{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 100,
  "global_step": 1275,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.47261681938543915,
      "epoch": 0.023543260741612712,
      "grad_norm": 0.53125,
      "learning_rate": 1.40625e-06,
      "loss": 0.7165036201477051,
      "mean_token_accuracy": 0.8025671623647213,
      "num_tokens": 796538.0,
      "step": 10
    },
    {
      "entropy": 0.5139306713826954,
      "epoch": 0.047086521483225424,
      "grad_norm": 0.388671875,
      "learning_rate": 2.96875e-06,
      "loss": 0.7095730304718018,
      "mean_token_accuracy": 0.7971717294305563,
      "num_tokens": 1544899.0,
      "step": 20
    },
    {
      "entropy": 0.5956933442503214,
      "epoch": 0.07062978222483814,
      "grad_norm": 0.2099609375,
      "learning_rate": 4.53125e-06,
      "loss": 0.6996610641479493,
      "mean_token_accuracy": 0.7971747264266014,
      "num_tokens": 2317683.0,
      "step": 30
    },
    {
      "entropy": 0.6512754812836647,
      "epoch": 0.09417304296645085,
      "grad_norm": 0.1474609375,
      "learning_rate": 6.093750000000001e-06,
      "loss": 0.6695662021636963,
      "mean_token_accuracy": 0.8040148165076971,
      "num_tokens": 3111422.0,
      "step": 40
    },
    {
      "entropy": 0.6480279197916389,
      "epoch": 0.11771630370806356,
      "grad_norm": 0.1083984375,
      "learning_rate": 7.656250000000001e-06,
      "loss": 0.6375884056091309,
      "mean_token_accuracy": 0.8059590008109808,
      "num_tokens": 3973704.0,
      "step": 50
    },
    {
      "entropy": 0.6220262819901109,
      "epoch": 0.14125956444967627,
      "grad_norm": 0.0986328125,
      "learning_rate": 9.21875e-06,
      "loss": 0.623339033126831,
      "mean_token_accuracy": 0.8087143875658512,
      "num_tokens": 4730669.0,
      "step": 60
    },
    {
      "entropy": 0.6261076383292675,
      "epoch": 0.164802825191289,
      "grad_norm": 0.08056640625,
      "learning_rate": 1.0781250000000001e-05,
      "loss": 0.6200876712799073,
      "mean_token_accuracy": 0.8125950843095779,
      "num_tokens": 5502078.0,
      "step": 70
    },
    {
      "entropy": 0.6319901453331113,
      "epoch": 0.1883460859329017,
      "grad_norm": 0.06591796875,
      "learning_rate": 1.234375e-05,
      "loss": 0.6182626247406006,
      "mean_token_accuracy": 0.8144623003900051,
      "num_tokens": 6266614.0,
      "step": 80
    },
    {
      "entropy": 0.602002534456551,
      "epoch": 0.21188934667451442,
      "grad_norm": 0.0625,
      "learning_rate": 1.3906250000000001e-05,
      "loss": 0.5683388710021973,
      "mean_token_accuracy": 0.8221714589744806,
      "num_tokens": 7059460.0,
      "step": 90
    },
    {
      "entropy": 0.5908293653279543,
      "epoch": 0.23543260741612712,
      "grad_norm": 0.06689453125,
      "learning_rate": 1.546875e-05,
      "loss": 0.5792154312133789,
      "mean_token_accuracy": 0.8240513067692519,
      "num_tokens": 7862593.0,
      "step": 100
    },
    {
      "epoch": 0.23543260741612712,
      "eval_entropy": 0.5910338105935624,
      "eval_loss": 0.5686901211738586,
      "eval_mean_token_accuracy": 0.823962179309163,
      "eval_num_tokens": 7862593.0,
      "eval_runtime": 139.6331,
      "eval_samples_per_second": 2.564,
      "eval_steps_per_second": 2.564,
      "step": 100
    },
    {
      "entropy": 0.5830874053761363,
      "epoch": 0.2589758681577399,
      "grad_norm": 0.058837890625,
      "learning_rate": 1.703125e-05,
      "loss": 0.5834993839263916,
      "mean_token_accuracy": 0.8253755442798137,
      "num_tokens": 8702391.0,
      "step": 110
    },
    {
      "entropy": 0.5565910406410695,
      "epoch": 0.28251912889935255,
      "grad_norm": 0.05859375,
      "learning_rate": 1.859375e-05,
      "loss": 0.5459747314453125,
      "mean_token_accuracy": 0.8344478290528059,
      "num_tokens": 9393136.0,
      "step": 120
    },
    {
      "entropy": 0.5650864312425256,
      "epoch": 0.30606238964096527,
      "grad_norm": 0.06103515625,
      "learning_rate": 1.9999962490381906e-05,
      "loss": 0.5637893676757812,
      "mean_token_accuracy": 0.829344104602933,
      "num_tokens": 10137523.0,
      "step": 130
    },
    {
      "entropy": 0.5473398096859455,
      "epoch": 0.329605650382578,
      "grad_norm": 0.0625,
      "learning_rate": 1.999546167668739e-05,
      "loss": 0.5404181003570556,
      "mean_token_accuracy": 0.8342546194791793,
      "num_tokens": 10946950.0,
      "step": 140
    },
    {
      "entropy": 0.5475172070786357,
      "epoch": 0.3531489111241907,
      "grad_norm": 0.0634765625,
      "learning_rate": 1.998346280806768e-05,
      "loss": 0.5369014263153076,
      "mean_token_accuracy": 0.8347291376441717,
      "num_tokens": 11724348.0,
      "step": 150
    },
    {
      "entropy": 0.5577760996297002,
      "epoch": 0.3766921718658034,
      "grad_norm": 0.06591796875,
      "learning_rate": 1.9963974885425267e-05,
      "loss": 0.5580552577972412,
      "mean_token_accuracy": 0.8297208815813064,
      "num_tokens": 12528416.0,
      "step": 160
    },
    {
      "entropy": 0.5457047978416085,
      "epoch": 0.4002354326074161,
      "grad_norm": 0.06103515625,
      "learning_rate": 1.9937012527546064e-05,
      "loss": 0.5416774749755859,
