1641,
      "theoretical_loss": 3.3263367122188137,
      "tokens_seen": 3285188608
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.613841524573721e-06,
      "loss": 1.213,
      "theoretical_loss": 3.3263317467538065,
      "tokens_seen": 3285254144
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.593781344032096e-06,
      "loss": 1.205,
      "theoretical_loss": 3.3263267814155864,
      "tokens_seen": 3285319680
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.573721163490472e-06,
      "loss": 1.3009,
      "theoretical_loss": 3.3263218162041484,
      "tokens_seen": 3285385216
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.5536609829488466e-06,
      "loss": 1.2798,
      "theoretical_loss": 3.3263168511194854,
      "tokens_seen": 3285450752
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.533600802407222e-06,
      "loss": 1.2068,
      "theoretical_loss": 3.3263118861615926,
      "tokens_seen": 3285516288
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.513540621865597e-06,
      "loss": 1.3391,
      "theoretical_loss": 3.3263069213304637,
      "tokens_seen": 3285581824
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.493480441323971e-06,
      "loss": 1.2679,
      "theoretical_loss": 3.326301956626093,
      "tokens_seen": 3285647360
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.473420260782347e-06,
      "loss": 1.2227,
      "theoretical_loss": 3.326296992048475,
      "tokens_seen": 3285712896
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.453360080240722e-06,
      "loss": 1.2722,
      "theoretical_loss": 3.326292027597604,
      "tokens_seen": 3285778432
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.433299899699097e-06,
      "loss": 1.2676,
      "theoretical_loss": 3.3262870632734742,
      "tokens_seen": 3285843968
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.413239719157473e-06,
      "loss": 1.2457,
      "theoretical_loss": 3.3262820990760793,
      "tokens_seen": 3285909504
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.393179538615848e-06,
      "loss": 1.2175,
      "theoretical_loss": 3.326277135005414,
      "tokens_seen": 3285975040
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.373119358074223e-06,
      "loss": 1.1674,
      "theoretical_loss": 3.3262721710614724,
      "tokens_seen": 3286040576
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.353059177532598e-06,
      "loss": 1.1571,
      "theoretical_loss": 3.326267207244249,
      "tokens_seen": 3286106112
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.332998996990973e-06,
      "loss": 1.2456,
      "theoretical_loss": 3.326262243553738,
      "tokens_seen": 3286171648
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.312938816449348e-06,
      "loss": 1.2905,
      "theoretical_loss": 3.326257279989933,
      "tokens_seen": 3286237184
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.292878635907723e-06,
      "loss": 1.2521,
      "theoretical_loss": 3.326252316552829,
      "tokens_seen": 3286302720
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.272818455366098e-06,
      "loss": 1.2651,
      "theoretical_loss": 3.3262473532424197,
      "tokens_seen": 3286368256
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.252758274824474e-06,
      "loss": 1.1995,
      "theoretical_loss": 3.3262423900586997,
      "tokens_seen": 3286433792
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.232698094282849e-06,
      "loss": 1.2482,
      "theoretical_loss": 3.3262374270016632,
      "tokens_seen": 3286499328
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.212637913741223e-06,
      "loss": 1.185,
      "theoretical_loss": 3.3262324640713046,
      "tokens_seen": 3286564864
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.192577733199599e-06,
      "loss": 1.247,
      "theoretical_loss": 3.326227501267618,
      "tokens_seen": 3286630400
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.172517552657974e-06,
      "loss": 1.2239,
      "theoretical_loss": 3.326222538590597,
      "tokens_seen": 3286695936
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.152457372116349e-06,
      "loss": 1.1768,
      "theoretical_loss": 3.3262175760402366,
      "tokens_seen": 3286761472
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.132397191574725e-06,
      "loss": 1.2245,
      "theoretical_loss": 3.326212613616531,
      "tokens_seen": 3286827008
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.1123370110331e-06,
      "loss": 1.3168,
      "theoretical_loss": 3.326207651319474,
      "tokens_seen": 3286892544
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.0922768304914746e-06,
      "loss": 1.301,
      "theoretical_loss": 3.3262026891490604,
      "tokens_seen": 3286958080
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.0722166499498495e-06,
      "loss": 1.1921,
      "theoretical_loss": 3.326197727105284,
      "tokens_seen": 3287023616
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.0521564694082245e-06,
      "loss": 1.2488,
      "theoretical_loss": 3.3261927651881393,
      "tokens_seen": 3287089152
