{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.041736227045075125,
  "eval_steps": 500,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "cache_hit_rate": 0.5591798931360245,
      "cache_hit_rate_size2": 0.3176217913627625,
      "cache_hit_rate_size4": 0.5634322067101797,
      "cache_hit_rate_size8": 0.7774784664312998,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.875,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 779.6,
      "completions/mean_length": 972.6,
      "completions/mean_terminated_length": 687.6933380126953,
      "completions/min_length": 624.6,
      "completions/min_terminated_length": 624.6,
      "entropy": 0.672938522696495,
      "epoch": 0.002086811352253756,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.14470750093460083,
      "kl": 0.09736248571425676,
      "learning_rate": 9e-05,
      "loss": 0.03134543597698212,
      "num_tokens": 188840.0,
      "reward": 0.5629066795110702,
      "reward_std": 0.11596517935395241,
      "rewards/cache_reward/mean": 0.5629066795110702,
      "rewards/cache_reward/std": 0.11596517935395241,
      "step": 10,
      "step_time": 98.73753785609733
    },
    {
      "cache_hit_rate": 0.549447825551033,
      "cache_hit_rate_size2": 0.32006935477256776,
      "cache_hit_rate_size4": 0.5575728058815003,
      "cache_hit_rate_size8": 0.7875463843345643,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.83125,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 636.3,
      "completions/mean_length": 944.1625,
      "completions/mean_terminated_length": 494.7774993896484,
      "completions/min_length": 582.6,
      "completions/min_terminated_length": 377.8,
      "entropy": 0.7167654812335968,
      "epoch": 0.004173622704507512,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.1720157116651535,
      "kl": 0.02486464837566018,
      "learning_rate": 0.0001,
      "loss": 0.003635849803686142,
      "num_tokens": 379194.0,
      "reward": 0.54392811357975,
      "reward_std": 0.10275287721306085,
      "rewards/cache_reward/mean": 0.54392811357975,
      "rewards/cache_reward/std": 0.10275287721306085,
      "step": 20,
      "step_time": 99.00200829666574
    },
    {
      "cache_hit_rate": 0.6325653582811356,
      "cache_hit_rate_size2": 0.35956130425135296,
      "cache_hit_rate_size4": 0.5539245963096618,
      "cache_hit_rate_size8": 0.7761083766818047,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.65625,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 787.7,
      "completions/mean_length": 871.475,
      "completions/mean_terminated_length": 635.165884399414,
      "completions/min_length": 477.4,
      "completions/min_terminated_length": 477.4,
      "entropy": 0.7135540544986725,
      "epoch": 0.006260434056761269,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.173299640417099,
      "kl": 0.024446052312850953,
      "learning_rate": 0.0001,
      "loss": 0.029578784108161928,
      "num_tokens": 562198.0,
      "reward": 0.64033243060112,
      "reward_std": 0.09460249152034521,
      "rewards/cache_reward/mean": 0.64033243060112,
      "rewards/cache_reward/std": 0.09460249152034521,
      "step": 30,
      "step_time": 98.59192049612757
    },
    {
      "cache_hit_rate": 0.5263916224241256,
      "cache_hit_rate_size2": 0.3331736276547114,
      "cache_hit_rate_size4": 0.5371630617550441,
      "cache_hit_rate_size8": 0.7860376387834549,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.4625,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 821.3,
      "completions/mean_length": 751.075,
      "completions/mean_terminated_length": 512.4199432373047,
      "completions/min_length": 246.3,
      "completions/min_terminated_length": 246.3,
      "entropy": 0.781426876783371,
      "epoch": 0.008347245409015025,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.17043574154376984,
      "kl": 0.03577139470726252,
      "learning_rate": 0.0001,
      "loss": 0.008154835551977158,
      "num_tokens": 720338.0,
      "reward": 0.5283390432596207,