      "mean_token_accuracy": 0.8351736020296812,
      "num_tokens": 13254942.0,
      "step": 170
    },
    {
      "entropy": 0.5473680015653372,
      "epoch": 0.42377869334902885,
      "grad_norm": 0.058837890625,
      "learning_rate": 1.9902595960133182e-05,
      "loss": 0.5401034355163574,
      "mean_token_accuracy": 0.8332272503525019,
      "num_tokens": 14035743.0,
      "step": 180
    },
    {
      "entropy": 0.5565412795171142,
      "epoch": 0.44732195409064157,
      "grad_norm": 0.052734375,
      "learning_rate": 1.986075100063469e-05,
      "loss": 0.5520064353942871,
      "mean_token_accuracy": 0.8305680822581053,
      "num_tokens": 14777254.0,
      "step": 190
    },
    {
      "entropy": 0.5201416643336415,
      "epoch": 0.47086521483225424,
      "grad_norm": 0.055908203125,
      "learning_rate": 1.9811509038876772e-05,
      "loss": 0.5139257431030273,
      "mean_token_accuracy": 0.8422017730772495,
      "num_tokens": 15582880.0,
      "step": 200
    },
    {
      "epoch": 0.47086521483225424,
      "eval_entropy": 0.5415721843695508,
      "eval_loss": 0.5158530473709106,
      "eval_mean_token_accuracy": 0.8355566641138918,
      "eval_num_tokens": 15582880.0,
      "eval_runtime": 142.3554,
      "eval_samples_per_second": 2.515,
      "eval_steps_per_second": 2.515,
      "step": 200
    },
    {
      "entropy": 0.5265132559463381,
      "epoch": 0.49440847557386697,
      "grad_norm": 0.05615234375,
      "learning_rate": 1.9754907013516774e-05,
      "loss": 0.5199352264404297,
      "mean_token_accuracy": 0.8380576658993959,
      "num_tokens": 16419990.0,
      "step": 210
    },
    {
      "entropy": 0.5467520544305444,
      "epoch": 0.5179517363154797,
      "grad_norm": 0.05859375,
      "learning_rate": 1.9690987384333823e-05,
      "loss": 0.5457062244415283,
      "mean_token_accuracy": 0.8330623578280211,
      "num_tokens": 17220133.0,
      "step": 220
    },
    {
      "entropy": 0.5202922957018018,
      "epoch": 0.5414949970570924,
      "grad_norm": 0.0732421875,
      "learning_rate": 1.961979810037778e-05,
      "loss": 0.5149586200714111,
      "mean_token_accuracy": 0.8408578157424926,
      "num_tokens": 17988703.0,
      "step": 230
    },
    {
      "entropy": 0.5157044745981694,
      "epoch": 0.5650382577987051,
      "grad_norm": 0.05859375,
      "learning_rate": 1.954139256400049e-05,
      "loss": 0.5135462760925293,
      "mean_token_accuracy": 0.8410524364560843,
      "num_tokens": 18760495.0,
      "step": 240
    },
    {
      "entropy": 0.5239256421104074,
      "epoch": 0.5885815185403178,
      "grad_norm": 0.0634765625,
      "learning_rate": 1.9455829590796184e-05,
      "loss": 0.5215154647827148,
      "mean_token_accuracy": 0.8391639325767756,
      "num_tokens": 19551514.0,
      "step": 250
    },
    {
      "entropy": 0.5295072626322508,
      "epoch": 0.6121247792819305,
      "grad_norm": 0.054443359375,
      "learning_rate": 1.9363173365481223e-05,
      "loss": 0.5391250133514405,
      "mean_token_accuracy": 0.8365894909948111,
      "num_tokens": 20367505.0,
      "step": 260
    },
    {
      "entropy": 0.513121047988534,
      "epoch": 0.6356680400235433,
      "grad_norm": 0.057861328125,
      "learning_rate": 1.92634933937462e-05,
      "loss": 0.5058225154876709,
      "mean_token_accuracy": 0.842748760432005,
      "num_tokens": 21071323.0,
      "step": 270
    },
    {
      "entropy": 0.5175997691228986,
      "epoch": 0.659211300765156,
      "grad_norm": 0.056396484375,
      "learning_rate": 1.915686445011651e-05,
      "loss": 0.5235993385314941,
      "mean_token_accuracy": 0.8399815000593662,
      "num_tokens": 21895697.0,
      "step": 280
    },
    {
      "entropy": 0.5231689540669322,
      "epoch": 0.6827545615067687,
      "grad_norm": 0.05126953125,
      "learning_rate": 1.9043366521860558e-05,
      "loss": 0.5182478904724122,
      "mean_token_accuracy": 0.8393962673842907,
      "num_tokens": 22679040.0,
      "step": 290
    },
    {
      "entropy": 0.5049041209742426,
      "epoch": 0.7062978222483814,
      "grad_norm": 0.062255859375,
      "learning_rate": 1.892308474898763e-05,
      "loss": 0.4974985599517822,
      "mean_token_accuracy": 0.8447154741734266,
      "num_tokens": 23386138.0,
      "step": 300
    },
    {
      "epoch": 0.7062978222483814,
      "eval_entropy": 0.5194896821702659,
      "eval_loss": 0.49936261773109436,
      "eval_mean_token_accuracy": 0.8389908150254681,
      "eval_num_tokens": 23386138.0,
      "eval_runtime": 136.5483,
      "eval_samples_per_second": 2.622,
      "eval_steps_per_second": 2.622,
      "step": 300
    },
    {
      "entropy": 0.5262435658834874,
      "epoch": 0.7298410829899941,
      "grad_norm": 0.0546875,
      "learning_rate": 1.879610936038046e-05,
      "loss": 0.5304554462432861,
      "mean_token_accuracy": 0.836079678311944,
      "num_tokens": 24170172.0,
      "step": 310
    },
    {
      "entropy": 0.519707421399653,
      "epoch": 0.7533843437316068,
      "grad_norm": 0.05419921875,
      "learning_rate": 1.86625356061104e-05,
      "loss": 0.525363826751709,
      "mean_token_accuracy": 0.8391638267785311,
      "num_tokens": 24962487.0,
      "step": 320
    },
    {
      "entropy": 0.5116234797984361,