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.032096288866599e-06,
      "loss": 1.2205,
      "theoretical_loss": 3.3261878033976204,
      "tokens_seen": 3287154688
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.012036108324975e-06,
      "loss": 1.2707,
      "theoretical_loss": 3.3261828417337216,
      "tokens_seen": 3287220224
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.99197592778335e-06,
      "loss": 1.1292,
      "theoretical_loss": 3.326177880196437,
      "tokens_seen": 3287285760
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.971915747241726e-06,
      "loss": 1.2631,
      "theoretical_loss": 3.3261729187857614,
      "tokens_seen": 3287351296
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.951855566700101e-06,
      "loss": 1.1814,
      "theoretical_loss": 3.3261679575016885,
      "tokens_seen": 3287416832
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.931795386158475e-06,
      "loss": 1.158,
      "theoretical_loss": 3.3261629963442125,
      "tokens_seen": 3287482368
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.911735205616851e-06,
      "loss": 1.1857,
      "theoretical_loss": 3.3261580353133278,
      "tokens_seen": 3287547904
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.891675025075226e-06,
      "loss": 1.2503,
      "theoretical_loss": 3.326153074409029,
      "tokens_seen": 3287613440
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.871614844533601e-06,
      "loss": 1.2186,
      "theoretical_loss": 3.3261481136313096,
      "tokens_seen": 3287678976
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.8515546639919765e-06,
      "loss": 1.1396,
      "theoretical_loss": 3.326143152980164,
      "tokens_seen": 3287744512
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.8314944834503506e-06,
      "loss": 1.1752,
      "theoretical_loss": 3.3261381924555877,
      "tokens_seen": 3287810048
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.8114343029087263e-06,
      "loss": 1.2233,
      "theoretical_loss": 3.326133232057573,
      "tokens_seen": 3287875584
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.7913741223671013e-06,
      "loss": 1.2527,
      "theoretical_loss": 3.3261282717861156,
      "tokens_seen": 3287941120
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.7713139418254767e-06,
      "loss": 1.1023,
      "theoretical_loss": 3.3261233116412092,
      "tokens_seen": 3288006656
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.7512537612838516e-06,
      "loss": 1.2622,
      "theoretical_loss": 3.3261183516228483,
      "tokens_seen": 3288072192
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.731193580742227e-06,
      "loss": 1.2262,
      "theoretical_loss": 3.3261133917310266,
      "tokens_seen": 3288137728
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.711133400200602e-06,
      "loss": 1.1073,
      "theoretical_loss": 3.326108431965739,
      "tokens_seen": 3288203264
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.691073219658977e-06,
      "loss": 1.232,
      "theoretical_loss": 3.326103472326979,
      "tokens_seen": 3288268800
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.6710130391173522e-06,
      "loss": 1.1887,
      "theoretical_loss": 3.326098512814742,
      "tokens_seen": 3288334336
    },
    {
      "debugging/Compilability": 1.0,
      "debugging/distinct-1-grams": 0.7477778926818737,
      "debugging/entropy-1-grams": 4.957593128654467,
      "debugging/length": 609.625,
      "debugging/num_segments": 8,
      "debugging/score": 0.006138877505802302,
      "debugging/score_std": 0.007441179575288503,
      "epoch": 0.99,
      "objective/train/docs_used": 1194491,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.221968650817871,
      "objective/train/theoretical_loss": 3.3260935534290206,
      "objective/train/tokens_used": 1658858976,
      "theoretical_loss": 3.3260935534290206,
      "tokens_seen": 3288399872
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.650952858575727e-06,
      "loss": 1.2151,
      "theoretical_loss": 3.3260935534290206,
      "tokens_seen": 3288399872
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.630892678034102e-06,
      "loss": 1.1775,
      "theoretical_loss": 3.3260885941698106,
      "tokens_seen": 3288465408
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.6108324974924775e-06,
      "loss": 1.2175,
      "theoretical_loss": 3.3260836350371057,
      "tokens_seen": 3288530944
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.590772316950853e-06,
      "loss": 1.2625,
      "theoretical_loss": 3.3260786760309,
      "tokens_seen": 3288596480
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.570712136409228e-06,
      "loss": 1.3513,
      "theoretical_loss": 3.326073717151188,
      "tokens_seen": 3288662016
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.5506519558676028e-06,
      "loss": 1.3692,
      "theoretical_loss": 3.326068758397964,
      "tokens_seen": 3288727552
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.530591775325978e-06,
      "loss": 1.329,
      "theoretical_loss": 3.3260637997712212,
      "tokens_seen": 3288793088