      "reward_std": 0.10703168977051973,
      "rewards/cache_reward/mean": 0.5283390432596207,
      "rewards/cache_reward/std": 0.10703168977051973,
      "step": 40,
      "step_time": 98.32788651273586
    },
    {
      "cache_hit_rate": 0.585143169760704,
      "cache_hit_rate_size2": 0.33264798671007156,
      "cache_hit_rate_size4": 0.5347203969955444,
      "cache_hit_rate_size8": 0.8004282712936401,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.68125,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 750.5,
      "completions/mean_length": 856.025,
      "completions/mean_terminated_length": 480.12476806640626,
      "completions/min_length": 361.6,
      "completions/min_terminated_length": 259.2,
      "entropy": 0.83377685546875,
      "epoch": 0.010434056761268781,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.9775862097740173,
      "kl": 0.6444699574261904,
      "learning_rate": 0.0001,
      "loss": 0.009068857133388519,
      "num_tokens": 899302.0,
      "reward": 0.5672911822795867,
      "reward_std": 0.17889709547162055,
      "rewards/cache_reward/mean": 0.5672911822795867,
      "rewards/cache_reward/std": 0.17889709547162055,
      "step": 50,
      "step_time": 99.10694038907532
    },
    {
      "cache_hit_rate": 0.467166006565094,
      "cache_hit_rate_size2": 0.36867357624901664,
      "cache_hit_rate_size4": 0.519642561674118,
      "cache_hit_rate_size8": 0.8164827227592468,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.7375,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 793.2,
      "completions/mean_length": 896.31875,
      "completions/mean_terminated_length": 569.6188934326171,
      "completions/min_length": 324.9,
      "completions/min_terminated_length": 324.9,
      "entropy": 0.7401006281375885,
      "epoch": 0.012520868113522538,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.18790468573570251,
      "kl": 1.2457976788282394,
      "learning_rate": 0.0001,
      "loss": 0.006084223091602325,
      "num_tokens": 1091545.0,
      "reward": 0.46716995537281036,
      "reward_std": 0.10014287233352662,
      "rewards/cache_reward/mean": 0.46716995537281036,
      "rewards/cache_reward/std": 0.10014287233352662,
      "step": 60,
      "step_time": 99.14377401731909
    },
    {
      "cache_hit_rate": 0.6927748620510101,
      "cache_hit_rate_size2": 0.38702807823816937,
      "cache_hit_rate_size4": 0.5830764522155126,
      "cache_hit_rate_size8": 0.8190080655945672,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.825,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 671.4,
      "completions/mean_length": 919.4375,
      "completions/mean_terminated_length": 488.41667022705076,
      "completions/min_length": 336.7,
      "completions/min_terminated_length": 336.7,
      "entropy": 0.8826560318470001,
      "epoch": 0.014607679465776294,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.9790771007537842,
      "kl": 0.11715116128325462,
      "learning_rate": 0.0001,
      "loss": 0.02142851650714874,
      "num_tokens": 1281711.0,
      "reward": 0.6808762818574905,
      "reward_std": 0.13245124705135822,
      "rewards/cache_reward/mean": 0.6808762818574905,
      "rewards/cache_reward/std": 0.13245124705135822,
      "step": 70,
      "step_time": 99.350616218918
    },
    {
      "cache_hit_rate": 0.5300061672925949,
      "cache_hit_rate_size2": 0.3520415276288986,
      "cache_hit_rate_size4": 0.618669467312949,
      "cache_hit_rate_size8": 0.8320682446161906,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.8875,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 517.4,
      "completions/mean_length": 966.375,
      "completions/mean_terminated_length": 425.7666717529297,
      "completions/min_length": 441.4,
      "completions/min_terminated_length": 339.0,
      "entropy": 0.9704236328601837,
      "epoch": 0.01669449081803005,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.17933516204357147,
      "kl": 0.1078528642654419,
      "learning_rate": 0.0001,