      "epoch": 0.7769276044732195,
      "grad_norm": 0.06591796875,
      "learning_rate": 1.852246368598597e-05,
      "loss": 0.5102458953857422,
      "mean_token_accuracy": 0.8417690549045801,
      "num_tokens": 25733914.0,
      "step": 330
    },
    {
      "entropy": 0.5263846382498741,
      "epoch": 0.8004708652148322,
      "grad_norm": 0.0546875,
      "learning_rate": 1.837599867438838e-05,
      "loss": 0.5220338821411132,
      "mean_token_accuracy": 0.8377181738615036,
      "num_tokens": 26579132.0,
      "step": 340
    },
    {
      "entropy": 0.5115921972319484,
      "epoch": 0.824014125956445,
      "grad_norm": 0.0615234375,
      "learning_rate": 1.822325044145041e-05,
      "loss": 0.49653100967407227,
      "mean_token_accuracy": 0.8414182372391223,
      "num_tokens": 27202803.0,
      "step": 350
    },
    {
      "entropy": 0.5089998429641127,
      "epoch": 0.8475573866980577,
      "grad_norm": 0.0673828125,
      "learning_rate": 1.8064333570637768e-05,
      "loss": 0.49803886413574217,
      "mean_token_accuracy": 0.842063357308507,
      "num_tokens": 27874601.0,
      "step": 360
    },
    {
      "entropy": 0.5171531289815903,
      "epoch": 0.8711006474396704,
      "grad_norm": 0.059814453125,
      "learning_rate": 1.7899367272794777e-05,
      "loss": 0.5236594676971436,
      "mean_token_accuracy": 0.8396157670766116,
      "num_tokens": 28758269.0,
      "step": 370
    },
    {
      "entropy": 0.5161923897452653,
      "epoch": 0.8946439081812831,
      "grad_norm": 0.047607421875,
      "learning_rate": 1.7728475296718846e-05,
      "loss": 0.5221555709838868,
      "mean_token_accuracy": 0.8399753466248512,
      "num_tokens": 29660317.0,
      "step": 380
    },
    {
      "entropy": 0.4955171117559075,
      "epoch": 0.9181871689228959,
      "grad_norm": 0.05517578125,
      "learning_rate": 1.7551785836330814e-05,
      "loss": 0.49817495346069335,
      "mean_token_accuracy": 0.846830865368247,
      "num_tokens": 30349463.0,
      "step": 390
    },
    {
      "entropy": 0.505057031661272,
      "epoch": 0.9417304296645085,
      "grad_norm": 0.064453125,
      "learning_rate": 1.7369431434510793e-05,
      "loss": 0.509951400756836,
      "mean_token_accuracy": 0.8432859953492879,
      "num_tokens": 31065185.0,
      "step": 400
    },
    {
      "epoch": 0.9417304296645085,
      "eval_entropy": 0.5111477355883774,
      "eval_loss": 0.4909190237522125,
      "eval_mean_token_accuracy": 0.8413028815272134,
      "eval_num_tokens": 31065185.0,
      "eval_runtime": 131.4779,
      "eval_samples_per_second": 2.723,
      "eval_steps_per_second": 2.723,
      "step": 400
    },
    {
      "entropy": 0.4994705172255635,
      "epoch": 0.9652736904061212,
      "grad_norm": 0.059814453125,
      "learning_rate": 1.718154888367167e-05,
      "loss": 0.5129815101623535,
      "mean_token_accuracy": 0.8450050104409457,
      "num_tokens": 31796395.0,
      "step": 410
    },
    {
      "entropy": 0.5047325070947408,
      "epoch": 0.9888169511477339,
      "grad_norm": 0.054443359375,
      "learning_rate": 1.698827912314484e-05,
      "loss": 0.4937889099121094,
      "mean_token_accuracy": 0.8428907435387373,
      "num_tokens": 32560011.0,
      "step": 420
    },
    {
      "entropy": 0.5052787409378932,
      "epoch": 1.0117716303708064,
      "grad_norm": 0.05908203125,
      "learning_rate": 1.678976713345513e-05,
      "loss": 0.5136929988861084,
      "mean_token_accuracy": 0.8436831063949145,
      "num_tokens": 33280194.0,
      "step": 430
    },
    {
      "entropy": 0.49020985178649423,
      "epoch": 1.035314891112419,
      "grad_norm": 0.05908203125,
      "learning_rate": 1.658616182756426e-05,
      "loss": 0.4995232582092285,
      "mean_token_accuracy": 0.847404608130455,
      "num_tokens": 34034683.0,
      "step": 440
    },
    {
      "entropy": 0.513733028434217,
      "epoch": 1.0588581518540319,
      "grad_norm": 0.0673828125,
      "learning_rate": 1.6377615939164387e-05,
      "loss": 0.524590253829956,
      "mean_token_accuracy": 0.8403749600052833,
      "num_tokens": 34873128.0,
      "step": 450
    },
    {
      "entropy": 0.5072762167081237,
      "epoch": 1.0824014125956445,
      "grad_norm": 0.06298828125,
      "learning_rate": 1.6164285908105526e-05,
      "loss": 0.520505142211914,
      "mean_token_accuracy": 0.8410047568380833,
      "num_tokens": 35734174.0,
      "step": 460
    },
    {
      "entropy": 0.49719911720603704,
      "epoch": 1.105944673337257,
      "grad_norm": 0.06298828125,
      "learning_rate": 1.594633176304287e-05,
      "loss": 0.494351863861084,
      "mean_token_accuracy": 0.8454386904835701,
      "num_tokens": 36393777.0,
      "step": 470
    },
    {
      "entropy": 0.49072135630995034,
      "epoch": 1.12948793407887,
      "grad_norm": 0.05517578125,
      "learning_rate": 1.5723917001391912e-05,
      "loss": 0.4863902568817139,
      "mean_token_accuracy": 0.8458908643573523,
      "num_tokens": 37251068.0,
      "step": 480
    },
    {
      "entropy": 0.4939659545198083,
      "epoch": 1.1530311948204826,
      "grad_norm": 0.06494140625,
      "learning_rate": 1.5497208466681547e-05,
      "loss": 0.4998356342315674,
      "mean_token_accuracy": 0.8443102762103081,
      "num_tokens": 38031375.0,
