{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9945822640433419,
  "eval_steps": 500,
  "global_step": 109,
  "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": -5.0,
      "completions/max_length": 964.0,
      "completions/max_terminated_length": 964.0,
      "completions/mean_length": 527.6614990234375,
      "completions/mean_terminated_length": 527.6614990234375,
      "completions/min_length": 293.0,
      "completions/min_terminated_length": 293.0,
      "entropy": 0.1096649169921875,
      "epoch": 0.009124607927003137,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.023836389183998108,
      "kl": 0.00021350383758544922,
      "learning_rate": 2e-05,
      "loss": -0.0088,
      "num_tokens": 411230.0,
      "reward": 1.1042730808258057,
      "reward_std": 0.233424112200737,
      "rewards/cross_solution_unittest_reward/mean": 0.481730118393898,
      "rewards/cross_solution_unittest_reward/std": 0.22199106961488724,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 920.0,
      "completions/max_terminated_length": 920.0,
      "completions/mean_length": 491.33856201171875,
      "completions/mean_terminated_length": 491.33856201171875,
      "completions/min_length": 261.0,
      "completions/min_terminated_length": 261.0,
      "entropy": 0.1144561767578125,
      "epoch": 0.018249215854006275,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.029492009431123734,
      "kl": 0.0002351701259613037,
      "learning_rate": 1.9995921928281893e-05,
      "loss": 0.0028,
      "num_tokens": 821440.0,
      "reward": 0.8423246741294861,
      "reward_std": 0.2633346915245056,
      "rewards/cross_solution_unittest_reward/mean": 0.371605783700943,
      "rewards/cross_solution_unittest_reward/std": 0.25612012296915054,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1480.0,
      "completions/max_terminated_length": 1480.0,
      "completions/mean_length": 573.1614990234375,
      "completions/mean_terminated_length": 573.1614990234375,
      "completions/min_length": 295.0,
      "completions/min_terminated_length": 295.0,
      "entropy": 0.0967254638671875,
      "epoch": 0.02737382378100941,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.026458488777279854,
      "kl": 0.00021409988403320312,
      "learning_rate": 1.9983691039261358e-05,
      "loss": -0.0114,
      "num_tokens": 1262302.0,
      "reward": 0.9348706603050232,
      "reward_std": 0.2640388309955597,
      "rewards/cross_solution_unittest_reward/mean": 0.40714406967163086,
      "rewards/cross_solution_unittest_reward/std": 0.22153814882040024,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 871.0,
      "completions/max_terminated_length": 871.0,
      "completions/mean_length": 506.05731201171875,
      "completions/mean_terminated_length": 506.05731201171875,
      "completions/min_length": 250.0,
      "completions/min_terminated_length": 250.0,
      "entropy": 0.12152099609375,
      "epoch": 0.03649843170801255,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.028014499694108963,
      "kl": 0.0002561211585998535,
      "learning_rate": 1.9963317308626916e-05,
      "loss": -0.0048,
      "num_tokens": 1703844.0,
      "reward": 0.6856307983398438,
      "reward_std": 0.24917131662368774,
      "rewards/cross_solution_unittest_reward/mean": 0.3077041059732437,
      "rewards/cross_solution_unittest_reward/std": 0.18602629750967026,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1224.0,
      "completions/max_terminated_length": 1224.0,
      "completions/mean_length": 483.484375,
      "completions/mean_terminated_length": 483.484375,
      "completions/min_length": 247.0,
      "completions/min_terminated_length": 247.0,
      "entropy": 0.1254425048828125,
      "epoch": 0.04562303963501568,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03355209901928902,
      "kl": 0.0002625584602355957,
      "learning_rate": 1.99348173534855e-05,
      "loss": -0.004,
      "num_tokens": 2084126.0,
      "reward": 1.038306474685669,
      "reward_std": 0.295562207698822,
      "rewards/cross_solution_unittest_reward/mean": 0.4519632160663605,
      "rewards/cross_solution_unittest_reward/std": 0.27628327906131744,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.96875,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1174.0,
      "completions/mean_length": 545.671875,
      "completions/mean_terminated_length": 529.764404296875,
      "completions/min_length": 293.0,
      "completions/min_terminated_length": 293.0,
      "entropy": 0.11370849609375,
      "epoch": 0.05474764756201882,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.027173170819878578,
      "kl": 0.00024318695068359375,
      "learning_rate": 1.989821441880933e-05,
      "loss": 0.0171,
      "num_tokens": 2507912.0,
      "reward": 0.9802842140197754,
      "reward_std": 0.19230906665325165,
      "rewards/cross_solution_unittest_reward/mean": 0.4297175034880638,
      "rewards/cross_solution_unittest_reward/std": 0.24670374393463135,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.953125,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1065.0,
      "completions/mean_length": 556.296875,
      "completions/mean_terminated_length": 532.4566650390625,
      "completions/min_length": 259.0,
      "completions/min_terminated_length": 259.0,
      "entropy": 0.1289520263671875,
      "epoch": 0.06387225548902195,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02603895403444767,
      "kl": 0.0002505183219909668,
      "learning_rate": 1.9853538358476933e-05,
      "loss": -0.0028,
      "num_tokens": 2962034.0,
      "reward": 0.9045166969299316,
      "reward_std": 0.2844803035259247,
      "rewards/cross_solution_unittest_reward/mean": 0.39844612032175064,
      "rewards/cross_solution_unittest_reward/std": 0.21336368471384048,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1615.0,
      "completions/mean_length": 534.5911865234375,
      "completions/mean_terminated_length": 526.6292724609375,
      "completions/min_length": 260.0,
      "completions/min_terminated_length": 260.0,
      "entropy": 0.111663818359375,
      "epoch": 0.0729968634160251,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02666051872074604,
      "kl": 0.00025323033332824707,
      "learning_rate": 1.9800825610923937e-05,
      "loss": -0.0147,
      "num_tokens": 3421973.0,
      "reward": 0.8588270545005798,
      "reward_std": 0.2657862901687622,
      "rewards/cross_solution_unittest_reward/mean": 0.37082912772893906,
      "rewards/cross_solution_unittest_reward/std": 0.2715388759970665,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1115.0,
      "completions/max_terminated_length": 1115.0,
      "completions/mean_length": 489.3724060058594,
      "completions/mean_terminated_length": 489.3724060058594,
      "completions/min_length": 249.0,
      "completions/min_terminated_length": 249.0,
      "entropy": 0.122955322265625,
      "epoch": 0.08212147134302823,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02615615725517273,
      "kl": 0.0002747476100921631,
      "learning_rate": 1.9740119169423337e-05,
      "loss": 0.0001,
      "num_tokens": 3813332.0,
      "reward": 0.886775016784668,
      "reward_std": 0.2901659309864044,
      "rewards/cross_solution_unittest_reward/mean": 0.39264198392629623,
      "rewards/cross_solution_unittest_reward/std": 0.22153189778327942,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 908.0,
      "completions/max_terminated_length": 908.0,
      "completions/mean_length": 511.34375,
      "completions/mean_terminated_length": 511.34375,
      "completions/min_length": 204.0,
      "completions/min_terminated_length": 204.0,
      "entropy": 0.1106414794921875,
      "epoch": 0.09124607927003137,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.027914274483919144,
      "kl": 0.0002880394458770752,
      "learning_rate": 1.9671468547019575e-05,
      "loss": 0.0019,
      "num_tokens": 4259928.0,
      "reward": 0.8995641469955444,
      "reward_std": 0.2562140226364136,
      "rewards/cross_solution_unittest_reward/mean": 0.395971842110157,
      "rewards/cross_solution_unittest_reward/std": 0.23649413138628006,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1339.0,
      "completions/max_terminated_length": 1339.0,
      "completions/mean_length": 548.9244995117188,
      "completions/mean_terminated_length": 548.9244995117188,
      "completions/min_length": 258.0,
      "completions/min_terminated_length": 258.0,
      "entropy": 0.1057586669921875,
      "epoch": 0.1003706871970345,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.028584128245711327,
      "kl": 0.0003082454204559326,
      "learning_rate": 1.9594929736144978e-05,
      "loss": 0.0015,
      "num_tokens": 4686539.0,
      "reward": 0.9073726534843445,
      "reward_std": 0.2739028036594391,
      "rewards/cross_solution_unittest_reward/mean": 0.40253733843564987,
      "rewards/cross_solution_unittest_reward/std": 0.2085171639919281,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1507.0,
      "completions/max_terminated_length": 1507.0,
      "completions/mean_length": 504.3645935058594,
      "completions/mean_terminated_length": 504.3645935058594,
      "completions/min_length": 263.0,
      "completions/min_terminated_length": 263.0,
      "entropy": 0.1241607666015625,
      "epoch": 0.10949529512403763,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03190146014094353,
      "kl": 0.00035694241523742676,
      "learning_rate": 1.9510565162951538e-05,
      "loss": -0.0026,
      "num_tokens": 5118263.0,
      "reward": 0.7852699160575867,
      "reward_std": 0.29856932163238525,
      "rewards/cross_solution_unittest_reward/mean": 0.3413052186369896,
      "rewards/cross_solution_unittest_reward/std": 0.2405463457107544,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1108.0,