      "loss": 0.0028869673609733582,
      "num_tokens": 1482467.0,
      "reward": 0.5285103291273117,
      "reward_std": 0.14326442833989858,
      "rewards/cache_reward/mean": 0.5285103291273117,
      "rewards/cache_reward/std": 0.14326442833989858,
      "step": 80,
      "step_time": 99.17187829778996
    },
    {
      "cache_hit_rate": 0.5446671068668365,
      "cache_hit_rate_size2": 0.39704080138887676,
      "cache_hit_rate_size4": 0.5744284888108572,
      "cache_hit_rate_size8": 0.7754806280136108,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.89375,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 348.8,
      "completions/mean_length": 977.05,
      "completions/mean_terminated_length": 268.01905517578126,
      "completions/min_length": 720.2,
      "completions/min_terminated_length": 208.2,
      "entropy": 0.9227084219455719,
      "epoch": 0.018781302170283808,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2052646279335022,
      "kl": 3.135468265414238,
      "learning_rate": 0.0001,
      "loss": 0.12404063940048218,
      "num_tokens": 1693243.0,
      "reward": 0.5458087146282196,
      "reward_std": 0.09762927331030369,
      "rewards/cache_reward/mean": 0.5458087146282196,
      "rewards/cache_reward/std": 0.09762927331030369,
      "step": 90,
      "step_time": 99.82454825101885
    },
    {
      "cache_hit_rate": 0.620132327079773,
      "cache_hit_rate_size2": 0.39693122506141665,
      "cache_hit_rate_size4": 0.6539293527603149,
      "cache_hit_rate_size8": 0.8750839233398438,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.95,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 172.1,
      "completions/mean_length": 995.375,
      "completions/mean_terminated_length": 114.43333435058594,
      "completions/min_length": 800.8,
      "completions/min_terminated_length": 84.0,
      "entropy": 1.0687272012233735,
      "epoch": 0.020868113522537562,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.1818697154521942,
      "kl": 0.11797114461660385,
      "learning_rate": 0.0001,
      "loss": -0.013124784827232361,
      "num_tokens": 1898175.0,
      "reward": 0.6175141602754592,
      "reward_std": 0.11044000014662743,
      "rewards/cache_reward/mean": 0.6175141602754592,
      "rewards/cache_reward/std": 0.11044000014662743,
      "step": 100,
      "step_time": 98.91498928379733
    },
    {
      "cache_hit_rate": 0.7145757257938385,
      "cache_hit_rate_size2": 0.400341796875,
      "cache_hit_rate_size4": 0.6684773763020834,
      "cache_hit_rate_size8": 0.8589237199889289,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.975,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 84.5,
      "completions/mean_length": 1004.74375,
      "completions/mean_terminated_length": 50.75,
      "completions/min_length": 836.2,
      "completions/min_terminated_length": 17.0,
      "entropy": 1.1998688220977782,
      "epoch": 0.022954924874791317,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.1931397169828415,
      "kl": 0.1271144889295101,
      "learning_rate": 0.0001,
      "loss": -0.01696862280368805,
      "num_tokens": 2105638.0,
      "reward": 0.7181334078311921,
      "reward_std": 0.1537357272580266,
      "rewards/cache_reward/mean": 0.7181334078311921,
      "rewards/cache_reward/std": 0.1537357272580266,
      "step": 110,
      "step_time": 98.86106825817842
    },
    {
      "cache_hit_rate": 0.6297692149877548,
      "cache_hit_rate_size2": 0.39751855134963987,
      "cache_hit_rate_size4": 0.6611618757247925,
      "cache_hit_rate_size8": 0.8319511016209921,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.94375,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 155.1,
      "completions/mean_length": 987.025,
      "completions/mean_terminated_length": 120.28333740234375,
      "completions/min_length": 693.4,
      "completions/min_terminated_length": 79.0,
      "entropy": 1.1744155049324037,
      "epoch": 0.025041736227045076,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.1738438606262207,
