{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.25,
  "eval_steps": 500,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "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": 0.0,
      "completions/max_length": 1080.0,
      "completions/max_terminated_length": 1080.0,
      "completions/mean_length": 414.15625,
      "completions/mean_terminated_length": 414.15625,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0025,
      "frac_reward_zero_std": 0.125,
      "grad_norm": 6.869231224060059,
      "learning_rate": 0.0,
      "loss": 0.0,
      "num_tokens": 120978.0,
      "reward": 0.16171874105930328,
      "reward_std": 0.29198843240737915,
      "rewards/format_reward/mean": 0.3671875,
      "rewards/format_reward/std": 0.23974503576755524,
      "rewards/mcq_exact_match_reward/mean": 0.125,
      "rewards/mcq_exact_match_reward/std": 0.3333333432674408,
      "step": 1
    },
    {
      "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": 0.0,
      "completions/max_length": 1573.0,
      "completions/max_terminated_length": 1573.0,
      "completions/mean_length": 449.25,
      "completions/mean_terminated_length": 449.25,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.005,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 8.82888412475586,
      "learning_rate": 2e-07,
      "loss": -0.0,
      "num_tokens": 233058.0,
      "reward": 0.21171875298023224,
      "reward_std": 0.28466087579727173,
      "rewards/format_reward/mean": 0.3984375,
      "rewards/format_reward/std": 0.28423789143562317,
      "rewards/mcq_exact_match_reward/mean": 0.171875,
      "rewards/mcq_exact_match_reward/std": 0.38025420904159546,
      "step": 2
    },
    {
      "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": 0.0,
      "completions/max_length": 1528.0,
      "completions/max_terminated_length": 1528.0,
      "completions/mean_length": 486.359375,
      "completions/mean_terminated_length": 486.359375,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0075,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 5.6109819412231445,
      "learning_rate": 4e-07,
      "loss": 0.0,
      "num_tokens": 345873.0,
      "reward": 0.3984374701976776,
      "reward_std": 0.3455878794193268,
      "rewards/format_reward/mean": 0.390625,
      "rewards/format_reward/std": 0.22658175230026245,
      "rewards/mcq_exact_match_reward/mean": 0.359375,
      "rewards/mcq_exact_match_reward/std": 0.4836103618144989,
      "step": 3
    },
    {
      "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": 0.0,
      "completions/max_length": 1972.0,
      "completions/max_terminated_length": 1972.0,
      "completions/mean_length": 600.75,
      "completions/mean_terminated_length": 600.75,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.01,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 8.43229866027832,
      "learning_rate": 6e-07,
      "loss": -0.0,
      "num_tokens": 473249.0,
      "reward": 0.3023437261581421,
      "reward_std": 0.3486832082271576,
      "rewards/format_reward/mean": 0.3671875,
      "rewards/format_reward/std": 0.28510910272598267,
      "rewards/mcq_exact_match_reward/mean": 0.265625,
      "rewards/mcq_exact_match_reward/std": 0.44515693187713623,
      "step": 4
    },
    {
      "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": 0.015625,
      "completions/max_length": 2048.0,
      "completions/max_terminated_length": 1276.0,
      "completions/mean_length": 553.796875,
      "completions/mean_terminated_length": 530.0794067382812,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 4.914008140563965,
      "learning_rate": 8e-07,
      "loss": 0.0,
      "num_tokens": 610444.0,
      "reward": 0.15859374403953552,
      "reward_std": 0.2671620845794678,
      "rewards/format_reward/mean": 0.3359375,
      "rewards/format_reward/std": 0.25283610820770264,
      "rewards/mcq_exact_match_reward/mean": 0.125,
      "rewards/mcq_exact_match_reward/std": 0.3333333432674408,
      "step": 5
    },
    {
      "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": 0.0,
      "completions/max_length": 1599.0,
      "completions/max_terminated_length": 1599.0,
      "completions/mean_length": 421.078125,
      "completions/mean_terminated_length": 421.078125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.015,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 8.73421859741211,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 738681.0,
      "reward": 0.1953125,
      "reward_std": 0.32575029134750366,
      "rewards/format_reward/mean": 0.390625,
      "rewards/format_reward/std": 0.22658175230026245,
      "rewards/mcq_exact_match_reward/mean": 0.15625,
      "rewards/mcq_exact_match_reward/std": 0.36596253514289856,
      "step": 6
    },
    {
      "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": 0.0625,
      "completions/max_length": 2048.0,
      "completions/max_terminated_length": 1297.0,
      "completions/mean_length": 559.015625,
      "completions/mean_terminated_length": 459.7500305175781,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0175,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 7.960887908935547,
      "learning_rate": 9.99726628670463e-07,
      "loss": 0.0,
      "num_tokens": 877346.0,
      "reward": 0.29218748211860657,
      "reward_std": 0.32155841588974,
      "rewards/format_reward/mean": 0.421875,
      "rewards/format_reward/std": 0.23935678601264954,
      "rewards/mcq_exact_match_reward/mean": 0.25,
      "rewards/mcq_exact_match_reward/std": 0.4364357888698578,
      "step": 7
    },
    {
      "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": 0.0,
      "completions/max_length": 1659.0,
      "completions/max_terminated_length": 1659.0,
      "completions/mean_length": 514.421875,
      "completions/mean_terminated_length": 514.421875,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.02,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 15.752873420715332,
      "learning_rate": 9.989068136093872e-07,
      "loss": 0.0,
      "num_tokens": 999109.0,
      "reward": 0.29218751192092896,
      "reward_std": 0.2708819806575775,
      "rewards/format_reward/mean": 0.421875,
      "rewards/format_reward/std": 0.23935678601264954,
      "rewards/mcq_exact_match_reward/mean": 0.25,
      "rewards/mcq_exact_match_reward/std": 0.4364357888698578,
      "step": 8
    },
    {
      "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": 0.0,
      "completions/max_length": 1567.0,
      "completions/max_terminated_length": 1567.0,
      "completions/mean_length": 518.625,
      "completions/mean_terminated_length": 518.625,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0225,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 9.001887321472168,
      "learning_rate": 9.975414512725056e-07,
      "loss": 0.0,
      "num_tokens": 1124677.0,
      "reward": 0.29218748211860657,
      "reward_std": 0.29666516184806824,
      "rewards/format_reward/mean": 0.421875,
      "rewards/format_reward/std": 0.23935678601264954,