      "step": 490
    },
    {
      "entropy": 0.4939275814220309,
      "epoch": 1.1765744555620954,
      "grad_norm": 0.055419921875,
      "learning_rate": 1.526637622339706e-05,
      "loss": 0.4955411911010742,
      "mean_token_accuracy": 0.8455841716378927,
      "num_tokens": 38860255.0,
      "step": 500
    },
    {
      "epoch": 1.1765744555620954,
      "eval_entropy": 0.5045043116008769,
      "eval_loss": 0.4862433075904846,
      "eval_mean_token_accuracy": 0.8424804836678106,
      "eval_num_tokens": 38860255.0,
      "eval_runtime": 135.2782,
      "eval_samples_per_second": 2.646,
      "eval_steps_per_second": 2.646,
      "step": 500
    },
    {
      "entropy": 0.4797388563863933,
      "epoch": 1.200117716303708,
      "grad_norm": 0.056396484375,
      "learning_rate": 1.5031593429406964e-05,
      "loss": 0.484775447845459,
      "mean_token_accuracy": 0.8496320709586144,
      "num_tokens": 39714269.0,
      "step": 510
    },
    {
      "entropy": 0.5136959448456764,
      "epoch": 1.2236609770453208,
      "grad_norm": 0.05419921875,
      "learning_rate": 1.4793036206069298e-05,
      "loss": 0.5073910236358643,
      "mean_token_accuracy": 0.839500718563795,
      "num_tokens": 40429331.0,
      "step": 520
    },
    {
      "entropy": 0.5001886915415525,
      "epoch": 1.2472042377869335,
      "grad_norm": 0.059814453125,
      "learning_rate": 1.455088350611494e-05,
      "loss": 0.4849071502685547,
      "mean_token_accuracy": 0.8452773235738278,
      "num_tokens": 41120160.0,
      "step": 530
    },
    {
      "entropy": 0.49588965652510525,
      "epoch": 1.270747498528546,
      "grad_norm": 0.05322265625,
      "learning_rate": 1.430531697940693e-05,
      "loss": 0.5028933525085449,
      "mean_token_accuracy": 0.8444558009505272,
      "num_tokens": 41875187.0,
      "step": 540
    },
    {
      "entropy": 0.4997696656733751,
      "epoch": 1.294290759270159,
      "grad_norm": 0.056884765625,
      "learning_rate": 1.4056520836676551e-05,
      "loss": 0.5174314498901367,
      "mean_token_accuracy": 0.8428490672260522,
      "num_tokens": 42701302.0,
      "step": 550
    },
    {
      "entropy": 0.49855835027992723,
      "epoch": 1.3178340200117717,
      "grad_norm": 0.059814453125,
      "learning_rate": 1.380468171133842e-05,
      "loss": 0.5048851013183594,
      "mean_token_accuracy": 0.8445437770336867,
      "num_tokens": 43490481.0,
      "step": 560
    },
    {
      "entropy": 0.4793679520487785,
      "epoch": 1.3413772807533844,
      "grad_norm": 0.058837890625,
      "learning_rate": 1.354998851948818e-05,
      "loss": 0.48621335029602053,
      "mean_token_accuracy": 0.8496838890016079,
      "num_tokens": 44257286.0,
      "step": 570
    },
    {
      "entropy": 0.5044341992586852,
      "epoch": 1.364920541494997,
      "grad_norm": 0.06591796875,
      "learning_rate": 1.3292632318187857e-05,
      "loss": 0.5028728485107422,
      "mean_token_accuracy": 0.8416947480291128,
      "num_tokens": 45031884.0,
      "step": 580
    },
    {
      "entropy": 0.4860644720494747,
      "epoch": 1.3884638022366098,
      "grad_norm": 0.054443359375,
      "learning_rate": 1.3032806162145193e-05,
      "loss": 0.48934264183044435,
      "mean_token_accuracy": 0.848303421586752,
      "num_tokens": 45805257.0,
      "step": 590
    },
    {
      "entropy": 0.49228332079946996,
      "epoch": 1.4120070629782224,
      "grad_norm": 0.051025390625,
      "learning_rate": 1.2770704958894424e-05,
      "loss": 0.48878173828125,
      "mean_token_accuracy": 0.8455546554178,
      "num_tokens": 46529684.0,
      "step": 600
    },
    {
      "epoch": 1.4120070629782224,
      "eval_entropy": 0.49637800954573646,
      "eval_loss": 0.4830160439014435,
      "eval_mean_token_accuracy": 0.8431555377371485,
      "eval_num_tokens": 46529684.0,
      "eval_runtime": 128.7956,
      "eval_samples_per_second": 2.78,
      "eval_steps_per_second": 2.78,
      "step": 600
    },
    {
      "entropy": 0.4935727553442121,
      "epoch": 1.4355503237198353,
      "grad_norm": 0.05712890625,
      "learning_rate": 1.2506525322587207e-05,
      "loss": 0.5051933765411377,
      "mean_token_accuracy": 0.8451346360147,
      "num_tokens": 47351703.0,
      "step": 610
    },
    {
      "entropy": 0.4920971740037203,
      "epoch": 1.4590935844614479,
      "grad_norm": 0.053466796875,
      "learning_rate": 1.2240465426503296e-05,
      "loss": 0.5081326484680175,
      "mean_token_accuracy": 0.8450745042413473,
      "num_tokens": 48111074.0,
      "step": 620
    },
    {
      "entropy": 0.4823517481796443,
      "epoch": 1.4826368452030607,
      "grad_norm": 0.052978515625,
      "learning_rate": 1.197272485439167e-05,
      "loss": 0.4746738910675049,
      "mean_token_accuracy": 0.8489425461739302,
      "num_tokens": 48899473.0,
      "step": 630
    },
    {
      "entropy": 0.4855606323108077,
      "epoch": 1.5061801059446733,
      "grad_norm": 0.0673828125,
      "learning_rate": 1.1703504450753604e-05,
      "loss": 0.4840294361114502,
      "mean_token_accuracy": 0.8471518792212009,
      "num_tokens": 49689664.0,
      "step": 640
    },
    {
      "entropy": 0.4964427137747407,
      "epoch": 1.529723366686286,
      "grad_norm": 0.055908203125,