      "completions/mean_length": 537.6640625,
      "completions/mean_terminated_length": 529.710205078125,
      "completions/min_length": 237.0,
      "completions/min_terminated_length": 237.0,
      "entropy": 0.1090240478515625,
      "epoch": 0.11861990305104078,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03350791707634926,
      "kl": 0.0005662441253662109,
      "learning_rate": 1.941844363639525e-05,
      "loss": -0.0074,
      "num_tokens": 5532270.0,
      "reward": 0.9594830870628357,
      "reward_std": 0.22333046793937683,
      "rewards/cross_solution_unittest_reward/mean": 0.42083391547203064,
      "rewards/cross_solution_unittest_reward/std": 0.24756141006946564,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 2111.0,
      "completions/max_terminated_length": 2111.0,
      "completions/mean_length": 610.9375,
      "completions/mean_terminated_length": 610.9375,
      "completions/min_length": 245.0,
      "completions/min_terminated_length": 245.0,
      "entropy": 0.10589599609375,
      "epoch": 0.1277445109780439,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.028112368658185005,
      "kl": 0.00034502148628234863,
      "learning_rate": 1.9318640292114526e-05,
      "loss": -0.0104,
      "num_tokens": 5998974.0,
      "reward": 1.0141568183898926,
      "reward_std": 0.2032373547554016,
      "rewards/cross_solution_unittest_reward/mean": 0.44281716644763947,
      "rewards/cross_solution_unittest_reward/std": 0.2426743358373642,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 858.0,
      "completions/max_terminated_length": 858.0,
      "completions/mean_length": 463.62762451171875,
      "completions/mean_terminated_length": 463.62762451171875,
      "completions/min_length": 226.0,
      "completions/min_terminated_length": 226.0,
      "entropy": 0.12762451171875,
      "epoch": 0.13686911890504705,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.029718179255723953,
      "kl": 0.00044339895248413086,
      "learning_rate": 1.92112365311485e-05,
      "loss": 0.0141,
      "num_tokens": 6343095.0,
      "reward": 1.0152561664581299,
      "reward_std": 0.24558740854263306,
      "rewards/cross_solution_unittest_reward/mean": 0.4407452195882797,
      "rewards/cross_solution_unittest_reward/std": 0.2384948655962944,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.953125,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 3444.0,
      "completions/mean_length": 578.6536865234375,
      "completions/mean_terminated_length": 554.989501953125,
      "completions/min_length": 256.0,
      "completions/min_terminated_length": 256.0,
      "entropy": 0.1071319580078125,
      "epoch": 0.1459937268320502,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.028195451945066452,
      "kl": 0.00040775537490844727,
      "learning_rate": 1.9096319953545186e-05,
      "loss": 0.0066,
      "num_tokens": 6764442.0,
      "reward": 0.8980576992034912,
      "reward_std": 0.28982824087142944,
      "rewards/cross_solution_unittest_reward/mean": 0.3970365896821022,
      "rewards/cross_solution_unittest_reward/std": 0.2071005031466484,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 2372.0,
      "completions/mean_length": 528.796875,
      "completions/mean_terminated_length": 520.81982421875,
      "completions/min_length": 235.0,
      "completions/min_terminated_length": 235.0,
      "entropy": 0.11444091796875,
      "epoch": 0.15511833475905332,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.030508190393447876,
      "kl": 0.00042378902435302734,
      "learning_rate": 1.8973984286913584e-05,
      "loss": 0.0154,
      "num_tokens": 7181420.0,
      "reward": 1.1027473211288452,
      "reward_std": 0.28832367062568665,
      "rewards/cross_solution_unittest_reward/mean": 0.46901851892471313,
      "rewards/cross_solution_unittest_reward/std": 0.2841140776872635,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1141.0,
      "completions/max_terminated_length": 1141.0,
      "completions/mean_length": 544.4661865234375,
      "completions/mean_terminated_length": 544.4661865234375,
      "completions/min_length": 251.0,
      "completions/min_terminated_length": 251.0,
      "entropy": 0.104888916015625,
      "epoch": 0.16424294268605646,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.028670065104961395,
      "kl": 0.00043770670890808105,
      "learning_rate": 1.8844329309978146e-05,
      "loss": 0.002,
      "num_tokens": 7620935.0,
      "reward": 0.9763619303703308,
      "reward_std": 0.243991419672966,
      "rewards/cross_solution_unittest_reward/mean": 0.4205486625432968,
      "rewards/cross_solution_unittest_reward/std": 0.23177453130483627,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1056.0,
      "completions/max_terminated_length": 1056.0,
      "completions/mean_length": 503.703125,
      "completions/mean_terminated_length": 503.703125,
      "completions/min_length": 276.0,
      "completions/min_terminated_length": 276.0,
      "entropy": 0.113372802734375,
      "epoch": 0.17336755061305958,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02729717455804348,
      "kl": 0.0005033016204833984,
      "learning_rate": 1.8707460771197773e-05,
      "loss": 0.0014,
      "num_tokens": 8032845.0,
      "reward": 0.9106826782226562,
      "reward_std": 0.27397438883781433,
      "rewards/cross_solution_unittest_reward/mean": 0.39500948786735535,
      "rewards/cross_solution_unittest_reward/std": 0.2559768706560135,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 960.0,
      "completions/max_terminated_length": 960.0,
      "completions/mean_length": 545.7396240234375,
      "completions/mean_terminated_length": 545.7396240234375,
      "completions/min_length": 322.0,
      "completions/min_terminated_length": 322.0,
      "entropy": 0.1036376953125,
      "epoch": 0.18249215854006273,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.027323100715875626,
      "kl": 0.0005331039428710938,
      "learning_rate": 1.856349030251589e-05,
      "loss": 0.0049,
      "num_tokens": 8492777.0,
      "reward": 0.9297716021537781,
      "reward_std": 0.29434794187545776,
      "rewards/cross_solution_unittest_reward/mean": 0.4003482013940811,
      "rewards/cross_solution_unittest_reward/std": 0.2353406548500061,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1354.0,
      "completions/max_terminated_length": 1354.0,
      "completions/mean_length": 492.17449951171875,
      "completions/mean_terminated_length": 492.17449951171875,
      "completions/min_length": 244.0,
      "completions/min_terminated_length": 244.0,
      "entropy": 0.105987548828125,
      "epoch": 0.19161676646706588,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02947384864091873,
      "kl": 0.0005356073379516602,
      "learning_rate": 1.8412535328311813e-05,
      "loss": 0.006,
      "num_tokens": 8880476.0,
      "reward": 0.9555484652519226,
      "reward_std": 0.23427900671958923,
      "rewards/cross_solution_unittest_reward/mean": 0.4216877445578575,
      "rewards/cross_solution_unittest_reward/std": 0.21247681975364685,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1018.0,
      "completions/max_terminated_length": 1018.0,
      "completions/mean_length": 506.7552185058594,
      "completions/mean_terminated_length": 506.7552185058594,
      "completions/min_length": 218.0,
      "completions/min_terminated_length": 218.0,
      "entropy": 0.1118011474609375,
      "epoch": 0.200741374394069,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03192589804530144,
      "kl": 0.0005907416343688965,
      "learning_rate": 1.825471896962774e-05,
      "loss": -0.0137,
      "num_tokens": 9256446.0,
      "reward": 0.9712496995925903,
      "reward_std": 0.2654087245464325,
      "rewards/cross_solution_unittest_reward/mean": 0.41771475970745087,
      "rewards/cross_solution_unittest_reward/std": 0.2567882165312767,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1568.0,
      "completions/max_terminated_length": 1568.0,
      "completions/mean_length": 527.6666870117188,
      "completions/mean_terminated_length": 527.6666870117188,
      "completions/min_length": 289.0,
      "completions/min_terminated_length": 289.0,
      "entropy": 0.108154296875,
      "epoch": 0.20986598232107215,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02674136683344841,
      "kl": 0.000629425048828125,
      "learning_rate": 1.8090169943749477e-05,
      "loss": -0.0106,
      "num_tokens": 9710614.0,
      "reward": 1.0353918075561523,
      "reward_std": 0.22365829348564148,
      "rewards/cross_solution_unittest_reward/mean": 0.4482730031013489,
      "rewards/cross_solution_unittest_reward/std": 0.2228257954120636,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 834.0,
      "completions/max_terminated_length": 834.0,
      "completions/mean_length": 509.2395935058594,
      "completions/mean_terminated_length": 509.2395935058594,
      "completions/min_length": 238.0,
      "completions/min_terminated_length": 238.0,
      "entropy": 0.1023101806640625,
      "epoch": 0.21899059024807527,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02758539468050003,
      "kl": 0.0006355643272399902,
      "learning_rate": 1.7919022459222754e-05,
      "loss": 0.0141,
      "num_tokens": 10133834.0,
      "reward": 0.8483362197875977,
      "reward_std": 0.2694234549999237,
      "rewards/cross_solution_unittest_reward/mean": 0.3766419366002083,
      "rewards/cross_solution_unittest_reward/std": 0.26132430136203766,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 980.0,
      "completions/mean_length": 494.6380310058594,
      "completions/mean_terminated_length": 486.5718078613281,
      "completions/min_length": 231.0,
      "completions/min_terminated_length": 231.0,
      "entropy": 0.11871337890625,