      "kl": 0.44582032933831217,
      "learning_rate": 0.0001,
      "loss": 0.008387177437543868,
      "num_tokens": 2307794.0,
      "reward": 0.6293900430202484,
      "reward_std": 0.12351493388414383,
      "rewards/cache_reward/mean": 0.6293900430202484,
      "rewards/cache_reward/std": 0.12351493388414383,
      "step": 120,
      "step_time": 98.64420467796735
    },
    {
      "cache_hit_rate": 0.6638946533203125,
      "cache_hit_rate_size2": 0.4562530517578125,
      "cache_hit_rate_size4": 0.7089080810546875,
      "cache_hit_rate_size8": 0.8859303792317709,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 1024.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 1024.0,
      "completions/min_terminated_length": 0.0,
      "entropy": 1.0602934896945952,
      "epoch": 0.02712854757929883,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.17686830461025238,
      "kl": 0.08731514140963555,
      "learning_rate": 0.0001,
      "loss": 0.0034926038235425947,
      "num_tokens": 2514962.0,
      "reward": 0.6638946533203125,
      "reward_std": 0.11878685541450977,
      "rewards/cache_reward/mean": 0.6638946533203125,
      "rewards/cache_reward/std": 0.11878685541450977,
      "step": 130,
      "step_time": 97.97791426731274
    },
    {
      "cache_hit_rate": 0.7039935708045959,
      "cache_hit_rate_size2": 0.4619227818080357,
      "cache_hit_rate_size4": 0.64324951171875,
      "cache_hit_rate_size8": 0.9138779789209366,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.99375,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 13.5,
      "completions/mean_length": 1018.44375,
      "completions/mean_terminated_length": 13.5,
      "completions/min_length": 935.1,
      "completions/min_terminated_length": 13.5,
      "entropy": 1.2289878249168396,
      "epoch": 0.02921535893155259,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.13990317285060883,
      "kl": 0.09567450881004333,
      "learning_rate": 0.0001,
      "loss": -0.007647792994976044,
      "num_tokens": 2714961.0,
      "reward": 0.7043542921543121,
      "reward_std": 0.18686594553291797,
      "rewards/cache_reward/mean": 0.7043542921543121,
      "rewards/cache_reward/std": 0.18686594553291797,
      "step": 140,
      "step_time": 98.28846779624
    },
    {
      "cache_hit_rate": 0.712518310546875,
      "cache_hit_rate_size2": 0.5248336791992188,
      "cache_hit_rate_size4": 0.74189453125,
      "cache_hit_rate_size8": 0.9121348063151041,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 1024.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 1024.0,
      "completions/min_terminated_length": 0.0,
      "entropy": 1.3492182612419128,
      "epoch": 0.031302170283806344,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.1376185417175293,
      "kl": 0.0961605504155159,
      "learning_rate": 0.0001,
      "loss": 0.00384642668068409,
      "num_tokens": 2917729.0,
      "reward": 0.712518310546875,
      "reward_std": 0.15230192057788372,
      "rewards/cache_reward/mean": 0.712518310546875,
      "rewards/cache_reward/std": 0.15230192057788372,
      "step": 150,
      "step_time": 98.02246708897874
    },
    {
      "cache_hit_rate": 0.7198211669921875,
      "cache_hit_rate_size2": 0.5438995361328125,
      "cache_hit_rate_size4": 0.81097412109375,
      "cache_hit_rate_size8": 0.8897552490234375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 1024.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 1024.0,
      "completions/min_terminated_length": 0.0,
      "entropy": 1.566550862789154,
      "epoch": 0.0333889816360601,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.1604277342557907,
      "kl": 0.08198406063020229,
      "learning_rate": 0.0001,
      "loss": 0.0032793622463941573,
      "num_tokens": 3123345.0,
      "reward": 0.7198211669921875,
      "reward_std": 0.12151376903057098,
      "rewards/cache_reward/mean": 0.7198211669921875,
      "rewards/cache_reward/std": 0.12151376903057098,
      "step": 160,