      "rewards/mcq_exact_match_reward/mean": 0.25,
      "rewards/mcq_exact_match_reward/std": 0.4364357888698578,
      "step": 9
    },
    {
      "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": 0.0,
      "completions/max_length": 1335.0,
      "completions/max_terminated_length": 1335.0,
      "completions/mean_length": 405.203125,
      "completions/mean_terminated_length": 405.203125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.025,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 20.01106071472168,
      "learning_rate": 9.956320346634875e-07,
      "loss": -0.0,
      "num_tokens": 1257138.0,
      "reward": 0.45937496423721313,
      "reward_std": 0.3208349049091339,
      "rewards/format_reward/mean": 0.375,
      "rewards/format_reward/std": 0.2182178944349289,
      "rewards/mcq_exact_match_reward/mean": 0.421875,
      "rewards/mcq_exact_match_reward/std": 0.49776285886764526,
      "step": 10
    },
    {
      "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": 0.0,
      "completions/max_length": 1375.0,
      "completions/max_terminated_length": 1375.0,
      "completions/mean_length": 256.203125,
      "completions/mean_terminated_length": 256.203125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0275,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 16.7104434967041,
      "learning_rate": 9.931806517013612e-07,
      "loss": -0.0,
      "num_tokens": 1357975.0,
      "reward": 0.3453124761581421,
      "reward_std": 0.3788047730922699,
      "rewards/format_reward/mean": 0.484375,
      "rewards/format_reward/std": 0.2514837086200714,
      "rewards/mcq_exact_match_reward/mean": 0.296875,
      "rewards/mcq_exact_match_reward/std": 0.4604927599430084,
      "step": 11
    },
    {
      "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": 0.0,
      "completions/max_length": 1265.0,
      "completions/max_terminated_length": 1265.0,
      "completions/mean_length": 127.234375,
      "completions/mean_terminated_length": 127.234375,
      "completions/min_length": 2.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.03,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 55.890071868896484,
      "learning_rate": 9.901899829374047e-07,
      "loss": -0.0,
      "num_tokens": 1466070.0,
      "reward": 0.48515623807907104,
      "reward_std": 0.2358182966709137,
      "rewards/format_reward/mean": 0.4765625,
      "rewards/format_reward/std": 0.10652101784944534,
      "rewards/mcq_exact_match_reward/mean": 0.4375,
      "rewards/mcq_exact_match_reward/std": 0.5,
      "step": 12
    },
    {
      "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": 0.0,
      "completions/max_length": 721.0,
      "completions/max_terminated_length": 721.0,
      "completions/mean_length": 103.8125,
      "completions/mean_terminated_length": 103.8125,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.0325,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 23.41650390625,
      "learning_rate": 9.866632986240029e-07,
      "loss": -0.0,
      "num_tokens": 1548482.0,
      "reward": 0.22109374403953552,
      "reward_std": 0.25605273246765137,
      "rewards/format_reward/mean": 0.4921875,
      "rewards/format_reward/std": 0.1406387835741043,
      "rewards/mcq_exact_match_reward/mean": 0.171875,
      "rewards/mcq_exact_match_reward/std": 0.38025420904159546,
      "step": 13
    },
    {
      "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": 0.015625,
      "completions/max_length": 2048.0,
      "completions/max_terminated_length": 481.0,
      "completions/mean_length": 74.25,
      "completions/mean_terminated_length": 42.92063903808594,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.035,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 0.2323995977640152,
      "learning_rate": 9.826044551386742e-07,
      "loss": 0.0,
      "num_tokens": 1667746.0,
      "reward": 0.42421871423721313,
      "reward_std": 0.00220970856025815,
      "rewards/format_reward/mean": 0.4921875,
      "rewards/format_reward/std": 0.0625,
      "rewards/mcq_exact_match_reward/mean": 0.375,
      "rewards/mcq_exact_match_reward/std": 0.48795005679130554,
      "step": 14
    },
    {
      "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": 0.0,
      "completions/max_length": 1505.0,
      "completions/max_terminated_length": 1505.0,
      "completions/mean_length": 157.109375,
      "completions/mean_terminated_length": 157.109375,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0375,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 19.204160690307617,
      "learning_rate": 9.780178907671788e-07,
      "loss": 0.0,
      "num_tokens": 1751393.0,
      "reward": 0.4140624701976776,
      "reward_std": 0.36057960987091064,
      "rewards/format_reward/mean": 0.546875,
      "rewards/format_reward/std": 0.14689241349697113,
      "rewards/mcq_exact_match_reward/mean": 0.359375,
      "rewards/mcq_exact_match_reward/std": 0.4836103618144989,
      "step": 15
    },
    {
      "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": 0.0,
      "completions/max_length": 765.0,
      "completions/max_terminated_length": 765.0,
      "completions/mean_length": 47.9375,
      "completions/mean_terminated_length": 47.9375,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.04,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 12.51365852355957,
      "learning_rate": 9.729086208503173e-07,
      "loss": -0.0,
      "num_tokens": 1835085.0,
      "reward": 0.26874998211860657,
      "reward_std": 0.2774261236190796,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.08908708393573761,
      "rewards/mcq_exact_match_reward/mean": 0.21875,
      "rewards/mcq_exact_match_reward/std": 0.4166666865348816,
      "step": 16
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0425,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 13.823043823242188,
      "learning_rate": 9.672822322997304e-07,
      "loss": -0.0,
      "num_tokens": 1928933.0,
      "reward": 0.17499999701976776,
      "reward_std": 0.17570310831069946,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.125,
      "rewards/mcq_exact_match_reward/std": 0.3333333432674408,
      "step": 17
    },
    {
      "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": 0.0,
      "completions/max_length": 74.0,
      "completions/max_terminated_length": 74.0,
      "completions/mean_length": 7.0625,
      "completions/mean_terminated_length": 7.0625,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.045,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 10.003605842590332,
      "learning_rate": 9.611448774886923e-07,
      "loss": 0.0,
      "num_tokens": 2027129.0,
      "reward": 0.20624998211860657,
      "reward_std": 0.0883883386850357,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.15625,
      "rewards/mcq_exact_match_reward/std": 0.36596253514289856,
      "step": 18
    },
    {
      "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": 0.0,
      "completions/max_length": 461.0,
      "completions/max_terminated_length": 461.0,