      "epoch": 0.22811519817507842,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.0348508358001709,
      "kl": 0.0009682774543762207,
      "learning_rate": 1.7741416106390828e-05,
      "loss": -0.0021,
      "num_tokens": 10537447.0,
      "reward": 0.9557062387466431,
      "reward_std": 0.22823719680309296,
      "rewards/cross_solution_unittest_reward/mean": 0.4154369905591011,
      "rewards/cross_solution_unittest_reward/std": 0.20701014250516891,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1757.0,
      "completions/max_terminated_length": 1757.0,
      "completions/mean_length": 497.625,
      "completions/mean_terminated_length": 497.625,
      "completions/min_length": 213.0,
      "completions/min_terminated_length": 213.0,
      "entropy": 0.11309814453125,
      "epoch": 0.23723980610208156,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.028710953891277313,
      "kl": 0.0007503032684326172,
      "learning_rate": 1.7557495743542586e-05,
      "loss": 0.0115,
      "num_tokens": 10895703.0,
      "reward": 0.8731624484062195,
      "reward_std": 0.24689321219921112,
      "rewards/cross_solution_unittest_reward/mean": 0.3826525807380676,
      "rewards/cross_solution_unittest_reward/std": 0.24787958711385727,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 902.0,
      "completions/max_terminated_length": 902.0,
      "completions/mean_length": 473.3671875,
      "completions/mean_terminated_length": 473.3671875,
      "completions/min_length": 276.0,
      "completions/min_terminated_length": 276.0,
      "entropy": 0.111358642578125,
      "epoch": 0.24636441402908468,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.030619844794273376,
      "kl": 0.0008200407028198242,
      "learning_rate": 1.736741137876405e-05,
      "loss": 0.002,
      "num_tokens": 11274364.0,
      "reward": 1.0031293630599976,
      "reward_std": 0.2526843547821045,
      "rewards/cross_solution_unittest_reward/mean": 0.44217514991760254,
      "rewards/cross_solution_unittest_reward/std": 0.21224447339773178,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 957.0,
      "completions/max_terminated_length": 957.0,
      "completions/mean_length": 506.1458435058594,
      "completions/mean_terminated_length": 506.1458435058594,
      "completions/min_length": 219.0,
      "completions/min_terminated_length": 219.0,
      "entropy": 0.1076812744140625,
      "epoch": 0.2554890219560878,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.026609111577272415,
      "kl": 0.0008921623229980469,
      "learning_rate": 1.7171318047589637e-05,
      "loss": -0.0014,
      "num_tokens": 11706916.0,
      "reward": 1.137014389038086,
      "reward_std": 0.2074248045682907,
      "rewards/cross_solution_unittest_reward/mean": 0.4929664433002472,
      "rewards/cross_solution_unittest_reward/std": 0.26993829011917114,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 952.0,
      "completions/max_terminated_length": 952.0,
      "completions/mean_length": 489.57293701171875,
      "completions/mean_terminated_length": 489.57293701171875,
      "completions/min_length": 256.0,
      "completions/min_terminated_length": 256.0,
      "entropy": 0.117828369140625,
      "epoch": 0.26461362988309095,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.031492725014686584,
      "kl": 0.001159071922302246,
      "learning_rate": 1.696937568655294e-05,
      "loss": -0.0028,
      "num_tokens": 12102904.0,
      "reward": 1.0333976745605469,
      "reward_std": 0.2310364693403244,
      "rewards/cross_solution_unittest_reward/mean": 0.4447094053030014,
      "rewards/cross_solution_unittest_reward/std": 0.29005488753318787,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.96875,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1168.0,
      "completions/mean_length": 544.9635620117188,
      "completions/mean_terminated_length": 529.0523681640625,
      "completions/min_length": 252.0,
      "completions/min_terminated_length": 252.0,
      "entropy": 0.103607177734375,
      "epoch": 0.2737382378100941,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.028927048668265343,
      "kl": 0.0009303092956542969,
      "learning_rate": 1.6761749002740195e-05,
      "loss": 0.0036,
      "num_tokens": 12531050.0,
      "reward": 0.8030904531478882,
      "reward_std": 0.2894265055656433,
      "rewards/cross_solution_unittest_reward/mean": 0.35538169741630554,
      "rewards/cross_solution_unittest_reward/std": 0.27038178592920303,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.96875,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 946.0,
      "completions/mean_length": 486.05731201171875,
      "completions/mean_terminated_length": 469.83770751953125,
      "completions/min_length": 236.0,
      "completions/min_terminated_length": 236.0,
      "entropy": 0.1143035888671875,
      "epoch": 0.28286284573709725,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.030626529827713966,
      "kl": 0.0011305809020996094,
      "learning_rate": 1.6548607339452853e-05,
      "loss": 0.0161,
      "num_tokens": 12901128.0,
      "reward": 1.0075210332870483,
      "reward_std": 0.2209041714668274,
      "rewards/cross_solution_unittest_reward/mean": 0.4487333372235298,
      "rewards/cross_solution_unittest_reward/std": 0.21068686991930008,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 954.0,
      "completions/max_terminated_length": 954.0,
      "completions/mean_length": 481.5677185058594,
      "completions/mean_terminated_length": 481.5677185058594,
      "completions/min_length": 237.0,
      "completions/min_terminated_length": 237.0,
      "entropy": 0.103546142578125,
      "epoch": 0.2919874536641004,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03601611778140068,
      "kl": 0.0013216733932495117,
      "learning_rate": 1.6330124538088705e-05,
      "loss": 0.0017,
      "num_tokens": 13354434.0,
      "reward": 1.051571011543274,
      "reward_std": 0.3614152669906616,
      "rewards/cross_solution_unittest_reward/mean": 0.4522690027952194,
      "rewards/cross_solution_unittest_reward/std": 0.24180202186107635,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 773.0,
      "completions/max_terminated_length": 773.0,
      "completions/mean_length": 469.72137451171875,
      "completions/mean_terminated_length": 469.72137451171875,
      "completions/min_length": 241.0,
      "completions/min_terminated_length": 241.0,
      "entropy": 0.115203857421875,
      "epoch": 0.3011120615911035,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02829568460583687,
      "kl": 0.0014170408248901367,
      "learning_rate": 1.6106478796354382e-05,
      "loss": 0.0034,
      "num_tokens": 13732783.0,
      "reward": 0.921411395072937,
      "reward_std": 0.17318974435329437,
      "rewards/cross_solution_unittest_reward/mean": 0.4217027798295021,
      "rewards/cross_solution_unittest_reward/std": 0.22807485610246658,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 839.0,
      "completions/max_terminated_length": 839.0,
      "completions/mean_length": 488.7578125,
      "completions/mean_terminated_length": 488.7578125,
      "completions/min_length": 217.0,
      "completions/min_terminated_length": 217.0,
      "entropy": 0.1075286865234375,
      "epoch": 0.31023666951810663,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.029855074360966682,
      "kl": 0.001324295997619629,
      "learning_rate": 1.5877852522924733e-05,
      "loss": -0.0055,
      "num_tokens": 14159730.0,
      "reward": 1.081244707107544,
      "reward_std": 0.18814358115196228,
      "rewards/cross_solution_unittest_reward/mean": 0.4768352806568146,
      "rewards/cross_solution_unittest_reward/std": 0.22413263469934464,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1815.0,
      "completions/max_terminated_length": 1815.0,
      "completions/mean_length": 510.74481201171875,
      "completions/mean_terminated_length": 510.74481201171875,
      "completions/min_length": 280.0,
      "completions/min_terminated_length": 280.0,
      "entropy": 0.1134490966796875,
      "epoch": 0.3193612774451098,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03141755983233452,
      "kl": 0.0015311241149902344,
      "learning_rate": 1.5644432188667695e-05,
      "loss": 0.0011,
      "num_tokens": 14547176.0,
      "reward": 0.9383748769760132,
      "reward_std": 0.2669951617717743,
      "rewards/cross_solution_unittest_reward/mean": 0.3990442007780075,
      "rewards/cross_solution_unittest_reward/std": 0.24042869359254837,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1357.0,
      "completions/mean_length": 537.3489990234375,
      "completions/mean_terminated_length": 529.394287109375,
      "completions/min_length": 273.0,
      "completions/min_terminated_length": 273.0,
      "entropy": 0.1122283935546875,
      "epoch": 0.32848588537211293,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.030957432463765144,
      "kl": 0.0015711784362792969,
      "learning_rate": 1.5406408174555978e-05,
      "loss": -0.0053,
      "num_tokens": 14975526.0,
      "reward": 0.9109244346618652,
      "reward_std": 0.25726160407066345,
      "rewards/cross_solution_unittest_reward/mean": 0.3915872722864151,
      "rewards/cross_solution_unittest_reward/std": 0.22853311151266098,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 2463.0,
      "completions/mean_length": 520.1276245117188,
      "completions/mean_terminated_length": 512.1279296875,
      "completions/min_length": 246.0,
      "completions/min_terminated_length": 246.0,
      "entropy": 0.097320556640625,
      "epoch": 0.3376104932991161,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.032772429287433624,
      "kl": 0.0015528202056884766,
      "learning_rate": 1.5163974616389621e-05,
      "loss": 0.0123,
      "num_tokens": 15386143.0,