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.510531594784353e-06,
      "loss": 1.2477,
      "theoretical_loss": 3.3260588412709553,
      "tokens_seen": 3288858624
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.490471414242728e-06,
      "loss": 1.2413,
      "theoretical_loss": 3.32605388289716,
      "tokens_seen": 3288924160
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.4704112337011034e-06,
      "loss": 1.197,
      "theoretical_loss": 3.3260489246498297,
      "tokens_seen": 3288989696
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.4503510531594788e-06,
      "loss": 1.2549,
      "theoretical_loss": 3.326043966528958,
      "tokens_seen": 3289055232
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.4302908726178537e-06,
      "loss": 1.2034,
      "theoretical_loss": 3.32603900853454,
      "tokens_seen": 3289120768
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.4102306920762287e-06,
      "loss": 1.3714,
      "theoretical_loss": 3.3260340506665695,
      "tokens_seen": 3289186304
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.390170511534604e-06,
      "loss": 1.216,
      "theoretical_loss": 3.326029092925041,
      "tokens_seen": 3289251840
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.370110330992979e-06,
      "loss": 1.1607,
      "theoretical_loss": 3.3260241353099484,
      "tokens_seen": 3289317376
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.350050150451354e-06,
      "loss": 1.1728,
      "theoretical_loss": 3.326019177821286,
      "tokens_seen": 3289382912
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.3299899699097293e-06,
      "loss": 1.2671,
      "theoretical_loss": 3.3260142204590486,
      "tokens_seen": 3289448448
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.3099297893681047e-06,
      "loss": 1.1539,
      "theoretical_loss": 3.32600926322323,
      "tokens_seen": 3289513984
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.2898696088264796e-06,
      "loss": 1.2499,
      "theoretical_loss": 3.3260043061138247,
      "tokens_seen": 3289579520
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.2698094282848546e-06,
      "loss": 1.2284,
      "theoretical_loss": 3.3259993491308264,
      "tokens_seen": 3289645056
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.24974924774323e-06,
      "loss": 1.3379,
      "theoretical_loss": 3.32599439227423,
      "tokens_seen": 3289710592
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.229689067201605e-06,
      "loss": 1.1807,
      "theoretical_loss": 3.325989435544029,
      "tokens_seen": 3289776128
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.20962888665998e-06,
      "loss": 1.186,
      "theoretical_loss": 3.325984478940219,
      "tokens_seen": 3289841664
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.189568706118355e-06,
      "loss": 1.24,
      "theoretical_loss": 3.3259795224627933,
      "tokens_seen": 3289907200
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.16950852557673e-06,
      "loss": 1.256,
      "theoretical_loss": 3.325974566111746,
      "tokens_seen": 3289972736
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.1494483450351055e-06,
      "loss": 1.1997,
      "theoretical_loss": 3.3259696098870712,
      "tokens_seen": 3290038272
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.1293881644934804e-06,
      "loss": 1.2308,
      "theoretical_loss": 3.3259646537887644,
      "tokens_seen": 3290103808
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.109327983951856e-06,
      "loss": 1.3138,
      "theoretical_loss": 3.3259596978168187,
      "tokens_seen": 3290169344
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.0892678034102308e-06,
      "loss": 1.2015,
      "theoretical_loss": 3.3259547419712288,
      "tokens_seen": 3290234880
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.0692076228686057e-06,
      "loss": 1.2385,
      "theoretical_loss": 3.325949786251989,
      "tokens_seen": 3290300416
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.049147442326981e-06,
      "loss": 1.1651,
      "theoretical_loss": 3.3259448306590933,
      "tokens_seen": 3290365952
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.029087261785356e-06,
      "loss": 1.2099,
      "theoretical_loss": 3.325939875192536,
      "tokens_seen": 3290431488
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.009027081243731e-06,
      "loss": 1.2478,
      "theoretical_loss": 3.325934919852312,
      "tokens_seen": 3290497024
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.9889669007021063e-06,
      "loss": 1.2222,
      "theoretical_loss": 3.3259299646384144,
      "tokens_seen": 3290562560
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.9689067201604817e-06,
      "loss": 1.245,
      "theoretical_loss": 3.325925009550838,
      "tokens_seen": 3290628096
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.9488465396188567e-06,
      "loss": 1.2625,
      "theoretical_loss": 3.3259200545895777,
      "tokens_seen": 3290693632
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.9287863590772316e-06,
      "loss": 1.2822,