      "step_time": 98.34183804546483
    },
    {
      "cache_hit_rate": 0.81680908203125,
      "cache_hit_rate_size2": 0.5521647135416666,
      "cache_hit_rate_size4": 0.842071533203125,
      "cache_hit_rate_size8": 0.894195556640625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 1024.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 1024.0,
      "completions/min_terminated_length": 0.0,
      "entropy": 1.6380006790161132,
      "epoch": 0.03547579298831385,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.12700115144252777,
      "kl": 0.08730730786919594,
      "learning_rate": 0.0001,
      "loss": 0.003492293506860733,
      "num_tokens": 3319857.0,
      "reward": 0.81680908203125,
      "reward_std": 0.10155810248106718,
      "rewards/cache_reward/mean": 0.81680908203125,
      "rewards/cache_reward/std": 0.10155810248106718,
      "step": 170,
      "step_time": 97.57386728890705
    },
    {
      "cache_hit_rate": 0.8318297088146209,
      "cache_hit_rate_size2": 0.6691019733746847,
      "cache_hit_rate_size4": 0.87476806640625,
      "cache_hit_rate_size8": 0.9261807998021444,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.9875,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 56.9,
      "completions/mean_length": 1014.88125,
      "completions/mean_terminated_length": 29.45,
      "completions/min_length": 923.6,
      "completions/min_terminated_length": 2.0,
      "entropy": 1.5399263143539428,
      "epoch": 0.037562604340567615,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.1375689059495926,
      "kl": 0.07477746270596981,
      "learning_rate": 0.0001,
      "loss": -0.010156913101673127,
      "num_tokens": 3523726.0,
      "reward": 0.8308217704296113,
      "reward_std": 0.10024527162313461,
      "rewards/cache_reward/mean": 0.8308217704296113,
      "rewards/cache_reward/std": 0.10024527162313461,
      "step": 180,
      "step_time": 98.64892102256417
    },
    {
      "cache_hit_rate": 0.8467194437980652,
      "cache_hit_rate_size2": 0.7021891276041666,
      "cache_hit_rate_size4": 0.8873421805245536,
      "cache_hit_rate_size8": 0.9641022086143494,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.99375,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 23.1,
      "completions/mean_length": 1019.04375,
      "completions/mean_terminated_length": 23.1,
      "completions/min_length": 944.7,
      "completions/min_terminated_length": 23.1,
      "entropy": 1.2291082501411439,
      "epoch": 0.03964941569282137,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.1418466866016388,
      "kl": 0.07442079186439514,
      "learning_rate": 0.0001,
      "loss": 0.005592674762010574,
      "num_tokens": 3716957.0,
      "reward": 0.8468945384025574,
      "reward_std": 0.10307033061981201,
      "rewards/cache_reward/mean": 0.8468945384025574,
      "rewards/cache_reward/std": 0.10307033061981201,
      "step": 190,
      "step_time": 98.52997402758338
    },
    {
      "cache_hit_rate": 0.834912109375,
      "cache_hit_rate_size2": 0.6823781331380209,
      "cache_hit_rate_size4": 0.9096527099609375,
      "cache_hit_rate_size8": 0.9377593994140625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 1024.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 1024.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 1024.0,
      "completions/min_terminated_length": 0.0,
      "entropy": 1.6369535088539124,
      "epoch": 0.041736227045075125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.10621531307697296,
      "kl": 0.07659099735319615,
      "learning_rate": 0.0001,
      "loss": 0.003063639998435974,
      "num_tokens": 3934149.0,
      "reward": 0.834912109375,
      "reward_std": 0.10910224989056587,
      "rewards/cache_reward/mean": 0.834912109375,
      "rewards/cache_reward/std": 0.10910224989056587,
      "step": 200,
      "step_time": 99.38287101071327
    }
  ],
  "logging_steps": 10,
  "max_steps": 200,
  "num_input_tokens_seen": 3934149,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}