      "reward": 0.9813830256462097,
      "reward_std": 0.22931739687919617,
      "rewards/cross_solution_unittest_reward/mean": 0.43425410240888596,
      "rewards/cross_solution_unittest_reward/std": 0.22800295799970627,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 887.0,
      "completions/max_terminated_length": 887.0,
      "completions/mean_length": 485.6380310058594,
      "completions/mean_terminated_length": 485.6380310058594,
      "completions/min_length": 205.0,
      "completions/min_terminated_length": 205.0,
      "entropy": 0.1133270263671875,
      "epoch": 0.34673510122611917,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03404747322201729,
      "kl": 0.0015780925750732422,
      "learning_rate": 1.491732924645604e-05,
      "loss": 0.0191,
      "num_tokens": 15770812.0,
      "reward": 0.9875783920288086,
      "reward_std": 0.2382659763097763,
      "rewards/cross_solution_unittest_reward/mean": 0.4320690929889679,
      "rewards/cross_solution_unittest_reward/std": 0.2256365790963173,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1240.0,
      "completions/max_terminated_length": 1240.0,
      "completions/mean_length": 473.8984375,
      "completions/mean_terminated_length": 473.8984375,
      "completions/min_length": 179.0,
      "completions/min_terminated_length": 179.0,
      "entropy": 0.116729736328125,
      "epoch": 0.3558597091531223,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03682932257652283,
      "kl": 0.003524303436279297,
      "learning_rate": 1.4666673232256738e-05,
      "loss": -0.0084,
      "num_tokens": 16082925.0,
      "reward": 0.8647634387016296,
      "reward_std": 0.3210960626602173,
      "rewards/cross_solution_unittest_reward/mean": 0.37845826894044876,
      "rewards/cross_solution_unittest_reward/std": 0.2232651561498642,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 2045.0,
      "completions/max_terminated_length": 2045.0,
      "completions/mean_length": 503.125,
      "completions/mean_terminated_length": 503.125,
      "completions/min_length": 202.0,
      "completions/min_terminated_length": 202.0,
      "entropy": 0.102325439453125,
      "epoch": 0.36498431708012546,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.04535383731126785,
      "kl": 0.003467082977294922,
      "learning_rate": 1.4412211012432213e-05,
      "loss": 0.0068,
      "num_tokens": 16496789.0,
      "reward": 0.8126609325408936,
      "reward_std": 0.26924118399620056,
      "rewards/cross_solution_unittest_reward/mean": 0.3569253906607628,
      "rewards/cross_solution_unittest_reward/std": 0.2864082232117653,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 958.0,
      "completions/max_terminated_length": 958.0,
      "completions/mean_length": 511.31512451171875,
      "completions/mean_terminated_length": 511.31512451171875,
      "completions/min_length": 248.0,
      "completions/min_terminated_length": 248.0,
      "entropy": 0.1136627197265625,
      "epoch": 0.3741089250071286,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.033672649413347244,
      "kl": 0.0028324127197265625,
      "learning_rate": 1.4154150130018867e-05,
      "loss": -0.0017,
      "num_tokens": 16912990.0,
      "reward": 0.7562817335128784,
      "reward_std": 0.20014888048171997,
      "rewards/cross_solution_unittest_reward/mean": 0.3420030549168587,
      "rewards/cross_solution_unittest_reward/std": 0.21658644825220108,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1215.0,
      "completions/max_terminated_length": 1215.0,
      "completions/mean_length": 506.51043701171875,
      "completions/mean_terminated_length": 506.51043701171875,
      "completions/min_length": 264.0,
      "completions/min_terminated_length": 264.0,
      "entropy": 0.093170166015625,
      "epoch": 0.38323353293413176,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03319159895181656,
      "kl": 0.0018248558044433594,
      "learning_rate": 1.3892701063173917e-05,
      "loss": -0.0041,
      "num_tokens": 17357914.0,
      "reward": 0.9257922172546387,
      "reward_std": 0.2104812115430832,
      "rewards/cross_solution_unittest_reward/mean": 0.4091387465596199,
      "rewards/cross_solution_unittest_reward/std": 0.23963715881109238,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 799.0,
      "completions/max_terminated_length": 799.0,
      "completions/mean_length": 478.70574951171875,
      "completions/mean_terminated_length": 478.70574951171875,
      "completions/min_length": 249.0,
      "completions/min_terminated_length": 249.0,
      "entropy": 0.105377197265625,
      "epoch": 0.39235814086113485,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03309491649270058,
      "kl": 0.00234222412109375,
      "learning_rate": 1.362807705350641e-05,
      "loss": -0.0027,
      "num_tokens": 17696089.0,
      "reward": 1.0697866678237915,
      "reward_std": 0.21351736783981323,
      "rewards/cross_solution_unittest_reward/mean": 0.46968914568424225,
      "rewards/cross_solution_unittest_reward/std": 0.25238344073295593,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 2798.0,
      "completions/mean_length": 519.0208740234375,
      "completions/mean_terminated_length": 511.0182800292969,
      "completions/min_length": 224.0,
      "completions/min_terminated_length": 224.0,
      "entropy": 0.095550537109375,
      "epoch": 0.401482748788138,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.029689131304621696,
      "kl": 0.0020627975463867188,
      "learning_rate": 1.3360493932154301e-05,
      "loss": -0.0021,
      "num_tokens": 18106297.0,
      "reward": 0.7189862728118896,
      "reward_std": 0.21532303094863892,
      "rewards/cross_solution_unittest_reward/mean": 0.3341630734503269,
      "rewards/cross_solution_unittest_reward/std": 0.210732139647007,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1472.0,
      "completions/max_terminated_length": 1472.0,
      "completions/mean_length": 467.91668701171875,
      "completions/mean_terminated_length": 467.91668701171875,
      "completions/min_length": 196.0,
      "completions/min_terminated_length": 196.0,
      "entropy": 0.1134033203125,
      "epoch": 0.41060735671514115,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.033839646726846695,
      "kl": 0.0030853748321533203,
      "learning_rate": 1.3090169943749475e-05,
      "loss": 0.021,
      "num_tokens": 18488529.0,
      "reward": 1.0240888595581055,
      "reward_std": 0.21602611243724823,
      "rewards/cross_solution_unittest_reward/mean": 0.4490320533514023,
      "rewards/cross_solution_unittest_reward/std": 0.2549639642238617,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1161.0,
      "completions/mean_length": 520.1119995117188,
      "completions/mean_terminated_length": 512.1123046875,
      "completions/min_length": 281.0,
      "completions/min_terminated_length": 281.0,
      "entropy": 0.0985565185546875,
      "epoch": 0.4197319646421443,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.031324755400419235,
      "kl": 0.0023124217987060547,
      "learning_rate": 1.2817325568414299e-05,
      "loss": -0.0063,
      "num_tokens": 18908484.0,
      "reward": 0.8984124064445496,
      "reward_std": 0.31652843952178955,
      "rewards/cross_solution_unittest_reward/mean": 0.394581638276577,
      "rewards/cross_solution_unittest_reward/std": 0.2338028848171234,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.921875,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1145.0,
      "completions/mean_length": 528.3125,
      "completions/mean_terminated_length": 488.0000305175781,
      "completions/min_length": 215.0,
      "completions/min_terminated_length": 215.0,
      "entropy": 0.11846923828125,
      "epoch": 0.42885657256914744,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03654487058520317,
      "kl": 0.003631591796875,
      "learning_rate": 1.2542183341934873e-05,
      "loss": 0.0342,
      "num_tokens": 19336532.0,
      "reward": 0.7277412414550781,
      "reward_std": 0.2056652158498764,
      "rewards/cross_solution_unittest_reward/mean": 0.3252406492829323,
      "rewards/cross_solution_unittest_reward/std": 0.24651427567005157,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1054.0,
      "completions/max_terminated_length": 1054.0,
      "completions/mean_length": 492.11199951171875,
      "completions/mean_terminated_length": 492.11199951171875,
      "completions/min_length": 262.0,
      "completions/min_terminated_length": 262.0,
      "entropy": 0.1041717529296875,
      "epoch": 0.43798118049615054,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03339504823088646,
      "kl": 0.0033910274505615234,
      "learning_rate": 1.2264967674257647e-05,
      "loss": 0.0055,
      "num_tokens": 19727471.0,
      "reward": 0.9818544387817383,
      "reward_std": 0.26466554403305054,
      "rewards/cross_solution_unittest_reward/mean": 0.4283819645643234,
      "rewards/cross_solution_unittest_reward/std": 0.22476153075695038,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1095.0,
      "completions/mean_length": 502.90625,
      "completions/mean_terminated_length": 494.86163330078125,
      "completions/min_length": 262.0,
      "completions/min_terminated_length": 262.0,
      "entropy": 0.1012725830078125,
      "epoch": 0.4471057884231537,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 18.39661979675293,
      "kl": 0.8277513980865479,
      "learning_rate": 1.1985904666457455e-05,
      "loss": 0.0468,
      "num_tokens": 20142435.0,
      "reward": 0.7730751037597656,
      "reward_std": 0.30047422647476196,
      "rewards/cross_solution_unittest_reward/mean": 0.33403923362493515,
      "rewards/cross_solution_unittest_reward/std": 0.25003235042095184,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1312.0,