      "learning_rate": 1.1433006170179961e-05,
      "loss": 0.499052095413208,
      "mean_token_accuracy": 0.8452402614057064,
      "num_tokens": 50432542.0,
      "step": 650
    },
    {
      "entropy": 0.48346656998619436,
      "epoch": 1.5532666274278988,
      "grad_norm": 0.059814453125,
      "learning_rate": 1.1161432925855792e-05,
      "loss": 0.49048147201538084,
      "mean_token_accuracy": 0.8482994694262743,
      "num_tokens": 51235911.0,
      "step": 660
    },
    {
      "entropy": 0.4915179075673223,
      "epoch": 1.5768098881695116,
      "grad_norm": 0.0673828125,
      "learning_rate": 1.0888988437345851e-05,
      "loss": 0.48571577072143557,
      "mean_token_accuracy": 0.8463525585830212,
      "num_tokens": 51914843.0,
      "step": 670
    },
    {
      "entropy": 0.5079520621336997,
      "epoch": 1.6003531489111242,
      "grad_norm": 0.0625,
      "learning_rate": 1.061587707777515e-05,
      "loss": 0.5094398975372314,
      "mean_token_accuracy": 0.8409215942025184,
      "num_tokens": 52788907.0,
      "step": 680
    },
    {
      "entropy": 0.4922261985018849,
      "epoch": 1.6238964096527368,
      "grad_norm": 0.06640625,
      "learning_rate": 1.0342303720519331e-05,
      "loss": 0.5027109146118164,
      "mean_token_accuracy": 0.8461885400116443,
      "num_tokens": 53610651.0,
      "step": 690
    },
    {
      "entropy": 0.4926283808425069,
      "epoch": 1.6474396703943497,
      "grad_norm": 0.052978515625,
      "learning_rate": 1.0068473585519714e-05,
      "loss": 0.48818025588989256,
      "mean_token_accuracy": 0.8448698211461305,
      "num_tokens": 54306359.0,
      "step": 700
    },
    {
      "epoch": 1.6474396703943497,
      "eval_entropy": 0.49725391719927337,
      "eval_loss": 0.48121583461761475,
      "eval_mean_token_accuracy": 0.8434386894356605,
      "eval_num_tokens": 54306359.0,
      "eval_runtime": 129.9385,
      "eval_samples_per_second": 2.755,
      "eval_steps_per_second": 2.755,
      "step": 700
    },
    {
      "entropy": 0.50206049811095,
      "epoch": 1.6709829311359623,
      "grad_norm": 0.05126953125,
      "learning_rate": 9.794592085338401e-06,
      "loss": 0.5090510845184326,
      "mean_token_accuracy": 0.8430146727710962,
      "num_tokens": 55162088.0,
      "step": 710
    },
    {
      "entropy": 0.4935040036216378,
      "epoch": 1.694526191877575,
      "grad_norm": 0.051513671875,
      "learning_rate": 9.520864671068875e-06,
      "loss": 0.5052507400512696,
      "mean_token_accuracy": 0.8460971619933844,
      "num_tokens": 55957773.0,
      "step": 720
    },
    {
      "entropy": 0.4884185861796141,
      "epoch": 1.7180694526191878,
      "grad_norm": 0.06005859375,
      "learning_rate": 9.247496678217704e-06,
      "loss": 0.4981977939605713,
      "mean_token_accuracy": 0.8484526939690114,
      "num_tokens": 56705800.0,
      "step": 730
    },
    {
      "entropy": 0.47191233560442924,
      "epoch": 1.7416127133608006,
      "grad_norm": 0.06396484375,
      "learning_rate": 8.974693172672936e-06,
      "loss": 0.48612565994262696,
      "mean_token_accuracy": 0.8515177499502897,
      "num_tokens": 57410830.0,
      "step": 740
    },
    {
      "entropy": 0.4871264295652509,
      "epoch": 1.7651559741024132,
      "grad_norm": 0.05615234375,
      "learning_rate": 8.70265879687476e-06,
      "loss": 0.4872490882873535,
      "mean_token_accuracy": 0.8473320409655571,
      "num_tokens": 58099041.0,
      "step": 750
    },
    {
      "entropy": 0.4919566001743078,
      "epoch": 1.7886992348440258,
      "grad_norm": 0.0625,
      "learning_rate": 8.431597616303818e-06,
      "loss": 0.5038920402526855,
      "mean_token_accuracy": 0.8457822546362876,
      "num_tokens": 58932723.0,
      "step": 760
    },
    {
      "entropy": 0.4820005618967116,
      "epoch": 1.8122424955856387,
      "grad_norm": 0.05810546875,
      "learning_rate": 8.16171296640231e-06,
      "loss": 0.4845709323883057,
      "mean_token_accuracy": 0.848111093044281,
      "num_tokens": 59693185.0,
      "step": 770
    },
    {
      "entropy": 0.4959249753504992,
      "epoch": 1.8357857563272513,
      "grad_norm": 0.048828125,
      "learning_rate": 7.89320730004274e-06,
      "loss": 0.5166537761688232,
      "mean_token_accuracy": 0.8443613056093454,
      "num_tokens": 60491010.0,
      "step": 780
    },
    {
      "entropy": 0.49144728798419235,
      "epoch": 1.8593290170688639,
      "grad_norm": 0.05908203125,
      "learning_rate": 7.626282035658703e-06,
      "loss": 0.49805240631103515,
      "mean_token_accuracy": 0.8461739920079708,
      "num_tokens": 61319991.0,
      "step": 790
    },
    {
      "entropy": 0.4859143828973174,
      "epoch": 1.8828722778104767,
      "grad_norm": 0.054931640625,
      "learning_rate": 7.361137406151706e-06,
      "loss": 0.4953606128692627,
      "mean_token_accuracy": 0.8474723279476166,
      "num_tokens": 62135256.0,
      "step": 800
    },
    {
      "epoch": 1.8828722778104767,
      "eval_entropy": 0.49684396469393255,
      "eval_loss": 0.48009854555130005,
      "eval_mean_token_accuracy": 0.8439133836570398,
      "eval_num_tokens": 62135256.0,
      "eval_runtime": 140.0838,
      "eval_samples_per_second": 2.556,
      "eval_steps_per_second": 2.556,