      "theoretical_loss": 3.3259150997546274,
      "tokens_seen": 3290759168
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.908726178535607e-06,
      "loss": 1.1807,
      "theoretical_loss": 3.325910145045981,
      "tokens_seen": 3290824704
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.888665997993982e-06,
      "loss": 1.2411,
      "theoretical_loss": 3.3259051904636325,
      "tokens_seen": 3290890240
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.868605817452357e-06,
      "loss": 1.2048,
      "theoretical_loss": 3.3259002360075773,
      "tokens_seen": 3290955776
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.8485456369107322e-06,
      "loss": 1.2314,
      "theoretical_loss": 3.3258952816778087,
      "tokens_seen": 3291021312
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.8284854563691076e-06,
      "loss": 1.1849,
      "theoretical_loss": 3.3258903274743212,
      "tokens_seen": 3291086848
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.8084252758274825e-06,
      "loss": 1.2415,
      "theoretical_loss": 3.325885373397109,
      "tokens_seen": 3291152384
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.7883650952858575e-06,
      "loss": 1.2716,
      "theoretical_loss": 3.325880419446167,
      "tokens_seen": 3291217920
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.768304914744233e-06,
      "loss": 1.1803,
      "theoretical_loss": 3.325875465621489,
      "tokens_seen": 3291283456
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.748244734202608e-06,
      "loss": 1.2979,
      "theoretical_loss": 3.3258705119230685,
      "tokens_seen": 3291348992
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.7281845536609828e-06,
      "loss": 1.2292,
      "theoretical_loss": 3.325865558350901,
      "tokens_seen": 3291414528
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.708124373119358e-06,
      "loss": 1.1924,
      "theoretical_loss": 3.32586060490498,
      "tokens_seen": 3291480064
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.6880641925777335e-06,
      "loss": 1.2563,
      "theoretical_loss": 3.3258556515853,
      "tokens_seen": 3291545600
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.6680040120361084e-06,
      "loss": 1.2346,
      "theoretical_loss": 3.325850698391856,
      "tokens_seen": 3291611136
    },
    {
      "epoch": 0.99,
      "objective/train/docs_used": 1195662,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2195073366165161,
      "objective/train/theoretical_loss": 3.325845745324641,
      "objective/train/tokens_used": 1662135776,
      "theoretical_loss": 3.325845745324641,
      "tokens_seen": 3291676672
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.6479438314944834e-06,
      "loss": 1.201,
      "theoretical_loss": 3.325845745324641,
      "tokens_seen": 3291676672
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.6278836509528588e-06,
      "loss": 1.1945,
      "theoretical_loss": 3.3258407923836497,
      "tokens_seen": 3291742208
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.6078234704112337e-06,
      "loss": 1.2287,
      "theoretical_loss": 3.325835839568877,
      "tokens_seen": 3291807744
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.5877632898696087e-06,
      "loss": 1.2077,
      "theoretical_loss": 3.3258308868803166,
      "tokens_seen": 3291873280
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.567703109327984e-06,
      "loss": 1.1069,
      "theoretical_loss": 3.3258259343179626,
      "tokens_seen": 3291938816
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.5476429287863594e-06,
      "loss": 1.2305,
      "theoretical_loss": 3.3258209818818094,
      "tokens_seen": 3292004352
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.5275827482447343e-06,
      "loss": 1.2144,
      "theoretical_loss": 3.3258160295718517,
      "tokens_seen": 3292069888
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.5075225677031093e-06,
      "loss": 1.2231,
      "theoretical_loss": 3.3258110773880833,
      "tokens_seen": 3292135424
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.4874623871614847e-06,
      "loss": 1.214,
      "theoretical_loss": 3.325806125330499,
      "tokens_seen": 3292200960
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.4674022066198596e-06,
      "loss": 1.3652,
      "theoretical_loss": 3.3258011733990926,
      "tokens_seen": 3292266496
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.4473420260782345e-06,
      "loss": 1.1702,
      "theoretical_loss": 3.3257962215938583,
      "tokens_seen": 3292332032
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.42728184553661e-06,
      "loss": 1.2422,
      "theoretical_loss": 3.3257912699147907,
      "tokens_seen": 3292397568
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.4072216649949853e-06,
      "loss": 1.247,
      "theoretical_loss": 3.325786318361884,
      "tokens_seen": 3292463104
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.38716148445336e-06,
      "loss": 1.4013,
      "theoretical_loss": 3.325781366935132,
      "tokens_seen": 3292528640
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.367101303911735e-06,