      "completions/max_terminated_length": 1312.0,
      "completions/mean_length": 509.84375,
      "completions/mean_terminated_length": 509.84375,
      "completions/min_length": 248.0,
      "completions/min_terminated_length": 248.0,
      "entropy": 0.09967041015625,
      "epoch": 0.45623039635015683,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03424382582306862,
      "kl": 0.0036292076110839844,
      "learning_rate": 1.170522192632624e-05,
      "loss": -0.0109,
      "num_tokens": 20554295.0,
      "reward": 0.9384050369262695,
      "reward_std": 0.2790818512439728,
      "rewards/cross_solution_unittest_reward/mean": 0.41122280806303024,
      "rewards/cross_solution_unittest_reward/std": 0.28970006108283997,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 909.0,
      "completions/max_terminated_length": 909.0,
      "completions/mean_length": 509.796875,
      "completions/mean_terminated_length": 509.796875,
      "completions/min_length": 173.0,
      "completions/min_terminated_length": 173.0,
      "entropy": 0.0945892333984375,
      "epoch": 0.46535500427716,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.030270231887698174,
      "kl": 0.003202676773071289,
      "learning_rate": 1.1423148382732854e-05,
      "loss": 0.0027,
      "num_tokens": 20974745.0,
      "reward": 1.0798063278198242,
      "reward_std": 0.2213771790266037,
      "rewards/cross_solution_unittest_reward/mean": 0.4737387150526047,
      "rewards/cross_solution_unittest_reward/std": 0.18860329687595367,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1354.0,
      "completions/mean_length": 513.109375,
      "completions/mean_terminated_length": 505.0914001464844,
      "completions/min_length": 251.0,
      "completions/min_terminated_length": 251.0,
      "entropy": 0.099945068359375,
      "epoch": 0.4744796122041631,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03432578966021538,
      "kl": 0.0032448768615722656,
      "learning_rate": 1.1139914098905406e-05,
      "loss": 0.0178,
      "num_tokens": 21378419.0,
      "reward": 1.0716464519500732,
      "reward_std": 0.2842645049095154,
      "rewards/cross_solution_unittest_reward/mean": 0.47282470762729645,
      "rewards/cross_solution_unittest_reward/std": 0.23804496228694916,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1391.0,
      "completions/max_terminated_length": 1391.0,
      "completions/mean_length": 487.3828125,
      "completions/mean_terminated_length": 487.3828125,
      "completions/min_length": 250.0,
      "completions/min_terminated_length": 250.0,
      "entropy": 0.0956878662109375,
      "epoch": 0.4836042201311662,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03196337819099426,
      "kl": 0.0026366710662841797,
      "learning_rate": 1.08557500847884e-05,
      "loss": -0.0004,
      "num_tokens": 21779142.0,
      "reward": 0.8248944282531738,
      "reward_std": 0.2299119085073471,
      "rewards/cross_solution_unittest_reward/mean": 0.3687998875975609,
      "rewards/cross_solution_unittest_reward/std": 0.22956574708223343,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1238.0,
      "completions/max_terminated_length": 1238.0,
      "completions/mean_length": 536.3932495117188,
      "completions/mean_terminated_length": 536.3932495117188,
      "completions/min_length": 250.0,
      "completions/min_terminated_length": 250.0,
      "entropy": 0.1009521484375,
      "epoch": 0.49272882805816937,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03287142142653465,
      "kl": 0.003429412841796875,
      "learning_rate": 1.0570888108627682e-05,
      "loss": -0.0033,
      "num_tokens": 22241477.0,
      "reward": 0.8158465623855591,
      "reward_std": 0.2910717725753784,
      "rewards/cross_solution_unittest_reward/mean": 0.358261875808239,
      "rewards/cross_solution_unittest_reward/std": 0.21051427721977234,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 898.0,
      "completions/max_terminated_length": 898.0,
      "completions/mean_length": 491.3802185058594,
      "completions/mean_terminated_length": 491.3802185058594,
      "completions/min_length": 219.0,
      "completions/min_terminated_length": 219.0,
      "entropy": 0.1052398681640625,
      "epoch": 0.5018534359851725,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03173370659351349,
      "kl": 0.0028574466705322266,
      "learning_rate": 1.0285560507936962e-05,
      "loss": -0.0091,
      "num_tokens": 22619007.0,
      "reward": 0.8488973379135132,
      "reward_std": 0.24665257334709167,
      "rewards/cross_solution_unittest_reward/mean": 0.37161923199892044,
      "rewards/cross_solution_unittest_reward/std": 0.21222124248743057,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 932.0,
      "completions/max_terminated_length": 932.0,
      "completions/mean_length": 494.5208435058594,
      "completions/mean_terminated_length": 494.5208435058594,
      "completions/min_length": 264.0,
      "completions/min_terminated_length": 264.0,
      "entropy": 0.1017913818359375,
      "epoch": 0.5109780439121756,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.045224472880363464,
      "kl": 0.00625157356262207,
      "learning_rate": 1e-05,
      "loss": 0.0093,
      "num_tokens": 22993871.0,
      "reward": 0.9006222486495972,
      "reward_std": 0.21599134802818298,
      "rewards/cross_solution_unittest_reward/mean": 0.40076694637537,
      "rewards/cross_solution_unittest_reward/std": 0.22537627071142197,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 981.0,
      "completions/mean_length": 510.2161560058594,
      "completions/mean_terminated_length": 502.19061279296875,
      "completions/min_length": 308.0,
      "completions/min_terminated_length": 308.0,
      "entropy": 0.1026153564453125,
      "epoch": 0.5201026518391788,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.033613454550504684,
      "kl": 0.0032994747161865234,
      "learning_rate": 9.71443949206304e-06,
      "loss": -0.0064,
      "num_tokens": 23442418.0,
      "reward": 0.9552373290061951,
      "reward_std": 0.21273988485336304,
      "rewards/cross_solution_unittest_reward/mean": 0.4271874502301216,
      "rewards/cross_solution_unittest_reward/std": 0.20565658062696457,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.96875,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 925.0,
      "completions/mean_length": 494.5208435058594,
      "completions/mean_terminated_length": 478.3455505371094,
      "completions/min_length": 258.0,
      "completions/min_terminated_length": 258.0,
      "entropy": 0.117401123046875,
      "epoch": 0.5292272597661819,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03614190220832825,
      "kl": 0.0036954879760742188,
      "learning_rate": 9.42911189137232e-06,
      "loss": 0.0155,
      "num_tokens": 23857530.0,
      "reward": 0.9477154016494751,
      "reward_std": 0.20985114574432373,
      "rewards/cross_solution_unittest_reward/mean": 0.41357675194740295,
      "rewards/cross_solution_unittest_reward/std": 0.2724653407931328,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 899.0,
      "completions/mean_length": 474.82293701171875,
      "completions/mean_terminated_length": 466.7049560546875,
      "completions/min_length": 242.0,
      "completions/min_terminated_length": 242.0,
      "entropy": 0.1109161376953125,
      "epoch": 0.5383518676931851,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03441567346453667,
      "kl": 0.00311279296875,
      "learning_rate": 9.144249915211605e-06,
      "loss": 0.0062,
      "num_tokens": 24261294.0,
      "reward": 0.9669393301010132,
      "reward_std": 0.22922192513942719,
      "rewards/cross_solution_unittest_reward/mean": 0.4128674417734146,
      "rewards/cross_solution_unittest_reward/std": 0.2560960501432419,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 854.0,
      "completions/max_terminated_length": 854.0,
      "completions/mean_length": 497.9609375,
      "completions/mean_terminated_length": 497.9609375,
      "completions/min_length": 250.0,
      "completions/min_terminated_length": 250.0,
      "entropy": 0.0975494384765625,
      "epoch": 0.5474764756201882,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.033020906150341034,
      "kl": 0.002737283706665039,
      "learning_rate": 8.860085901094595e-06,
      "loss": -0.0018,
      "num_tokens": 24696543.0,
      "reward": 0.9655224680900574,
      "reward_std": 0.28828054666519165,
      "rewards/cross_solution_unittest_reward/mean": 0.4253334328532219,
      "rewards/cross_solution_unittest_reward/std": 0.23384006321430206,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 999.0,
      "completions/max_terminated_length": 999.0,
      "completions/mean_length": 524.890625,
      "completions/mean_terminated_length": 524.890625,
      "completions/min_length": 263.0,
      "completions/min_terminated_length": 263.0,
      "entropy": 0.096435546875,
      "epoch": 0.5566010835471913,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03234833851456642,
      "kl": 0.003005504608154297,
      "learning_rate": 8.576851617267151e-06,
      "loss": 0.0086,
      "num_tokens": 25143405.0,
      "reward": 1.010263442993164,
      "reward_std": 0.23946420848369598,
      "rewards/cross_solution_unittest_reward/mean": 0.44504720717668533,
      "rewards/cross_solution_unittest_reward/std": 0.2548332065343857,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1025.0,
      "completions/mean_length": 530.8958740234375,
      "completions/mean_terminated_length": 522.92431640625,
      "completions/min_length": 232.0,
      "completions/min_terminated_length": 232.0,
      "entropy": 0.10467529296875,
      "epoch": 0.5657256914741945,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03223549574613571,
      "kl": 0.002941131591796875,