      "step": 800
    },
    {
      "entropy": 0.4975118497386575,
      "epoch": 1.9064155385520896,
      "grad_norm": 0.055908203125,
      "learning_rate": 7.0979723086872245e-06,
      "loss": 0.5027891159057617,
      "mean_token_accuracy": 0.8441127471625804,
      "num_tokens": 62886746.0,
      "step": 810
    },
    {
      "entropy": 0.48534935507923366,
      "epoch": 1.9299587992937022,
      "grad_norm": 0.060546875,
      "learning_rate": 6.8369841554928226e-06,
      "loss": 0.48329763412475585,
      "mean_token_accuracy": 0.8473416466265917,
      "num_tokens": 63579402.0,
      "step": 820
    },
    {
      "entropy": 0.484352495148778,
      "epoch": 1.9535020600353148,
      "grad_norm": 0.05029296875,
      "learning_rate": 6.578368725770182e-06,
      "loss": 0.48952536582946776,
      "mean_token_accuracy": 0.8478636931627989,
      "num_tokens": 64391578.0,
      "step": 830
    },
    {
      "entropy": 0.48771984670311214,
      "epoch": 1.9770453207769276,
      "grad_norm": 0.0498046875,
      "learning_rate": 6.322320018832125e-06,
      "loss": 0.4877884864807129,
      "mean_token_accuracy": 0.8477956593036652,
      "num_tokens": 65101582.0,
      "step": 840
    },
    {
      "entropy": 0.4792339194279451,
      "epoch": 2.0,
      "grad_norm": 0.0693359375,
      "learning_rate": 6.069030108574806e-06,
      "loss": 0.4753533363342285,
      "mean_token_accuracy": 0.8492753834296496,
      "num_tokens": 65847956.0,
      "step": 850
    },
    {
      "entropy": 0.4897763991728425,
      "epoch": 2.023543260741613,
      "grad_norm": 0.0595703125,
      "learning_rate": 5.8186889993942704e-06,
      "loss": 0.49657087326049804,
      "mean_token_accuracy": 0.8467563409358263,
      "num_tokens": 66639763.0,
      "step": 860
    },
    {
      "entropy": 0.4955088282003999,
      "epoch": 2.0470865214832252,
      "grad_norm": 0.064453125,
      "learning_rate": 5.571484483655443e-06,
      "loss": 0.5085082530975342,
      "mean_token_accuracy": 0.8449712213128805,
      "num_tokens": 67385166.0,
      "step": 870
    },
    {
      "entropy": 0.4888547468930483,
      "epoch": 2.070629782224838,
      "grad_norm": 0.05810546875,
      "learning_rate": 5.327602000820413e-06,
      "loss": 0.4888143539428711,
      "mean_token_accuracy": 0.846753703430295,
      "num_tokens": 68099723.0,
      "step": 880
    },
    {
      "entropy": 0.49544552695006133,
      "epoch": 2.094173042966451,
      "grad_norm": 0.054443359375,
      "learning_rate": 5.087224498341815e-06,
      "loss": 0.49994626045227053,
      "mean_token_accuracy": 0.843446821346879,
      "num_tokens": 68788696.0,
      "step": 890
    },
    {
      "entropy": 0.4897686781361699,
      "epoch": 2.1177163037080637,
      "grad_norm": 0.062255859375,
      "learning_rate": 4.8505322944255096e-06,
      "loss": 0.49230360984802246,
      "mean_token_accuracy": 0.8458703383803368,
      "num_tokens": 69602614.0,
      "step": 900
    },
    {
      "epoch": 2.1177163037080637,
      "eval_entropy": 0.494421407497129,
      "eval_loss": 0.47960278391838074,
      "eval_mean_token_accuracy": 0.8439072530362859,
      "eval_num_tokens": 69602614.0,
      "eval_runtime": 136.7014,
      "eval_samples_per_second": 2.619,
      "eval_steps_per_second": 2.619,
      "step": 900
    },
    {
      "entropy": 0.4872464815154672,
      "epoch": 2.141259564449676,
      "grad_norm": 0.059326171875,
      "learning_rate": 4.617702942765641e-06,
      "loss": 0.49840707778930665,
      "mean_token_accuracy": 0.847221752628684,
      "num_tokens": 70400747.0,
      "step": 910
    },
    {
      "entropy": 0.4936610346660018,
      "epoch": 2.164802825191289,
      "grad_norm": 0.064453125,
      "learning_rate": 4.388911099353446e-06,
      "loss": 0.5067010402679444,
      "mean_token_accuracy": 0.8448445748537778,
      "num_tokens": 71161256.0,
      "step": 920
    },
    {
      "entropy": 0.4865023369900882,
      "epoch": 2.188346085932902,
      "grad_norm": 0.05712890625,
      "learning_rate": 4.164328391459752e-06,
      "loss": 0.48999724388122556,
      "mean_token_accuracy": 0.8471139241009951,
      "num_tokens": 71912573.0,
      "step": 930
    },
    {
      "entropy": 0.48061597514897586,
      "epoch": 2.211889346674514,
      "grad_norm": 0.05810546875,
      "learning_rate": 3.944123288889516e-06,
      "loss": 0.4725522518157959,
      "mean_token_accuracy": 0.8494672421365976,
      "num_tokens": 72707432.0,
      "step": 940
    },
    {
      "entropy": 0.5059708185493946,
      "epoch": 2.235432607416127,
      "grad_norm": 0.05419921875,
      "learning_rate": 3.728460977604852e-06,
      "loss": 0.5131529331207275,
      "mean_token_accuracy": 0.8409805115312338,
      "num_tokens": 73449621.0,
      "step": 950
    },
    {
      "entropy": 0.4863934010267258,
      "epoch": 2.25897586815774,
      "grad_norm": 0.055419921875,
      "learning_rate": 3.517503235811468e-06,
      "loss": 0.5088976860046387,
      "mean_token_accuracy": 0.846961434558034,
      "num_tokens": 74347033.0,
      "step": 960
    },
    {
      "entropy": 0.4784310838207603,
      "epoch": 2.2825191288993527,
      "grad_norm": 0.0634765625,
      "learning_rate": 3.3114083126014097e-06,
      "loss": 0.47273902893066405,