      "completions/mean_length": 13.109375,
      "completions/mean_terminated_length": 13.109375,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0475,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 12.742534637451172,
      "learning_rate": 9.545032675245813e-07,
      "loss": -0.0,
      "num_tokens": 2123488.0,
      "reward": 0.25312498211860657,
      "reward_std": 0.11100947111845016,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.203125,
      "rewards/mcq_exact_match_reward/std": 0.40550529956817627,
      "step": 19
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.05,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 12.020237922668457,
      "learning_rate": 9.473646649103817e-07,
      "loss": 0.0,
      "num_tokens": 2223872.0,
      "reward": 0.5499999523162842,
      "reward_std": 0.0883883386850357,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.5,
      "rewards/mcq_exact_match_reward/std": 0.5039526224136353,
      "step": 20
    },
    {
      "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": 0.0,
      "completions/max_length": 94.0,
      "completions/max_terminated_length": 94.0,
      "completions/mean_length": 7.375,
      "completions/mean_terminated_length": 7.375,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0525,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 15.074679374694824,
      "learning_rate": 9.397368756032444e-07,
      "loss": -0.0,
      "num_tokens": 2317880.0,
      "reward": 0.40937498211860657,
      "reward_std": 0.16887322068214417,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.359375,
      "rewards/mcq_exact_match_reward/std": 0.4836103618144989,
      "step": 21
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.055,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 15.53366470336914,
      "learning_rate": 9.316282404787869e-07,
      "loss": -0.0,
      "num_tokens": 2398664.0,
      "reward": 0.40937498211860657,
      "reward_std": 0.16887322068214417,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.359375,
      "rewards/mcq_exact_match_reward/std": 0.4836103618144989,
      "step": 22
    },
    {
      "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": 0.0,
      "completions/max_length": 15.0,
      "completions/max_terminated_length": 15.0,
      "completions/mean_length": 6.28125,
      "completions/mean_terminated_length": 6.28125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0575,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 18.223918914794922,
      "learning_rate": 9.230476262104676e-07,
      "loss": -0.0,
      "num_tokens": 2485018.0,
      "reward": 0.40625,
      "reward_std": 0.2714304029941559,
      "rewards/format_reward/mean": 0.46875,
      "rewards/format_reward/std": 0.12198751419782639,
      "rewards/mcq_exact_match_reward/mean": 0.359375,
      "rewards/mcq_exact_match_reward/std": 0.4836103618144989,
      "step": 23
    },
    {
      "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": 0.0,
      "completions/max_length": 285.0,
      "completions/max_terminated_length": 285.0,
      "completions/mean_length": 10.359375,
      "completions/mean_terminated_length": 10.359375,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.06,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 20.070579528808594,
      "learning_rate": 9.1400441557401e-07,
      "loss": -0.0,
      "num_tokens": 2573769.0,
      "reward": 0.6078124642372131,
      "reward_std": 0.19909577071666718,
      "rewards/format_reward/mean": 0.453125,
      "rewards/format_reward/std": 0.14689241349697113,
      "rewards/mcq_exact_match_reward/mean": 0.5625,
      "rewards/mcq_exact_match_reward/std": 0.5,
      "step": 24
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0625,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 29.562118530273438,
      "learning_rate": 9.045084971874737e-07,
      "loss": -0.0,
      "num_tokens": 2676177.0,
      "reward": 0.5031249523162842,
      "reward_std": 0.29826050996780396,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.453125,
      "rewards/mcq_exact_match_reward/std": 0.501733124256134,
      "step": 25
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.065,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 13.510916709899902,
      "learning_rate": 8.945702546981968e-07,
      "loss": -0.0,
      "num_tokens": 2760681.0,
      "reward": 0.5031249523162842,
      "reward_std": 0.16887323558330536,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.453125,
      "rewards/mcq_exact_match_reward/std": 0.501733124256134,
      "step": 26
    },
    {
      "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": 0.0,
      "completions/max_length": 230.0,
      "completions/max_terminated_length": 230.0,
      "completions/mean_length": 9.5,
      "completions/mean_terminated_length": 9.5,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0675,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 17.294832229614258,
      "learning_rate": 8.842005554284295e-07,
      "loss": -0.0,
      "num_tokens": 2846193.0,
      "reward": 0.42499998211860657,
      "reward_std": 0.2041158527135849,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.375,
      "rewards/mcq_exact_match_reward/std": 0.48795005679130554,
      "step": 27
    },
    {
      "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": 0.0,
      "completions/max_length": 96.0,
      "completions/max_terminated_length": 96.0,
      "completions/mean_length": 7.40625,
      "completions/mean_terminated_length": 7.40625,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.07,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 17.21614646911621,
      "learning_rate": 8.734107384920769e-07,
      "loss": -0.0,
      "num_tokens": 2948131.0,
      "reward": 0.31562498211860657,
      "reward_std": 0.16887322068214417,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.265625,
      "rewards/mcq_exact_match_reward/std": 0.44515693187713623,
      "step": 28
    },
    {
      "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": 0.0,
      "completions/max_length": 604.0,
      "completions/max_terminated_length": 604.0,
      "completions/mean_length": 22.453125,
      "completions/mean_terminated_length": 22.453125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0725,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 28.476659774780273,
      "learning_rate": 8.622126023955445e-07,
      "loss": -0.0,
      "num_tokens": 3048088.0,
      "reward": 0.4406249523162842,
      "reward_std": 0.34246447682380676,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.390625,
      "rewards/mcq_exact_match_reward/std": 0.4917473793029785,
      "step": 29
    },
    {
      "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": 0.0,
      "completions/max_length": 11.0,
      "completions/max_terminated_length": 11.0,
      "completions/mean_length": 6.078125,
      "completions/mean_terminated_length": 6.078125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.075,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 15.044946670532227,