      "learning_rate": 8.294778073673762e-06,
      "loss": 0.007,
      "num_tokens": 25572309.0,
      "reward": 0.7772404551506042,
      "reward_std": 0.2317599058151245,
      "rewards/cross_solution_unittest_reward/mean": 0.35900117829442024,
      "rewards/cross_solution_unittest_reward/std": 0.1959025040268898,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1132.0,
      "completions/mean_length": 478.97137451171875,
      "completions/mean_terminated_length": 470.8642272949219,
      "completions/min_length": 156.0,
      "completions/min_terminated_length": 156.0,
      "entropy": 0.1028900146484375,
      "epoch": 0.5748502994011976,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.034286729991436005,
      "kl": 0.0025823116302490234,
      "learning_rate": 8.014095333542548e-06,
      "loss": 0.0114,
      "num_tokens": 25945338.0,
      "reward": 0.8456071615219116,
      "reward_std": 0.2458418607711792,
      "rewards/cross_solution_unittest_reward/mean": 0.37883561104536057,
      "rewards/cross_solution_unittest_reward/std": 0.22371377050876617,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 855.0,
      "completions/max_terminated_length": 855.0,
      "completions/mean_length": 491.22137451171875,
      "completions/mean_terminated_length": 491.22137451171875,
      "completions/min_length": 309.0,
      "completions/min_terminated_length": 309.0,
      "entropy": 0.0994110107421875,
      "epoch": 0.5839749073282008,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.030411940068006516,
      "kl": 0.0029740333557128906,
      "learning_rate": 7.735032325742355e-06,
      "loss": 0.0071,
      "num_tokens": 26337343.0,
      "reward": 1.2452019453048706,
      "reward_std": 0.28348508477211,
      "rewards/cross_solution_unittest_reward/mean": 0.5320437103509903,
      "rewards/cross_solution_unittest_reward/std": 0.20597976446151733,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1100.0,
      "completions/max_terminated_length": 1100.0,
      "completions/mean_length": 481.04949951171875,
      "completions/mean_terminated_length": 481.04949951171875,
      "completions/min_length": 232.0,
      "completions/min_terminated_length": 232.0,
      "entropy": 0.1112823486328125,
      "epoch": 0.5930995152552039,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03709479793906212,
      "kl": 0.0034236907958984375,
      "learning_rate": 7.4578166580651335e-06,
      "loss": 0.0018,
      "num_tokens": 26727642.0,
      "reward": 1.0363634824752808,
      "reward_std": 0.2708354592323303,
      "rewards/cross_solution_unittest_reward/mean": 0.4516971707344055,
      "rewards/cross_solution_unittest_reward/std": 0.30414242297410965,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.953125,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1223.0,
      "completions/mean_length": 540.4427490234375,
      "completions/mean_terminated_length": 516.4776611328125,
      "completions/min_length": 225.0,
      "completions/min_terminated_length": 225.0,
      "entropy": 0.102630615234375,
      "epoch": 0.602224123182207,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.031675394624471664,
      "kl": 0.0034074783325195312,
      "learning_rate": 7.182674431585703e-06,
      "loss": -0.0053,
      "num_tokens": 27207164.0,
      "reward": 1.1464284658432007,
      "reward_std": 0.23834940791130066,
      "rewards/cross_solution_unittest_reward/mean": 0.4887780696153641,
      "rewards/cross_solution_unittest_reward/std": 0.26997726410627365,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1062.0,
      "completions/max_terminated_length": 1062.0,
      "completions/mean_length": 492.33856201171875,
      "completions/mean_terminated_length": 492.33856201171875,
      "completions/min_length": 217.0,
      "completions/min_terminated_length": 217.0,
      "entropy": 0.098785400390625,
      "epoch": 0.6113487311092102,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03432027995586395,
      "kl": 0.004305839538574219,
      "learning_rate": 6.909830056250527e-06,
      "loss": 0.0091,
      "num_tokens": 27661350.0,
      "reward": 1.2568825483322144,
      "reward_std": 0.3327004909515381,
      "rewards/cross_solution_unittest_reward/mean": 0.5392695367336273,
      "rewards/cross_solution_unittest_reward/std": 0.24271728843450546,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 953.0,
      "completions/max_terminated_length": 953.0,
      "completions/mean_length": 515.4479370117188,
      "completions/mean_terminated_length": 515.4479370117188,
      "completions/min_length": 270.0,
      "completions/min_terminated_length": 270.0,
      "entropy": 0.0980682373046875,
      "epoch": 0.6204733390362133,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03175772726535797,
      "kl": 0.0025773048400878906,
      "learning_rate": 6.639506067845698e-06,
      "loss": 0.0052,
      "num_tokens": 28061562.0,
      "reward": 1.0176376104354858,
      "reward_std": 0.2250736653804779,
      "rewards/cross_solution_unittest_reward/mean": 0.44822562485933304,
      "rewards/cross_solution_unittest_reward/std": 0.2024165391921997,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 937.0,
      "completions/max_terminated_length": 937.0,
      "completions/mean_length": 493.3567810058594,
      "completions/mean_terminated_length": 493.3567810058594,
      "completions/min_length": 291.0,
      "completions/min_terminated_length": 291.0,
      "entropy": 0.0994415283203125,
      "epoch": 0.6295979469632165,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.033092062920331955,
      "kl": 0.0027437210083007812,
      "learning_rate": 6.3719229464935915e-06,
      "loss": 0.0023,
      "num_tokens": 28497043.0,
      "reward": 1.0380425453186035,
      "reward_std": 0.23625622689723969,
      "rewards/cross_solution_unittest_reward/mean": 0.4579319655895233,
      "rewards/cross_solution_unittest_reward/std": 0.20026205480098724,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1090.0,
      "completions/max_terminated_length": 1090.0,
      "completions/mean_length": 518.9921875,
      "completions/mean_terminated_length": 518.9921875,
      "completions/min_length": 243.0,
      "completions/min_terminated_length": 243.0,
      "entropy": 0.0950927734375,
      "epoch": 0.6387225548902196,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03034958802163601,
      "kl": 0.0027964115142822266,
      "learning_rate": 6.107298936826086e-06,
      "loss": -0.001,
      "num_tokens": 28942272.0,
      "reward": 1.1260254383087158,
      "reward_std": 0.3392567038536072,
      "rewards/cross_solution_unittest_reward/mean": 0.4806702435016632,
      "rewards/cross_solution_unittest_reward/std": 0.2181769162416458,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.96875,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 983.0,
      "completions/mean_length": 520.0963745117188,
      "completions/mean_terminated_length": 504.05499267578125,
      "completions/min_length": 282.0,
      "completions/min_terminated_length": 282.0,
      "entropy": 0.1016082763671875,
      "epoch": 0.6478471628172227,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03144148737192154,
      "kl": 0.0028443336486816406,
      "learning_rate": 5.845849869981137e-06,
      "loss": 0.0139,
      "num_tokens": 29354421.0,
      "reward": 0.683957576751709,
      "reward_std": 0.19793997704982758,
      "rewards/cross_solution_unittest_reward/mean": 0.31807853654026985,
      "rewards/cross_solution_unittest_reward/std": 0.18481089919805527,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1151.0,
      "completions/max_terminated_length": 1151.0,
      "completions/mean_length": 519.0833740234375,
      "completions/mean_terminated_length": 519.0833740234375,
      "completions/min_length": 273.0,
      "completions/min_terminated_length": 273.0,
      "entropy": 0.10321044921875,
      "epoch": 0.6569717707442259,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.42725464701652527,
      "kl": 0.03642606735229492,
      "learning_rate": 5.587788987567785e-06,
      "loss": 0.0002,
      "num_tokens": 29764661.0,
      "reward": 1.0907951593399048,
      "reward_std": 0.25230711698532104,
      "rewards/cross_solution_unittest_reward/mean": 0.47279611229896545,
      "rewards/cross_solution_unittest_reward/std": 0.2695095092058182,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 791.0,
      "completions/max_terminated_length": 791.0,
      "completions/mean_length": 481.8125,
      "completions/mean_terminated_length": 481.8125,
      "completions/min_length": 261.0,
      "completions/min_terminated_length": 261.0,
      "entropy": 0.1062469482421875,
      "epoch": 0.666096378671229,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.033676039427518845,
      "kl": 0.003365039825439453,
      "learning_rate": 5.333326767743263e-06,
      "loss": 0.0049,
      "num_tokens": 30167253.0,
      "reward": 1.1704354286193848,
      "reward_std": 0.26924002170562744,
      "rewards/cross_solution_unittest_reward/mean": 0.503605455160141,
      "rewards/cross_solution_unittest_reward/std": 0.24150928109884262,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1028.0,
      "completions/max_terminated_length": 1028.0,
      "completions/mean_length": 531.5989990234375,
      "completions/mean_terminated_length": 531.5989990234375,
      "completions/min_length": 228.0,
      "completions/min_terminated_length": 228.0,
      "entropy": 0.098541259765625,
      "epoch": 0.6752209865982322,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.029249846935272217,
      "kl": 0.0035860538482666016,
      "learning_rate": 5.082670753543961e-06,
      "loss": 0.0039,
      "num_tokens": 30577827.0,
      "reward": 1.0187840461730957,
      "reward_std": 0.27179187536239624,