      "loss": 1.2293,
      "theoretical_loss": 3.32577641563453,
      "tokens_seen": 3292594176
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.3470411233701105e-06,
      "loss": 1.2132,
      "theoretical_loss": 3.3257714644600713,
      "tokens_seen": 3292659712
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.3269809428284855e-06,
      "loss": 1.1991,
      "theoretical_loss": 3.3257665134117507,
      "tokens_seen": 3292725248
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.3069207622868604e-06,
      "loss": 1.1438,
      "theoretical_loss": 3.3257615624895625,
      "tokens_seen": 3292790784
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.286860581745236e-06,
      "loss": 1.2986,
      "theoretical_loss": 3.3257566116935005,
      "tokens_seen": 3292856320
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.266800401203611e-06,
      "loss": 1.1599,
      "theoretical_loss": 3.3257516610235593,
      "tokens_seen": 3292921856
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.2467402206619857e-06,
      "loss": 1.2845,
      "theoretical_loss": 3.325746710479733,
      "tokens_seen": 3292987392
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.226680040120361e-06,
      "loss": 1.1509,
      "theoretical_loss": 3.3257417600620163,
      "tokens_seen": 3293052928
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.2066198595787364e-06,
      "loss": 1.0711,
      "theoretical_loss": 3.325736809770403,
      "tokens_seen": 3293118464
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.1865596790371114e-06,
      "loss": 1.2054,
      "theoretical_loss": 3.325731859604888,
      "tokens_seen": 3293184000
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.1664994984954863e-06,
      "loss": 1.2846,
      "theoretical_loss": 3.3257269095654647,
      "tokens_seen": 3293249536
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.1464393179538617e-06,
      "loss": 1.1699,
      "theoretical_loss": 3.325721959652128,
      "tokens_seen": 3293315072
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.126379137412237e-06,
      "loss": 1.1783,
      "theoretical_loss": 3.325717009864872,
      "tokens_seen": 3293380608
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.1063189568706116e-06,
      "loss": 1.2862,
      "theoretical_loss": 3.325712060203691,
      "tokens_seen": 3293446144
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.086258776328987e-06,
      "loss": 1.1791,
      "theoretical_loss": 3.325707110668579,
      "tokens_seen": 3293511680
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0661985957873623e-06,
      "loss": 1.2801,
      "theoretical_loss": 3.325702161259531,
      "tokens_seen": 3293577216
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0461384152457373e-06,
      "loss": 1.1293,
      "theoretical_loss": 3.325697211976541,
      "tokens_seen": 3293642752
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0260782347041122e-06,
      "loss": 1.3476,
      "theoretical_loss": 3.3256922628196026,
      "tokens_seen": 3293708288
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0060180541624876e-06,
      "loss": 1.2636,
      "theoretical_loss": 3.325687313788711,
      "tokens_seen": 3293773824
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.985957873620863e-06,
      "loss": 1.1937,
      "theoretical_loss": 3.32568236488386,
      "tokens_seen": 3293839360
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.9658976930792375e-06,
      "loss": 1.2964,
      "theoretical_loss": 3.325677416105044,
      "tokens_seen": 3293904896
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.945837512537613e-06,
      "loss": 1.1777,
      "theoretical_loss": 3.325672467452257,
      "tokens_seen": 3293970432
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.9257773319959882e-06,
      "loss": 1.1471,
      "theoretical_loss": 3.325667518925494,
      "tokens_seen": 3294035968
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.9057171514543632e-06,
      "loss": 1.1783,
      "theoretical_loss": 3.3256625705247482,
      "tokens_seen": 3294101504
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8856569709127383e-06,
      "loss": 1.2824,
      "theoretical_loss": 3.3256576222500147,
      "tokens_seen": 3294167040
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8655967903711135e-06,
      "loss": 1.233,
      "theoretical_loss": 3.3256526741012875,
      "tokens_seen": 3294232576
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8455366098294884e-06,
      "loss": 1.2736,
      "theoretical_loss": 3.3256477260785613,
      "tokens_seen": 3294298112
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8254764292878636e-06,
      "loss": 1.2584,
      "theoretical_loss": 3.3256427781818303,
      "tokens_seen": 3294363648
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8054162487462388e-06,
      "loss": 1.322,
      "theoretical_loss": 3.325637830411088,
      "tokens_seen": 3294429184
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.785356068204614e-06,
      "loss": 1.2184,
      "theoretical_loss": 3.325632882766329,