      "learning_rate": 8.506183921362442e-07,
      "loss": -0.0,
      "num_tokens": 3148061.0,
      "reward": 0.5812499523162842,
      "reward_std": 0.12467905879020691,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.53125,
      "rewards/mcq_exact_match_reward/std": 0.5029674172401428,
      "step": 30
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0775,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 11.75574016571045,
      "learning_rate": 8.386407858128706e-07,
      "loss": 0.0,
      "num_tokens": 3248325.0,
      "reward": 0.6906249523162842,
      "reward_std": 0.13258251547813416,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.640625,
      "rewards/mcq_exact_match_reward/std": 0.4836103618144989,
      "step": 31
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.08,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 19.9006404876709,
      "learning_rate": 8.262928807620843e-07,
      "loss": 0.0,
      "num_tokens": 3331157.0,
      "reward": 0.3468749523162842,
      "reward_std": 0.31512531638145447,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.296875,
      "rewards/mcq_exact_match_reward/std": 0.4604927599430084,
      "step": 32
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0825,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 21.77203941345215,
      "learning_rate": 8.135881792367685e-07,
      "loss": -0.0,
      "num_tokens": 3407133.0,
      "reward": 0.6281249523162842,
      "reward_std": 0.2382849007844925,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.578125,
      "rewards/mcq_exact_match_reward/std": 0.49776285886764526,
      "step": 33
    },
    {
      "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": 0.0,
      "completions/max_length": 154.0,
      "completions/max_terminated_length": 154.0,
      "completions/mean_length": 13.828125,
      "completions/mean_terminated_length": 13.828125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.085,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 9.613591194152832,
      "learning_rate": 8.005405736415125e-07,
      "loss": 0.0,
      "num_tokens": 3505250.0,
      "reward": 0.4718749523162842,
      "reward_std": 0.10205791890621185,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.421875,
      "rewards/mcq_exact_match_reward/std": 0.49776285886764526,
      "step": 34
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0875,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 16.91990852355957,
      "learning_rate": 7.871643313414718e-07,
      "loss": -0.0,
      "num_tokens": 3591986.0,
      "reward": 0.37812498211860657,
      "reward_std": 0.11100947111845016,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.328125,
      "rewards/mcq_exact_match_reward/std": 0.4732423722743988,
      "step": 35
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.09,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 18.04999542236328,
      "learning_rate": 7.734740790612136e-07,
      "loss": -0.0,
      "num_tokens": 3668962.0,
      "reward": 0.36249998211860657,
      "reward_std": 0.15520364046096802,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.3125,
      "rewards/mcq_exact_match_reward/std": 0.467176616191864,
      "step": 36
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0925,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 26.474571228027344,
      "learning_rate": 7.594847868906076e-07,
      "loss": -0.0,
      "num_tokens": 3753034.0,
      "reward": 0.6593749523162842,
      "reward_std": 0.12255740165710449,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.609375,
      "rewards/mcq_exact_match_reward/std": 0.4917473793029785,
      "step": 37
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.095,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 5.568378925323486,
      "learning_rate": 7.452117519152541e-07,
      "loss": -0.0,
      "num_tokens": 3879226.0,
      "reward": 0.33124998211860657,
      "reward_std": 0.0578637532889843,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.28125,
      "rewards/mcq_exact_match_reward/std": 0.4531635046005249,
      "step": 38
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0975,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 22.415874481201172,
      "learning_rate": 7.306705814893439e-07,
      "loss": -0.0,
      "num_tokens": 3957258.0,
      "reward": 0.5187499523162842,
      "reward_std": 0.25513991713523865,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.46875,
      "rewards/mcq_exact_match_reward/std": 0.5029674172401428,
      "step": 39
    },
    {
      "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": 0.0,
      "completions/max_length": 78.0,
      "completions/max_terminated_length": 78.0,
      "completions/mean_length": 7.125,
      "completions/mean_terminated_length": 7.125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 19.496747970581055,
      "learning_rate": 7.158771761692464e-07,
      "loss": -0.0,
      "num_tokens": 4064962.0,
      "reward": 0.6437499523162842,
      "reward_std": 0.1735912561416626,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.59375,
      "rewards/mcq_exact_match_reward/std": 0.49501484632492065,
      "step": 40
    },
    {
      "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": 0.0,
      "completions/max_length": 177.0,
      "completions/max_terminated_length": 177.0,
      "completions/mean_length": 8.75,
      "completions/mean_terminated_length": 8.75,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1025,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 21.04681396484375,
      "learning_rate": 7.008477123264847e-07,
      "loss": 0.0,
      "num_tokens": 4160314.0,
      "reward": 0.8624999523162842,
      "reward_std": 0.16675156354904175,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.8125,
      "rewards/mcq_exact_match_reward/std": 0.39339789748191833,
      "step": 41
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.105,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 16.286230087280273,
      "learning_rate": 6.855986244591103e-07,
      "loss": -0.0,
      "num_tokens": 4247658.0,
      "reward": 0.33124998211860657,
      "reward_std": 0.10888782143592834,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.28125,
      "rewards/mcq_exact_match_reward/std": 0.4531635046005249,
      "step": 42
    },
    {
      "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": 0.0,
      "completions/max_length": 88.0,
      "completions/max_terminated_length": 88.0,
      "completions/mean_length": 7.28125,
      "completions/mean_terminated_length": 7.28125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1075,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 15.285970687866211,
      "learning_rate": 6.701465872208216e-07,
      "loss": 0.0,
      "num_tokens": 4329100.0,
      "reward": 0.5499999523162842,
      "reward_std": 0.16675156354904175,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.5,
      "rewards/mcq_exact_match_reward/std": 0.5039526224136353,
      "step": 43