      "rewards/cross_solution_unittest_reward/mean": 0.4446810930967331,
      "rewards/cross_solution_unittest_reward/std": 0.2518850564956665,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1103.0,
      "completions/max_terminated_length": 1103.0,
      "completions/mean_length": 513.2943115234375,
      "completions/mean_terminated_length": 513.2943115234375,
      "completions/min_length": 238.0,
      "completions/min_terminated_length": 238.0,
      "entropy": 0.1102142333984375,
      "epoch": 0.6843455945252352,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.029439395293593407,
      "kl": 0.002677440643310547,
      "learning_rate": 4.836025383610382e-06,
      "loss": 0.0061,
      "num_tokens": 30956420.0,
      "reward": 1.0420805215835571,
      "reward_std": 0.2407897710800171,
      "rewards/cross_solution_unittest_reward/mean": 0.4538075774908066,
      "rewards/cross_solution_unittest_reward/std": 0.22261689603328705,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.96875,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 958.0,
      "completions/mean_length": 490.7734375,
      "completions/mean_terminated_length": 474.57855224609375,
      "completions/min_length": 233.0,
      "completions/min_terminated_length": 233.0,
      "entropy": 0.107452392578125,
      "epoch": 0.6934702024522383,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.032771483063697815,
      "kl": 0.0031430721282958984,
      "learning_rate": 4.593591825444028e-06,
      "loss": 0.0142,
      "num_tokens": 31310765.0,
      "reward": 0.8610422015190125,
      "reward_std": 0.2242804616689682,
      "rewards/cross_solution_unittest_reward/mean": 0.38095080852508545,
      "rewards/cross_solution_unittest_reward/std": 0.2320489063858986,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 2310.0,
      "completions/max_terminated_length": 2310.0,
      "completions/mean_length": 502.9661560058594,
      "completions/mean_terminated_length": 502.9661560058594,
      "completions/min_length": 240.0,
      "completions/min_terminated_length": 240.0,
      "entropy": 0.10906982421875,
      "epoch": 0.7025948103792415,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.036506183445453644,
      "kl": 0.0029113292694091797,
      "learning_rate": 4.355567811332311e-06,
      "loss": -0.0079,
      "num_tokens": 31718384.0,
      "reward": 1.0806396007537842,
      "reward_std": 0.2699912488460541,
      "rewards/cross_solution_unittest_reward/mean": 0.4672403484582901,
      "rewards/cross_solution_unittest_reward/std": 0.2118784487247467,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1254.0,
      "completions/max_terminated_length": 1254.0,
      "completions/mean_length": 517.2578125,
      "completions/mean_terminated_length": 517.2578125,
      "completions/min_length": 228.0,
      "completions/min_terminated_length": 228.0,
      "entropy": 0.09716796875,
      "epoch": 0.7117194183062446,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.034147728234529495,
      "kl": 0.0033721923828125,
      "learning_rate": 4.12214747707527e-06,
      "loss": 0.0056,
      "num_tokens": 32162219.0,
      "reward": 1.0363221168518066,
      "reward_std": 0.3166886568069458,
      "rewards/cross_solution_unittest_reward/mean": 0.4502222388982773,
      "rewards/cross_solution_unittest_reward/std": 0.25721245259046555,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.953125,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 3357.0,
      "completions/mean_length": 533.2734375,
      "completions/mean_terminated_length": 509.251953125,
      "completions/min_length": 263.0,
      "completions/min_terminated_length": 263.0,
      "entropy": 0.1002655029296875,
      "epoch": 0.7208440262332478,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.029844792559742928,
      "kl": 0.0024819374084472656,
      "learning_rate": 3.893521203645618e-06,
      "loss": 0.0208,
      "num_tokens": 32565772.0,
      "reward": 1.0131210088729858,
      "reward_std": 0.24623362720012665,
      "rewards/cross_solution_unittest_reward/mean": 0.44952526688575745,
      "rewards/cross_solution_unittest_reward/std": 0.233152836561203,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 822.0,
      "completions/max_terminated_length": 822.0,
      "completions/mean_length": 475.51824951171875,
      "completions/mean_terminated_length": 475.51824951171875,
      "completions/min_length": 223.0,
      "completions/min_terminated_length": 223.0,
      "entropy": 0.099822998046875,
      "epoch": 0.7299686341602509,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03284022584557533,
      "kl": 0.003690004348754883,
      "learning_rate": 3.6698754619112974e-06,
      "loss": -0.0014,
      "num_tokens": 32926955.0,
      "reward": 0.9502054452896118,
      "reward_std": 0.23533646762371063,
      "rewards/cross_solution_unittest_reward/mean": 0.43129874020814896,
      "rewards/cross_solution_unittest_reward/std": 0.2330498918890953,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 962.0,
      "completions/max_terminated_length": 962.0,
      "completions/mean_length": 487.8177185058594,
      "completions/mean_terminated_length": 487.8177185058594,
      "completions/min_length": 200.0,
      "completions/min_terminated_length": 200.0,
      "entropy": 0.1024932861328125,
      "epoch": 0.739093242087254,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.033087749034166336,
      "kl": 0.002620697021484375,
      "learning_rate": 3.4513926605471504e-06,
      "loss": -0.0118,
      "num_tokens": 33311189.0,
      "reward": 1.0504975318908691,
      "reward_std": 0.2200670838356018,
      "rewards/cross_solution_unittest_reward/mean": 0.4615302234888077,
      "rewards/cross_solution_unittest_reward/std": 0.20018253475427628,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1020.0,
      "completions/max_terminated_length": 1020.0,
      "completions/mean_length": 479.33074951171875,
      "completions/mean_terminated_length": 479.33074951171875,
      "completions/min_length": 203.0,
      "completions/min_terminated_length": 203.0,
      "entropy": 0.1038665771484375,
      "epoch": 0.7482178500142572,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03461228311061859,
      "kl": 0.002687692642211914,
      "learning_rate": 3.2382509972598087e-06,
      "loss": 0.0073,
      "num_tokens": 33701588.0,
      "reward": 0.9868416786193848,
      "reward_std": 0.23717579245567322,
      "rewards/cross_solution_unittest_reward/mean": 0.4351993724703789,
      "rewards/cross_solution_unittest_reward/std": 0.2459121197462082,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1788.0,
      "completions/max_terminated_length": 1788.0,
      "completions/mean_length": 484.34375,
      "completions/mean_terminated_length": 484.34375,
      "completions/min_length": 252.0,
      "completions/min_terminated_length": 252.0,
      "entropy": 0.1046905517578125,
      "epoch": 0.7573424579412603,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.032532352954149246,
      "kl": 0.0031991004943847656,
      "learning_rate": 3.0306243134470668e-06,
      "loss": 0.0136,
      "num_tokens": 34096832.0,
      "reward": 0.9118785858154297,
      "reward_std": 0.27310025691986084,
      "rewards/cross_solution_unittest_reward/mean": 0.39596662670373917,
      "rewards/cross_solution_unittest_reward/std": 0.23184479027986526,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.953125,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 814.0,
      "completions/mean_length": 497.4140625,
      "completions/mean_terminated_length": 473.1102294921875,
      "completions/min_length": 261.0,
      "completions/min_terminated_length": 261.0,
      "entropy": 0.112152099609375,
      "epoch": 0.7664670658682635,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.030421627685427666,
      "kl": 0.0025551319122314453,
      "learning_rate": 2.8286819524103657e-06,
      "loss": 0.015,
      "num_tokens": 34443807.0,
      "reward": 1.0494276285171509,
      "reward_std": 0.24891619384288788,
      "rewards/cross_solution_unittest_reward/mean": 0.4558604210615158,
      "rewards/cross_solution_unittest_reward/std": 0.21075250953435898,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 963.0,
      "completions/mean_length": 498.6770935058594,
      "completions/mean_terminated_length": 490.6214294433594,
      "completions/min_length": 267.0,
      "completions/min_terminated_length": 267.0,
      "entropy": 0.1007080078125,
      "epoch": 0.7755916737952666,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.030757782980799675,
      "kl": 0.0036520957946777344,
      "learning_rate": 2.6325886212359496e-06,
      "loss": 0.0108,
      "num_tokens": 34874923.0,
      "reward": 1.0011277198791504,
      "reward_std": 0.2030499428510666,
      "rewards/cross_solution_unittest_reward/mean": 0.44105569273233414,
      "rewards/cross_solution_unittest_reward/std": 0.19068162888288498,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1026.0,
      "completions/max_terminated_length": 1026.0,
      "completions/mean_length": 480.3359375,
      "completions/mean_terminated_length": 480.3359375,
      "completions/min_length": 243.0,
      "completions/min_terminated_length": 243.0,
      "entropy": 0.108489990234375,
      "epoch": 0.7847162817222697,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.034772977232933044,
      "kl": 0.0031616687774658203,
      "learning_rate": 2.4425042564574186e-06,
      "loss": -0.0031,
      "num_tokens": 35288116.0,
      "reward": 0.9796751737594604,
      "reward_std": 0.23475846648216248,
      "rewards/cross_solution_unittest_reward/mean": 0.43246912956237793,
      "rewards/cross_solution_unittest_reward/std": 0.19666722416877747,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 3303.0,