      "mean_token_accuracy": 0.8500492256134748,
      "num_tokens": 75011148.0,
      "step": 970
    },
    {
      "entropy": 0.48275542035698893,
      "epoch": 2.306062389640965,
      "grad_norm": 0.06982421875,
      "learning_rate": 3.110330809243134e-06,
      "loss": 0.4815514087677002,
      "mean_token_accuracy": 0.8492911938577891,
      "num_tokens": 75701952.0,
      "step": 980
    },
    {
      "entropy": 0.4942383628338575,
      "epoch": 2.329605650382578,
      "grad_norm": 0.05029296875,
      "learning_rate": 2.914421563208003e-06,
      "loss": 0.5011926174163819,
      "mean_token_accuracy": 0.8445728730410338,
      "num_tokens": 76495004.0,
      "step": 990
    },
    {
      "entropy": 0.4953793972730637,
      "epoch": 2.353148911124191,
      "grad_norm": 0.050537109375,
      "learning_rate": 2.7238275350201603e-06,
      "loss": 0.49687824249267576,
      "mean_token_accuracy": 0.8450685154646635,
      "num_tokens": 77212538.0,
      "step": 1000
    },
    {
      "epoch": 2.353148911124191,
      "eval_entropy": 0.49301343573538287,
      "eval_loss": 0.4794597625732422,
      "eval_mean_token_accuracy": 0.8439149758336264,
      "eval_num_tokens": 77212538.0,
      "eval_runtime": 126.3492,
      "eval_samples_per_second": 2.833,
      "eval_steps_per_second": 2.833,
      "step": 1000
    },
    {
      "entropy": 0.4894041134044528,
      "epoch": 2.3766921718658036,
      "grad_norm": 0.04931640625,
      "learning_rate": 2.5386916980147126e-06,
      "loss": 0.49138665199279785,
      "mean_token_accuracy": 0.8464279536157846,
      "num_tokens": 78026138.0,
      "step": 1010
    },
    {
      "entropy": 0.4856866804882884,
      "epoch": 2.400235432607416,
      "grad_norm": 0.057861328125,
      "learning_rate": 2.3591529310868376e-06,
      "loss": 0.4937007904052734,
      "mean_token_accuracy": 0.8472504287958145,
      "num_tokens": 78780085.0,
      "step": 1020
    },
    {
      "entropy": 0.4839438125491142,
      "epoch": 2.423778693349029,
      "grad_norm": 0.05126953125,
      "learning_rate": 2.18534591451236e-06,
      "loss": 0.4821192264556885,
      "mean_token_accuracy": 0.8487556364387274,
      "num_tokens": 79570654.0,
      "step": 1030
    },
    {
      "entropy": 0.4880734058097005,
      "epoch": 2.4473219540906417,
      "grad_norm": 0.06298828125,
      "learning_rate": 2.0174010289178914e-06,
      "loss": 0.49988112449645994,
      "mean_token_accuracy": 0.8469519291073084,
      "num_tokens": 80250823.0,
      "step": 1040
    },
    {
      "entropy": 0.480747524369508,
      "epoch": 2.470865214832254,
      "grad_norm": 0.05859375,
      "learning_rate": 1.8554442574763298e-06,
      "loss": 0.47515578269958497,
      "mean_token_accuracy": 0.8491792384535074,
      "num_tokens": 81007390.0,
      "step": 1050
    },
    {
      "entropy": 0.4821536695584655,
      "epoch": 2.494408475573867,
      "grad_norm": 0.05712890625,
      "learning_rate": 1.699597091401093e-06,
      "loss": 0.4729485511779785,
      "mean_token_accuracy": 0.8483801148831844,
      "num_tokens": 81740993.0,
      "step": 1060
    },
    {
      "entropy": 0.472045193053782,
      "epoch": 2.5179517363154797,
      "grad_norm": 0.060546875,
      "learning_rate": 1.549976438809978e-06,
      "loss": 0.4712677955627441,
      "mean_token_accuracy": 0.8525977376848459,
      "num_tokens": 82419756.0,
      "step": 1070
    },
    {
      "entropy": 0.4656744490377605,
      "epoch": 2.541494997057092,
      "grad_norm": 0.058349609375,
      "learning_rate": 1.4066945370270225e-06,
      "loss": 0.47236061096191406,
      "mean_token_accuracy": 0.8537530150264502,
      "num_tokens": 83196283.0,
      "step": 1080
    },
    {
      "entropy": 0.478969038836658,
      "epoch": 2.565038257798705,
      "grad_norm": 0.0556640625,
      "learning_rate": 1.2698588683881185e-06,
      "loss": 0.48878769874572753,
      "mean_token_accuracy": 0.8492098618298769,
      "num_tokens": 84036439.0,
      "step": 1090
    },
    {
      "entropy": 0.4828123481944203,
      "epoch": 2.588581518540318,
      "grad_norm": 0.0556640625,
      "learning_rate": 1.1395720796135868e-06,
      "loss": 0.4844477653503418,
      "mean_token_accuracy": 0.8475308798253536,
      "num_tokens": 84820180.0,
      "step": 1100
    },
    {
      "epoch": 2.588581518540318,
      "eval_entropy": 0.4935321879453499,
      "eval_loss": 0.4793815016746521,
      "eval_mean_token_accuracy": 0.8439810825459784,
      "eval_num_tokens": 84820180.0,
      "eval_runtime": 126.7867,
      "eval_samples_per_second": 2.824,
      "eval_steps_per_second": 2.824,
      "step": 1100
    },
    {
      "entropy": 0.49042358435690403,
      "epoch": 2.6121247792819307,
      "grad_norm": 0.06005859375,
      "learning_rate": 1.0159319048081463e-06,
      "loss": 0.49592008590698244,
      "mean_token_accuracy": 0.8456685792654752,
      "num_tokens": 85635100.0,
      "step": 1110
    },
    {
      "entropy": 0.4801861563697457,
      "epoch": 2.6356680400235435,
      "grad_norm": 0.0498046875,
      "learning_rate": 8.990310921460865e-07,
      "loss": 0.48504209518432617,
      "mean_token_accuracy": 0.8492361098527909,
      "num_tokens": 86360259.0,
      "step": 1120
    },
    {