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.11,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 26.266441345214844,
      "learning_rate": 6.545084971874736e-07,
      "loss": -0.0,
      "num_tokens": 4435308.0,
      "reward": 0.6437499523162842,
      "reward_std": 0.17570312321186066,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.59375,
      "rewards/mcq_exact_match_reward/std": 0.49501484632492065,
      "step": 44
    },
    {
      "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": 0.0,
      "completions/max_length": 230.0,
      "completions/max_terminated_length": 230.0,
      "completions/mean_length": 10.578125,
      "completions/mean_terminated_length": 10.578125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1125,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 24.9106502532959,
      "learning_rate": 6.387014543809223e-07,
      "loss": -0.0,
      "num_tokens": 4520425.0,
      "reward": 0.6593749523162842,
      "reward_std": 0.21778541803359985,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.609375,
      "rewards/mcq_exact_match_reward/std": 0.4917473793029785,
      "step": 45
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.115,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 25.48346710205078,
      "learning_rate": 6.227427435703995e-07,
      "loss": -0.0,
      "num_tokens": 4602065.0,
      "reward": 0.5343749523162842,
      "reward_std": 0.2109457552433014,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.484375,
      "rewards/mcq_exact_match_reward/std": 0.5037065148353577,
      "step": 46
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1175,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 5.884969711303711,
      "learning_rate": 6.066498153718734e-07,
      "loss": 0.0,
      "num_tokens": 4671777.0,
      "reward": 0.6906249523162842,
      "reward_std": 0.04419416934251785,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.640625,
      "rewards/mcq_exact_match_reward/std": 0.4836103618144989,
      "step": 47
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.12,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 47.58740997314453,
      "learning_rate": 5.90440267166055e-07,
      "loss": -0.0,
      "num_tokens": 4768801.0,
      "reward": 0.6593749523162842,
      "reward_std": 0.12255740165710449,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.609375,
      "rewards/mcq_exact_match_reward/std": 0.4917473793029785,
      "step": 48
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1225,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 27.00326919555664,
      "learning_rate": 5.741318238559209e-07,
      "loss": -0.0,
      "num_tokens": 4852049.0,
      "reward": 0.5031249523162842,
      "reward_std": 0.2867125868797302,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.453125,
      "rewards/mcq_exact_match_reward/std": 0.501733124256134,
      "step": 49
    },
    {
      "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": 0.0,
      "completions/max_length": 16.0,
      "completions/max_terminated_length": 16.0,
      "completions/mean_length": 6.15625,
      "completions/mean_terminated_length": 6.15625,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.125,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 19.381776809692383,
      "learning_rate": 5.577423184847931e-07,
      "loss": -0.0,
      "num_tokens": 4952019.0,
      "reward": 0.45546871423721313,
      "reward_std": 0.2919526696205139,
      "rewards/format_reward/mean": 0.4921875,
      "rewards/format_reward/std": 0.0625,
      "rewards/mcq_exact_match_reward/mean": 0.40625,
      "rewards/mcq_exact_match_reward/std": 0.49501484632492065,
      "step": 50
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1275,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 25.298965454101562,
      "learning_rate": 5.412896727361662e-07,
      "loss": -0.0,
      "num_tokens": 5060339.0,
      "reward": 0.5343749523162842,
      "reward_std": 0.2198973000049591,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.484375,
      "rewards/mcq_exact_match_reward/std": 0.5037065148353577,
      "step": 51
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.13,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 9.572460174560547,
      "learning_rate": 5.247918773366111e-07,
      "loss": -0.0,
      "num_tokens": 5170843.0,
      "reward": 0.3468749523162842,
      "reward_std": 0.0646936446428299,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.296875,
      "rewards/mcq_exact_match_reward/std": 0.4604927599430084,
      "step": 52
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1325,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 15.549805641174316,
      "learning_rate": 5.082669723831793e-07,
      "loss": 0.0,
      "num_tokens": 5284787.0,
      "reward": 0.23671874403953552,
      "reward_std": 0.14846180379390717,
      "rewards/format_reward/mean": 0.4921875,
      "rewards/format_reward/std": 0.0625,
      "rewards/mcq_exact_match_reward/mean": 0.1875,
      "rewards/mcq_exact_match_reward/std": 0.39339789748191833,
      "step": 53
    },
    {
      "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": 0.0,
      "completions/max_length": 8.0,
      "completions/max_terminated_length": 8.0,
      "completions/mean_length": 6.03125,
      "completions/mean_terminated_length": 6.03125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.135,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 16.681224822998047,
      "learning_rate": 4.917330276168208e-07,
      "loss": 0.0,
      "num_tokens": 5385909.0,
      "reward": 0.31484371423721313,
      "reward_std": 0.13479220867156982,
      "rewards/format_reward/mean": 0.4921875,
      "rewards/format_reward/std": 0.0625,
      "rewards/mcq_exact_match_reward/mean": 0.265625,
      "rewards/mcq_exact_match_reward/std": 0.44515693187713623,
      "step": 54
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1375,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 19.34429359436035,
      "learning_rate": 4.752081226633888e-07,
      "loss": 0.0,
      "num_tokens": 5477261.0,
      "reward": 0.48749998211860657,
      "reward_std": 0.1462520956993103,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.4375,
      "rewards/mcq_exact_match_reward/std": 0.5,
      "step": 55
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.14,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 43.70222091674805,
      "learning_rate": 4.5871032726383385e-07,
      "loss": -0.0,
      "num_tokens": 5564381.0,
      "reward": 0.8781249523162842,
      "reward_std": 0.16887323558330536,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.828125,
      "rewards/mcq_exact_match_reward/std": 0.38025420904159546,
      "step": 56
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1425,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 14.921404838562012,
      "learning_rate": 4.4225768151520694e-07,
      "loss": 0.0,
      "num_tokens": 5680501.0,
      "reward": 0.4249999523162842,