      "tokens_seen": 3294494720
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.765295887662989e-06,
      "loss": 1.1413,
      "theoretical_loss": 3.3256279352475486,
      "tokens_seen": 3294560256
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.745235707121364e-06,
      "loss": 1.0844,
      "theoretical_loss": 3.32562298785474,
      "tokens_seen": 3294625792
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.7251755265797394e-06,
      "loss": 1.2721,
      "theoretical_loss": 3.3256180405878975,
      "tokens_seen": 3294691328
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.7051153460381143e-06,
      "loss": 1.2063,
      "theoretical_loss": 3.3256130934470156,
      "tokens_seen": 3294756864
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.6850551654964895e-06,
      "loss": 1.2552,
      "theoretical_loss": 3.325608146432089,
      "tokens_seen": 3294822400
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.6649949849548646e-06,
      "loss": 1.3633,
      "theoretical_loss": 3.3256031995431115,
      "tokens_seen": 3294887936
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 1196765,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.2364988327026367,
      "objective/train/theoretical_loss": 3.325598252780078,
      "objective/train/tokens_used": 1665412576,
      "theoretical_loss": 3.325598252780078,
      "tokens_seen": 3294953472
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.6449348044132398e-06,
      "loss": 1.1791,
      "theoretical_loss": 3.325598252780078,
      "tokens_seen": 3294953472
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.624874623871615e-06,
      "loss": 1.3507,
      "theoretical_loss": 3.3255933061429817,
      "tokens_seen": 3295019008
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.60481444332999e-06,
      "loss": 1.1827,
      "theoretical_loss": 3.325588359631818,
      "tokens_seen": 3295084544
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.584754262788365e-06,
      "loss": 1.093,
      "theoretical_loss": 3.32558341324658,
      "tokens_seen": 3295150080
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.5646940822467402e-06,
      "loss": 1.2426,
      "theoretical_loss": 3.3255784669872632,
      "tokens_seen": 3295215616
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.5446339017051154e-06,
      "loss": 1.212,
      "theoretical_loss": 3.3255735208538613,
      "tokens_seen": 3295281152
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.5245737211634905e-06,
      "loss": 1.2461,
      "theoretical_loss": 3.325568574846369,
      "tokens_seen": 3295346688
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.5045135406218655e-06,
      "loss": 1.2043,
      "theoretical_loss": 3.3255636289647796,
      "tokens_seen": 3295412224
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.4844533600802409e-06,
      "loss": 1.2506,
      "theoretical_loss": 3.3255586832090884,
      "tokens_seen": 3295477760
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.4643931795386158e-06,
      "loss": 1.188,
      "theoretical_loss": 3.3255537375792894,
      "tokens_seen": 3295543296
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.444332998996991e-06,
      "loss": 1.2367,
      "theoretical_loss": 3.3255487920753772,
      "tokens_seen": 3295608832
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.4242728184553661e-06,
      "loss": 1.1824,
      "theoretical_loss": 3.3255438466973453,
      "tokens_seen": 3295674368
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.4042126379137413e-06,
      "loss": 1.2212,
      "theoretical_loss": 3.325538901445188,
      "tokens_seen": 3295739904
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3841524573721164e-06,
      "loss": 1.2779,
      "theoretical_loss": 3.3255339563189006,
      "tokens_seen": 3295805440
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3640922768304914e-06,
      "loss": 1.1365,
      "theoretical_loss": 3.325529011318477,
      "tokens_seen": 3295870976
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3440320962888667e-06,
      "loss": 1.2008,
      "theoretical_loss": 3.325524066443911,
      "tokens_seen": 3295936512
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3239719157472417e-06,
      "loss": 1.3011,
      "theoretical_loss": 3.325519121695197,
      "tokens_seen": 3296002048
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3039117352056169e-06,
      "loss": 1.3223,
      "theoretical_loss": 3.3255141770723298,
      "tokens_seen": 3296067584
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.283851554663992e-06,
      "loss": 1.1882,
      "theoretical_loss": 3.3255092325753033,
      "tokens_seen": 3296133120
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.2637913741223672e-06,
      "loss": 1.2385,
      "theoretical_loss": 3.3255042882041117,
      "tokens_seen": 3296198656
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.2437311935807423e-06,
      "loss": 1.2445,
      "theoretical_loss": 3.32549934395875,
      "tokens_seen": 3296264192
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.2236710130391173e-06,
      "loss": 1.2896,