      "entropy": 0.47704076319932937,
      "epoch": 2.659211300765156,
      "grad_norm": 0.056640625,
      "learning_rate": 7.889573342965884e-07,
      "loss": 0.4823050022125244,
      "mean_token_accuracy": 0.8498645894229412,
      "num_tokens": 87089379.0,
      "step": 1130
    },
    {
      "entropy": 0.491375151835382,
      "epoch": 2.6827545615067687,
      "grad_norm": 0.055419921875,
      "learning_rate": 6.857932026414305e-07,
      "loss": 0.4903556823730469,
      "mean_token_accuracy": 0.8461613107472659,
      "num_tokens": 87800946.0,
      "step": 1140
    },
    {
      "entropy": 0.49147001672536134,
      "epoch": 2.7062978222483816,
      "grad_norm": 0.05029296875,
      "learning_rate": 5.896160853344079e-07,
      "loss": 0.49661922454833984,
      "mean_token_accuracy": 0.8448879312723875,
      "num_tokens": 88653378.0,
      "step": 1150
    },
    {
      "entropy": 0.48449703082442286,
      "epoch": 2.729841082989994,
      "grad_norm": 0.0537109375,
      "learning_rate": 5.00498129248913e-07,
      "loss": 0.49818434715271,
      "mean_token_accuracy": 0.8479541372507811,
      "num_tokens": 89456233.0,
      "step": 1160
    },
    {
      "entropy": 0.5064624701626599,
      "epoch": 2.753384343731607,
      "grad_norm": 0.061279296875,
      "learning_rate": 4.1850618585726477e-07,
      "loss": 0.5106511592864991,
      "mean_token_accuracy": 0.8413057509809733,
      "num_tokens": 90273983.0,
      "step": 1170
    },
    {
      "entropy": 0.47172180134803054,
      "epoch": 2.7769276044732196,
      "grad_norm": 0.046875,
      "learning_rate": 3.437017610823323e-07,
      "loss": 0.4803746223449707,
      "mean_token_accuracy": 0.8520814351737499,
      "num_tokens": 91088834.0,
      "step": 1180
    },
    {
      "entropy": 0.48386119250208137,
      "epoch": 2.800470865214832,
      "grad_norm": 0.05908203125,
      "learning_rate": 2.7614096915913325e-07,
      "loss": 0.5014448642730713,
      "mean_token_accuracy": 0.848398182913661,
      "num_tokens": 91850372.0,
      "step": 1190
    },
    {
      "entropy": 0.473122557811439,
      "epoch": 2.824014125956445,
      "grad_norm": 0.0537109375,
      "learning_rate": 2.1587449054095933e-07,
      "loss": 0.46098814010620115,
      "mean_token_accuracy": 0.8519114807248116,
      "num_tokens": 92604266.0,
      "step": 1200
    },
    {
      "epoch": 2.824014125956445,
      "eval_entropy": 0.49339639424611736,
      "eval_loss": 0.4793107807636261,
      "eval_mean_token_accuracy": 0.8440321680553798,
      "eval_num_tokens": 92604266.0,
      "eval_runtime": 129.0476,
      "eval_samples_per_second": 2.774,
      "eval_steps_per_second": 2.774,
      "step": 1200
    },
    {
      "entropy": 0.4882160659879446,
      "epoch": 2.8475573866980577,
      "grad_norm": 0.0576171875,
      "learning_rate": 1.6294753388163886e-07,
      "loss": 0.4841885566711426,
      "mean_token_accuracy": 0.8477579735219478,
      "num_tokens": 93383250.0,
      "step": 1210
    },
    {
      "entropy": 0.4832633834332228,
      "epoch": 2.8711006474396705,
      "grad_norm": 0.05859375,
      "learning_rate": 1.1739980212245983e-07,
      "loss": 0.49979395866394044,
      "mean_token_accuracy": 0.849021689593792,
      "num_tokens": 94238149.0,
      "step": 1220
    },
    {
      "entropy": 0.5064708769321442,
      "epoch": 2.8946439081812834,
      "grad_norm": 0.05517578125,
      "learning_rate": 7.926546270915603e-08,
      "loss": 0.523139476776123,
      "mean_token_accuracy": 0.8406466241925955,
      "num_tokens": 95043273.0,
      "step": 1230
    },
    {
      "entropy": 0.47456965055316686,
      "epoch": 2.9181871689228958,
      "grad_norm": 0.0478515625,
      "learning_rate": 4.8573121961347226e-08,
      "loss": 0.4729292392730713,
      "mean_token_accuracy": 0.8510482702404261,
      "num_tokens": 95859784.0,
      "step": 1240
    },
    {
      "entropy": 0.4945228073745966,
      "epoch": 2.9417304296645086,
      "grad_norm": 0.0576171875,
      "learning_rate": 2.5345803613617826e-08,
      "loss": 0.49330801963806153,
      "mean_token_accuracy": 0.8454886879771948,
      "num_tokens": 96662576.0,
      "step": 1250
    },
    {
      "entropy": 0.48427404602989554,
      "epoch": 2.9652736904061214,
      "grad_norm": 0.051513671875,
      "learning_rate": 9.600931544362369e-09,
      "loss": 0.4869691848754883,
      "mean_token_accuracy": 0.8485142230987549,
      "num_tokens": 97491833.0,
      "step": 1260
    },
    {
      "entropy": 0.48224083045497534,
      "epoch": 2.988816951147734,
      "grad_norm": 0.052734375,
      "learning_rate": 1.3503167053285916e-09,
      "loss": 0.4842352390289307,
      "mean_token_accuracy": 0.8483625873923302,
      "num_tokens": 98341011.0,
      "step": 1270
    },
    {
      "epoch": 3.0,
      "eval_entropy": 0.49332542625885434,
      "eval_loss": 0.4793912470340729,
      "eval_mean_token_accuracy": 0.8440024263365975,
      "eval_num_tokens": 98771934.0,
      "eval_runtime": 125.3024,
      "eval_samples_per_second": 2.857,
      "eval_steps_per_second": 2.857,
      "step": 1275
    }
  ],
  "logging_steps": 10,
  "max_steps": 1275,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 8.35260937439957e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}