      "reward_std": 0.0883883386850357,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.375,
      "rewards/mcq_exact_match_reward/std": 0.48795005679130554,
      "step": 57
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.145,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 0.0,
      "learning_rate": 4.258681761440789e-07,
      "loss": 0.0,
      "num_tokens": 5784693.0,
      "reward": 0.7999999523162842,
      "reward_std": 0.0,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.75,
      "rewards/mcq_exact_match_reward/std": 0.4364357888698578,
      "step": 58
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1475,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 16.13064956665039,
      "learning_rate": 4.095597328339452e-07,
      "loss": 0.0,
      "num_tokens": 5864573.0,
      "reward": 0.5343749523162842,
      "reward_std": 0.10205791890621185,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.484375,
      "rewards/mcq_exact_match_reward/std": 0.5037065148353577,
      "step": 59
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.15,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 24.483049392700195,
      "learning_rate": 3.9335018462812664e-07,
      "loss": -0.0,
      "num_tokens": 5946637.0,
      "reward": 0.5812499523162842,
      "reward_std": 0.1872510462999344,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.53125,
      "rewards/mcq_exact_match_reward/std": 0.5029674172401428,
      "step": 60
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1525,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 13.92232608795166,
      "learning_rate": 3.772572564296004e-07,
      "loss": 0.0,
      "num_tokens": 6068045.0,
      "reward": 0.6749999523162842,
      "reward_std": 0.0883883386850357,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.625,
      "rewards/mcq_exact_match_reward/std": 0.48795005679130554,
      "step": 61
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.155,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 12.822912216186523,
      "learning_rate": 3.612985456190778e-07,
      "loss": 0.0,
      "num_tokens": 6163925.0,
      "reward": 0.5187499523162842,
      "reward_std": 0.1462520956993103,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.46875,
      "rewards/mcq_exact_match_reward/std": 0.5029674172401428,
      "step": 62
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1575,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 14.21402645111084,
      "learning_rate": 3.454915028125263e-07,
      "loss": -0.0,
      "num_tokens": 6280653.0,
      "reward": 0.4562499523162842,
      "reward_std": 0.0578637532889843,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.40625,
      "rewards/mcq_exact_match_reward/std": 0.49501484632492065,
      "step": 63
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.16,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 50.95706558227539,
      "learning_rate": 3.2985341277917846e-07,
      "loss": 0.0,
      "num_tokens": 6375885.0,
      "reward": 0.3624999523162842,
      "reward_std": 0.2756393849849701,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.3125,
      "rewards/mcq_exact_match_reward/std": 0.467176616191864,
      "step": 64
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1625,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 19.622344970703125,
      "learning_rate": 3.1440137554088953e-07,
      "loss": 0.0,
      "num_tokens": 6479389.0,
      "reward": 0.6437499523162842,
      "reward_std": 0.10888781398534775,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.59375,
      "rewards/mcq_exact_match_reward/std": 0.49501484632492065,
      "step": 65
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.165,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 16.24167823791504,
      "learning_rate": 2.9915228767351535e-07,
      "loss": 0.0,
      "num_tokens": 6565821.0,
      "reward": 0.4562499523162842,
      "reward_std": 0.1293872892856598,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.40625,
      "rewards/mcq_exact_match_reward/std": 0.49501484632492065,
      "step": 66
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1675,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 28.129180908203125,
      "learning_rate": 2.841228238307536e-07,
      "loss": -0.0,
      "num_tokens": 6674269.0,
      "reward": 0.8468749523162842,
      "reward_std": 0.12255740165710449,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.796875,
      "rewards/mcq_exact_match_reward/std": 0.40550529956817627,
      "step": 67
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.17,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 36.77799606323242,
      "learning_rate": 2.6932941851065615e-07,
      "loss": -0.0,
      "num_tokens": 6777461.0,
      "reward": 0.5187499523162842,
      "reward_std": 0.24511480331420898,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.46875,
      "rewards/mcq_exact_match_reward/std": 0.5029674172401428,
      "step": 68
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1725,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 14.383221626281738,
      "learning_rate": 2.547882480847461e-07,
      "loss": 0.0,
      "num_tokens": 6884381.0,
      "reward": 0.4093749523162842,
      "reward_std": 0.1530819833278656,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.359375,
      "rewards/mcq_exact_match_reward/std": 0.4836103618144989,
      "step": 69
    },
    {
      "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": 0.0,
      "completions/max_length": 8.0,
      "completions/max_terminated_length": 8.0,
      "completions/mean_length": 6.03125,
      "completions/mean_terminated_length": 6.03125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.175,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 23.195758819580078,
      "learning_rate": 2.4051521310939254e-07,
      "loss": -0.0,
      "num_tokens": 7002799.0,
      "reward": 0.25312498211860657,
      "reward_std": 0.23144522309303284,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.203125,
      "rewards/mcq_exact_match_reward/std": 0.40550529956817627,
      "step": 70
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1775,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 19.705732345581055,
      "learning_rate": 2.2652592093878665e-07,
      "loss": -0.0,
      "num_tokens": 7095623.0,
      "reward": 0.3468749523162842,
      "reward_std": 0.1530819833278656,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.296875,
      "rewards/mcq_exact_match_reward/std": 0.4604927599430084,
      "step": 71
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.18,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 27.19195556640625,
      "learning_rate": 2.128356686585282e-07,
      "loss": -0.0,
      "num_tokens": 7207847.0,
      "reward": 0.6906249523162842,
      "reward_std": 0.1315089464187622,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.640625,
      "rewards/mcq_exact_match_reward/std": 0.4836103618144989,
      "step": 72