      "theoretical_loss": 3.3254943998392115,
      "tokens_seen": 3296329728
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.2036108324974926e-06,
      "loss": 1.225,
      "theoretical_loss": 3.3254894558454913,
      "tokens_seen": 3296395264
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.1835506519558676e-06,
      "loss": 1.2776,
      "theoretical_loss": 3.3254845119775833,
      "tokens_seen": 3296460800
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.1634904714142427e-06,
      "loss": 1.2901,
      "theoretical_loss": 3.3254795682354814,
      "tokens_seen": 3296526336
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.143430290872618e-06,
      "loss": 1.2501,
      "theoretical_loss": 3.325474624619181,
      "tokens_seen": 3296591872
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.1233701103309928e-06,
      "loss": 1.2448,
      "theoretical_loss": 3.3254696811286752,
      "tokens_seen": 3296657408
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.1033099297893682e-06,
      "loss": 1.2308,
      "theoretical_loss": 3.3254647377639595,
      "tokens_seen": 3296722944
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0832497492477432e-06,
      "loss": 1.2936,
      "theoretical_loss": 3.325459794525027,
      "tokens_seen": 3296788480
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0631895687061185e-06,
      "loss": 1.2781,
      "theoretical_loss": 3.3254548514118727,
      "tokens_seen": 3296854016
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0431293881644935e-06,
      "loss": 1.283,
      "theoretical_loss": 3.325449908424491,
      "tokens_seen": 3296919552
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0230692076228686e-06,
      "loss": 1.2684,
      "theoretical_loss": 3.325444965562876,
      "tokens_seen": 3296985088
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0030090270812438e-06,
      "loss": 1.1707,
      "theoretical_loss": 3.3254400228270216,
      "tokens_seen": 3297050624
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.829488465396187e-07,
      "loss": 1.325,
      "theoretical_loss": 3.3254350802169226,
      "tokens_seen": 3297116160
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.628886659979941e-07,
      "loss": 1.1086,
      "theoretical_loss": 3.3254301377325732,
      "tokens_seen": 3297181696
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.428284854563692e-07,
      "loss": 1.2355,
      "theoretical_loss": 3.3254251953739677,
      "tokens_seen": 3297247232
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.227683049147442e-07,
      "loss": 1.2311,
      "theoretical_loss": 3.3254202531411003,
      "tokens_seen": 3297312768
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.027081243731194e-07,
      "loss": 1.1702,
      "theoretical_loss": 3.3254153110339653,
      "tokens_seen": 3297378304
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.826479438314945e-07,
      "loss": 1.3253,
      "theoretical_loss": 3.325410369052557,
      "tokens_seen": 3297443840
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.625877632898697e-07,
      "loss": 1.2886,
      "theoretical_loss": 3.32540542719687,
      "tokens_seen": 3297509376
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.425275827482447e-07,
      "loss": 1.1456,
      "theoretical_loss": 3.3254004854668984,
      "tokens_seen": 3297574912
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.224674022066199e-07,
      "loss": 1.1835,
      "theoretical_loss": 3.325395543862636,
      "tokens_seen": 3297640448
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.02407221664995e-07,
      "loss": 1.2055,
      "theoretical_loss": 3.3253906023840782,
      "tokens_seen": 3297705984
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.823470411233701e-07,
      "loss": 1.2654,
      "theoretical_loss": 3.3253856610312185,
      "tokens_seen": 3297771520
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.622868605817453e-07,
      "loss": 1.0963,
      "theoretical_loss": 3.3253807198040515,
      "tokens_seen": 3297837056
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.422266800401204e-07,
      "loss": 1.2342,
      "theoretical_loss": 3.325375778702571,
      "tokens_seen": 3297902592
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.221664994984955e-07,
      "loss": 1.2497,
      "theoretical_loss": 3.325370837726772,
      "tokens_seen": 3297968128
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.021063189568706e-07,
      "loss": 1.2119,
      "theoretical_loss": 3.325365896876648,
      "tokens_seen": 3298033664
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.820461384152457e-07,
      "loss": 1.2671,
      "theoretical_loss": 3.3253609561521946,
      "tokens_seen": 3298099200
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.619859578736208e-07,
      "loss": 1.2433,
      "theoretical_loss": 3.325356015553405,
      "tokens_seen": 3298164736
    },
    {
      "epoch": 1.0,
      "objective/train/docs_used": 1197928,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 1.270293951034546,
      "objective/train/theoretical_loss": 3.325351075080274,