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1825,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 14.726841926574707,
      "learning_rate": 1.9945942635848745e-07,
      "loss": 0.0,
      "num_tokens": 7323079.0,
      "reward": 0.23749998211860657,
      "reward_std": 0.1462520956993103,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.1875,
      "rewards/mcq_exact_match_reward/std": 0.39339789748191833,
      "step": 73
    },
    {
      "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": 0.0,
      "completions/max_length": 36.0,
      "completions/max_terminated_length": 36.0,
      "completions/mean_length": 6.46875,
      "completions/mean_terminated_length": 6.46875,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.185,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 21.405261993408203,
      "learning_rate": 1.8641182076323148e-07,
      "loss": 0.0,
      "num_tokens": 7408869.0,
      "reward": 0.5343749523162842,
      "reward_std": 0.1530819833278656,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.484375,
      "rewards/mcq_exact_match_reward/std": 0.5037065148353577,
      "step": 74
    },
    {
      "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": 0.0,
      "completions/max_length": 171.0,
      "completions/max_terminated_length": 171.0,
      "completions/mean_length": 8.578125,
      "completions/mean_terminated_length": 8.578125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1875,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 20.04102325439453,
      "learning_rate": 1.7370711923791564e-07,
      "loss": -0.0,
      "num_tokens": 7498050.0,
      "reward": 0.5968749523162842,
      "reward_std": 0.1804211586713791,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.546875,
      "rewards/mcq_exact_match_reward/std": 0.501733124256134,
      "step": 75
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.19,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 25.5474796295166,
      "learning_rate": 1.6135921418712955e-07,
      "loss": -0.0,
      "num_tokens": 7595218.0,
      "reward": 0.7218749523162842,
      "reward_std": 0.2109457552433014,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.671875,
      "rewards/mcq_exact_match_reward/std": 0.4732423722743988,
      "step": 76
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1925,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 11.988852500915527,
      "learning_rate": 1.493816078637557e-07,
      "loss": 0.0,
      "num_tokens": 7701818.0,
      "reward": 0.7999999523162842,
      "reward_std": 0.0883883386850357,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.75,
      "rewards/mcq_exact_match_reward/std": 0.4364357888698578,
      "step": 77
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.195,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 20.20587730407715,
      "learning_rate": 1.3778739760445552e-07,
      "loss": -0.0,
      "num_tokens": 7792362.0,
      "reward": 0.3781249523162842,
      "reward_std": 0.22673697769641876,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.328125,
      "rewards/mcq_exact_match_reward/std": 0.4732423722743988,
      "step": 78
    },
    {
      "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": 0.0,
      "completions/max_length": 22.0,
      "completions/max_terminated_length": 22.0,
      "completions/mean_length": 6.3125,
      "completions/mean_terminated_length": 6.3125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.1975,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 37.22200393676758,
      "learning_rate": 1.2658926150792322e-07,
      "loss": -0.0,
      "num_tokens": 7902286.0,
      "reward": 0.5632811784744263,
      "reward_std": 0.17463430762290955,
      "rewards/format_reward/mean": 0.4765625,
      "rewards/format_reward/std": 0.10652101784944534,
      "rewards/mcq_exact_match_reward/mean": 0.515625,
      "rewards/mcq_exact_match_reward/std": 0.5037065148353577,
      "step": 79
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 30.08717155456543,
      "learning_rate": 1.1579944457157059e-07,
      "loss": 0.0,
      "num_tokens": 7996974.0,
      "reward": 0.6906249523162842,
      "reward_std": 0.23144522309303284,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.640625,
      "rewards/mcq_exact_match_reward/std": 0.4836103618144989,
      "step": 80
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2025,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 6.288850784301758,
      "learning_rate": 1.0542974530180327e-07,
      "loss": -0.0,
      "num_tokens": 8092638.0,
      "reward": 0.6593749523162842,
      "reward_std": 0.04419417306780815,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.609375,
      "rewards/mcq_exact_match_reward/std": 0.4917473793029785,
      "step": 81
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.205,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 22.16109275817871,
      "learning_rate": 9.549150281252632e-08,
      "loss": -0.0,
      "num_tokens": 8195990.0,
      "reward": 0.37812498211860657,
      "reward_std": 0.15992167592048645,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.328125,
      "rewards/mcq_exact_match_reward/std": 0.4732423722743988,
      "step": 82
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2075,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 55.5490608215332,
      "learning_rate": 8.599558442598998e-08,
      "loss": -0.0,
      "num_tokens": 8293598.0,
      "reward": 0.7999999523162842,
      "reward_std": 0.23356686532497406,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.75,
      "rewards/mcq_exact_match_reward/std": 0.4364357888698578,
      "step": 83
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.21,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 4.853342533111572,
      "learning_rate": 7.695237378953224e-08,
      "loss": 0.0,
      "num_tokens": 8380158.0,
      "reward": 0.9406249523162842,
      "reward_std": 0.04419416934251785,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.890625,
      "rewards/mcq_exact_match_reward/std": 0.3145764470100403,
      "step": 84
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2125,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 24.126224517822266,
      "learning_rate": 6.837175952121304e-08,
      "loss": -0.0,
      "num_tokens": 8501366.0,
      "reward": 0.5031249523162842,
      "reward_std": 0.0646936446428299,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.453125,
      "rewards/mcq_exact_match_reward/std": 0.501733124256134,
      "step": 85
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.215,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 19.379676818847656,
      "learning_rate": 6.026312439675551e-08,
      "loss": -0.0,
      "num_tokens": 8594558.0,
      "reward": 0.5031249523162842,
      "reward_std": 0.0646936446428299,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.453125,
      "rewards/mcq_exact_match_reward/std": 0.501733124256134,
      "step": 86
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2175,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 16.495403289794922,