      "objective/train/tokens_used": 1668689376,
      "theoretical_loss": 3.325351075080274,
      "tokens_seen": 3298230272
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.41925777331996e-07,
      "loss": 1.2633,
      "theoretical_loss": 3.325351075080274,
      "tokens_seen": 3298230272
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.218655967903712e-07,
      "loss": 1.1949,
      "theoretical_loss": 3.3253461347327953,
      "tokens_seen": 3298295808
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.018054162487463e-07,
      "loss": 1.1414,
      "theoretical_loss": 3.325341194510964,
      "tokens_seen": 3298361344
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.817452357071214e-07,
      "loss": 1.2067,
      "theoretical_loss": 3.3253362544147738,
      "tokens_seen": 3298426880
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.616850551654964e-07,
      "loss": 1.2156,
      "theoretical_loss": 3.3253313144442194,
      "tokens_seen": 3298492416
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.416248746238716e-07,
      "loss": 1.3,
      "theoretical_loss": 3.325326374599295,
      "tokens_seen": 3298557952
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.215646940822467e-07,
      "loss": 1.3118,
      "theoretical_loss": 3.325321434879995,
      "tokens_seen": 3298623488
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.015045135406219e-07,
      "loss": 1.3131,
      "theoretical_loss": 3.3253164952863132,
      "tokens_seen": 3298689024
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.814443329989971e-07,
      "loss": 1.1684,
      "theoretical_loss": 3.325311555818245,
      "tokens_seen": 3298754560
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.613841524573721e-07,
      "loss": 1.2351,
      "theoretical_loss": 3.325306616475783,
      "tokens_seen": 3298820096
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.4132397191574727e-07,
      "loss": 1.2381,
      "theoretical_loss": 3.3253016772589232,
      "tokens_seen": 3298885632
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.2126379137412237e-07,
      "loss": 1.1597,
      "theoretical_loss": 3.3252967381676592,
      "tokens_seen": 3298951168
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.012036108324975e-07,
      "loss": 1.2651,
      "theoretical_loss": 3.325291799201985,
      "tokens_seen": 3299016704
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.8114343029087263e-07,
      "loss": 1.224,
      "theoretical_loss": 3.3252868603618957,
      "tokens_seen": 3299082240
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.6108324974924774e-07,
      "loss": 1.2995,
      "theoretical_loss": 3.3252819216473846,
      "tokens_seen": 3299147776
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.4102306920762284e-07,
      "loss": 1.2241,
      "theoretical_loss": 3.325276983058447,
      "tokens_seen": 3299213312
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.20962888665998e-07,
      "loss": 1.1384,
      "theoretical_loss": 3.3252720445950765,
      "tokens_seen": 3299278848
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.0090270812437316e-07,
      "loss": 1.31,
      "theoretical_loss": 3.325267106257268,
      "tokens_seen": 3299344384
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.808425275827482e-07,
      "loss": 1.2418,
      "theoretical_loss": 3.3252621680450156,
      "tokens_seen": 3299409920
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.6078234704112337e-07,
      "loss": 1.2549,
      "theoretical_loss": 3.3252572299583134,
      "tokens_seen": 3299475456
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.4072216649949853e-07,
      "loss": 1.228,
      "theoretical_loss": 3.3252522919971557,
      "tokens_seen": 3299540992
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.2066198595787363e-07,
      "loss": 1.3123,
      "theoretical_loss": 3.325247354161537,
      "tokens_seen": 3299606528
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0060180541624874e-07,
      "loss": 1.199,
      "theoretical_loss": 3.3252424164514514,
      "tokens_seen": 3299672064
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8054162487462387e-07,
      "loss": 1.2633,
      "theoretical_loss": 3.3252374788668937,
      "tokens_seen": 3299737600
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.60481444332999e-07,
      "loss": 1.1791,
      "theoretical_loss": 3.325232541407858,
      "tokens_seen": 3299803136
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.404212637913741e-07,
      "loss": 1.1447,
      "theoretical_loss": 3.325227604074338,
      "tokens_seen": 3299868672
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.2036108324974926e-07,
      "loss": 1.26,
      "theoretical_loss": 3.325222666866329,
      "tokens_seen": 3299934208
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0030090270812437e-07,
      "loss": 1.2552,
      "theoretical_loss": 3.325217729783825,
      "tokens_seen": 3299999744
    }
  ],
  "max_steps": 25177,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 8.42054334676992e+17,
  "trial_name": null,
  "trial_params": null
}