      "learning_rate": 5.263533508961826e-08,
      "loss": -0.0,
      "num_tokens": 8700366.0,
      "reward": 0.6437499523162842,
      "reward_std": 0.0883883461356163,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.59375,
      "rewards/mcq_exact_match_reward/std": 0.49501484632492065,
      "step": 87
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.22,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 18.12753677368164,
      "learning_rate": 4.549673247541874e-08,
      "loss": -0.0,
      "num_tokens": 8817238.0,
      "reward": 0.5804687142372131,
      "reward_std": 0.1255451887845993,
      "rewards/format_reward/mean": 0.4921875,
      "rewards/format_reward/std": 0.0625,
      "rewards/mcq_exact_match_reward/mean": 0.53125,
      "rewards/mcq_exact_match_reward/std": 0.5029674172401428,
      "step": 88
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2225,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 32.31092834472656,
      "learning_rate": 3.8855122511307626e-08,
      "loss": 0.0,
      "num_tokens": 8920630.0,
      "reward": 0.4718749523162842,
      "reward_std": 0.3266732394695282,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.421875,
      "rewards/mcq_exact_match_reward/std": 0.49776285886764526,
      "step": 89
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.225,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 8.409988403320312,
      "learning_rate": 3.271776770026963e-08,
      "loss": -0.0,
      "num_tokens": 9014870.0,
      "reward": 0.7687499523162842,
      "reward_std": 0.0578637532889843,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.71875,
      "rewards/mcq_exact_match_reward/std": 0.4531635046005249,
      "step": 90
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2275,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 12.464180946350098,
      "learning_rate": 2.7091379149682682e-08,
      "loss": -0.0,
      "num_tokens": 9106758.0,
      "reward": 0.7374999523162842,
      "reward_std": 0.10888782143592834,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.6875,
      "rewards/mcq_exact_match_reward/std": 0.467176616191864,
      "step": 91
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.23,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 35.47895431518555,
      "learning_rate": 2.1982109232821176e-08,
      "loss": 0.0,
      "num_tokens": 9198654.0,
      "reward": 0.5031249523162842,
      "reward_std": 0.1530819833278656,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.453125,
      "rewards/mcq_exact_match_reward/std": 0.501733124256134,
      "step": 92
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2325,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 14.290452003479004,
      "learning_rate": 1.7395544861325718e-08,
      "loss": -0.0,
      "num_tokens": 9293702.0,
      "reward": 0.7218749523162842,
      "reward_std": 0.0646936446428299,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.671875,
      "rewards/mcq_exact_match_reward/std": 0.4732423722743988,
      "step": 93
    },
    {
      "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": 0.0,
      "completions/max_length": 26.0,
      "completions/max_terminated_length": 26.0,
      "completions/mean_length": 6.3125,
      "completions/mean_terminated_length": 6.3125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.235,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 39.53792953491211,
      "learning_rate": 1.3336701375997127e-08,
      "loss": -0.0,
      "num_tokens": 9405314.0,
      "reward": 0.3624999523162842,
      "reward_std": 0.22461532056331635,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.3125,
      "rewards/mcq_exact_match_reward/std": 0.467176616191864,
      "step": 94
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2375,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 14.538268089294434,
      "learning_rate": 9.810017062595321e-09,
      "loss": -0.0,
      "num_tokens": 9497994.0,
      "reward": 0.7999999523162842,
      "reward_std": 0.1157275065779686,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.75,
      "rewards/mcq_exact_match_reward/std": 0.4364357888698578,
      "step": 95
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.24,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 13.064372062683105,
      "learning_rate": 6.819348298638839e-09,
      "loss": -0.0,
      "num_tokens": 9600186.0,
      "reward": 0.5031249523162842,
      "reward_std": 0.10205792635679245,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.453125,
      "rewards/mcq_exact_match_reward/std": 0.501733124256134,
      "step": 96
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2425,
      "frac_reward_zero_std": 0.625,
      "grad_norm": 28.892200469970703,
      "learning_rate": 4.367965336512403e-09,
      "loss": 0.0,
      "num_tokens": 9689682.0,
      "reward": 0.4406249523162842,
      "reward_std": 0.17358145117759705,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.390625,
      "rewards/mcq_exact_match_reward/std": 0.4917473793029785,
      "step": 97
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.245,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 26.061214447021484,
      "learning_rate": 2.458548727494292e-09,
      "loss": 0.0,
      "num_tokens": 9783594.0,
      "reward": 0.5812499523162842,
      "reward_std": 0.2177756279706955,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.53125,
      "rewards/mcq_exact_match_reward/std": 0.5029674172401428,
      "step": 98
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.2475,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 31.908193588256836,
      "learning_rate": 1.0931863906127325e-09,
      "loss": -0.0,
      "num_tokens": 9877682.0,
      "reward": 0.5656249523162842,
      "reward_std": 0.2198972851037979,
      "rewards/format_reward/mean": 0.5,
      "rewards/format_reward/std": 0.0,
      "rewards/mcq_exact_match_reward/mean": 0.515625,
      "rewards/mcq_exact_match_reward/std": 0.5037065148353577,
      "step": 99
    },
    {
      "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": 0.0,
      "completions/max_length": 6.0,
      "completions/max_terminated_length": 6.0,
      "completions/mean_length": 6.0,
      "completions/mean_terminated_length": 6.0,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.25,
      "frac_reward_zero_std": 0.875,
      "grad_norm": 15.086002349853516,
      "learning_rate": 2.733713295369755e-10,
      "loss": 0.0,
      "num_tokens": 9962522.0,
      "reward": 0.4195312261581421,
      "reward_std": 0.00220970856025815,
      "rewards/format_reward/mean": 0.4453125,
      "rewards/format_reward/std": 0.15728822350502014,
      "rewards/mcq_exact_match_reward/mean": 0.375,
      "rewards/mcq_exact_match_reward/std": 0.48795005679130554,
      "step": 100
    }
  ],
  "logging_steps": 1,
  "max_steps": 100,
  "num_input_tokens_seen": 9962522,
  "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": 1,
  "trial_name": null,
  "trial_params": null
}