      "completions/mean_length": 531.0416870117188,
      "completions/mean_terminated_length": 523.0704956054688,
      "completions/min_length": 260.0,
      "completions/min_terminated_length": 260.0,
      "entropy": 0.10565185546875,
      "epoch": 0.7938408896492729,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.039444971829652786,
      "kl": 0.0033414363861083984,
      "learning_rate": 2.2585838936091753e-06,
      "loss": 0.0083,
      "num_tokens": 35724444.0,
      "reward": 1.1217626333236694,
      "reward_std": 0.223866805434227,
      "rewards/cross_solution_unittest_reward/mean": 0.48780302703380585,
      "rewards/cross_solution_unittest_reward/std": 0.22542807459831238,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 878.0,
      "completions/max_terminated_length": 878.0,
      "completions/mean_length": 442.8828125,
      "completions/mean_terminated_length": 442.8828125,
      "completions/min_length": 235.0,
      "completions/min_terminated_length": 235.0,
      "entropy": 0.1145172119140625,
      "epoch": 0.802965497576276,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03573368489742279,
      "kl": 0.0030510425567626953,
      "learning_rate": 2.0809775407772505e-06,
      "loss": 0.0084,
      "num_tokens": 36081911.0,
      "reward": 1.0845204591751099,
      "reward_std": 0.2791823744773865,
      "rewards/cross_solution_unittest_reward/mean": 0.461356058716774,
      "rewards/cross_solution_unittest_reward/std": 0.23068440705537796,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 852.0,
      "completions/max_terminated_length": 852.0,
      "completions/mean_length": 518.3568115234375,
      "completions/mean_terminated_length": 518.3568115234375,
      "completions/min_length": 313.0,
      "completions/min_terminated_length": 313.0,
      "entropy": 0.093414306640625,
      "epoch": 0.8120901055032792,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03187531977891922,
      "kl": 0.002918720245361328,
      "learning_rate": 1.9098300562505266e-06,
      "loss": 0.0067,
      "num_tokens": 36510360.0,
      "reward": 0.9905633330345154,
      "reward_std": 0.20145916938781738,
      "rewards/cross_solution_unittest_reward/mean": 0.4334751144051552,
      "rewards/cross_solution_unittest_reward/std": 0.23444491624832153,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 986.0,
      "completions/max_terminated_length": 986.0,
      "completions/mean_length": 490.1953125,
      "completions/mean_terminated_length": 490.1953125,
      "completions/min_length": 258.0,
      "completions/min_terminated_length": 258.0,
      "entropy": 0.0989532470703125,
      "epoch": 0.8212147134302823,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.05914483591914177,
      "kl": 0.008898019790649414,
      "learning_rate": 1.74528103037226e-06,
      "loss": 0.0025,
      "num_tokens": 36925595.0,
      "reward": 0.9185196757316589,
      "reward_std": 0.3170742094516754,
      "rewards/cross_solution_unittest_reward/mean": 0.3952532112598419,
      "rewards/cross_solution_unittest_reward/std": 0.2790588438510895,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 980.0,
      "completions/max_terminated_length": 980.0,
      "completions/mean_length": 483.3567810058594,
      "completions/mean_terminated_length": 483.3567810058594,
      "completions/min_length": 230.0,
      "completions/min_terminated_length": 230.0,
      "entropy": 0.10723876953125,
      "epoch": 0.8303393213572854,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03479889780282974,
      "kl": 0.0028510093688964844,
      "learning_rate": 1.587464671688187e-06,
      "loss": 0.001,
      "num_tokens": 37317868.0,
      "reward": 0.9646342992782593,
      "reward_std": 0.18958084285259247,
      "rewards/cross_solution_unittest_reward/mean": 0.4295145273208618,
      "rewards/cross_solution_unittest_reward/std": 0.2108893096446991,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 956.0,
      "completions/max_terminated_length": 956.0,
      "completions/mean_length": 501.15887451171875,
      "completions/mean_terminated_length": 501.15887451171875,
      "completions/min_length": 237.0,
      "completions/min_terminated_length": 237.0,
      "entropy": 0.1046600341796875,
      "epoch": 0.8394639292842886,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.030385170131921768,
      "kl": 0.0032634735107421875,
      "learning_rate": 1.436509697484111e-06,
      "loss": 0.0069,
      "num_tokens": 37734865.0,
      "reward": 0.9471985697746277,
      "reward_std": 0.24106745421886444,
      "rewards/cross_solution_unittest_reward/mean": 0.4128854274749756,
      "rewards/cross_solution_unittest_reward/std": 0.24737022817134857,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.96875,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1438.0,
      "completions/mean_length": 564.9948120117188,
      "completions/mean_terminated_length": 549.1884765625,
      "completions/min_length": 280.0,
      "completions/min_terminated_length": 280.0,
      "entropy": 0.087799072265625,
      "epoch": 0.8485885372112917,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2799060642719269,
      "kl": 0.017560720443725586,
      "learning_rate": 1.2925392288022299e-06,
      "loss": 0.0064,
      "num_tokens": 38171079.0,
      "reward": 0.9303452372550964,
      "reward_std": 0.21974782645702362,
      "rewards/cross_solution_unittest_reward/mean": 0.41144028306007385,
      "rewards/cross_solution_unittest_reward/std": 0.23167048394680023,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.96875,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 1033.0,
      "completions/mean_length": 545.6354370117188,
      "completions/mean_terminated_length": 529.727783203125,
      "completions/min_length": 204.0,
      "completions/min_terminated_length": 204.0,
      "entropy": 0.092437744140625,
      "epoch": 0.8577131451382949,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02881380170583725,
      "kl": 0.0027322769165039062,
      "learning_rate": 1.1556706900218572e-06,
      "loss": 0.0152,
      "num_tokens": 38596603.0,
      "reward": 0.9431265592575073,
      "reward_std": 0.3126261830329895,
      "rewards/cross_solution_unittest_reward/mean": 0.4096960350871086,
      "rewards/cross_solution_unittest_reward/std": 0.274701751768589,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 775.0,
      "completions/max_terminated_length": 775.0,
      "completions/mean_length": 492.7734375,
      "completions/mean_terminated_length": 492.7734375,
      "completions/min_length": 255.0,
      "completions/min_terminated_length": 255.0,
      "entropy": 0.1008758544921875,
      "epoch": 0.866837753065298,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.030817339196801186,
      "kl": 0.003123760223388672,
      "learning_rate": 1.0260157130864178e-06,
      "loss": -0.0065,
      "num_tokens": 39042508.0,
      "reward": 0.98272705078125,
      "reward_std": 0.2458946406841278,
      "rewards/cross_solution_unittest_reward/mean": 0.42810578644275665,
      "rewards/cross_solution_unittest_reward/std": 0.19119084626436234,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 1011.0,
      "completions/max_terminated_length": 1011.0,
      "completions/mean_length": 520.4036865234375,
      "completions/mean_terminated_length": 520.4036865234375,
      "completions/min_length": 241.0,
      "completions/min_terminated_length": 241.0,
      "entropy": 0.10015869140625,
      "epoch": 0.8759623609923011,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03871850296854973,
      "kl": 0.0033326148986816406,
      "learning_rate": 9.036800464548157e-07,
      "loss": -0.0045,
      "num_tokens": 39430183.0,
      "reward": 0.944609522819519,
      "reward_std": 0.2243545949459076,
      "rewards/cross_solution_unittest_reward/mean": 0.40971074998378754,
      "rewards/cross_solution_unittest_reward/std": 0.22074688971042633,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 832.0,
      "completions/mean_length": 457.82293701171875,
      "completions/mean_terminated_length": 449.66058349609375,
      "completions/min_length": 227.0,
      "completions/min_terminated_length": 227.0,
      "entropy": 0.1109771728515625,
      "epoch": 0.8850869689193043,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.035750389099121094,
      "kl": 0.0031118392944335938,
      "learning_rate": 7.887634688515e-07,
      "loss": -0.0005,
      "num_tokens": 39751387.0,
      "reward": 1.1470212936401367,
      "reward_std": 0.19520321488380432,
      "rewards/cross_solution_unittest_reward/mean": 0.5024754554033279,
      "rewards/cross_solution_unittest_reward/std": 0.2238490730524063,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 856.0,
      "completions/max_terminated_length": 856.0,
      "completions/mean_length": 528.5989990234375,
      "completions/mean_terminated_length": 528.5989990234375,
      "completions/min_length": 285.0,
      "completions/min_terminated_length": 285.0,
      "entropy": 0.0987396240234375,
      "epoch": 0.8942115768463074,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02971605211496353,
      "kl": 0.002849102020263672,
      "learning_rate": 6.813597078854772e-07,
      "loss": 0.0069,
      "num_tokens": 40197641.0,
      "reward": 1.0776381492614746,
      "reward_std": 0.2849399149417877,
      "rewards/cross_solution_unittest_reward/mean": 0.4694649428129196,
      "rewards/cross_solution_unittest_reward/std": 0.28404664993286133,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 911.0,
      "completions/max_terminated_length": 911.0,
      "completions/mean_length": 493.4036560058594,
      "completions/mean_terminated_length": 493.4036560058594,
      "completions/min_length": 209.0,
      "completions/min_terminated_length": 209.0,
      "entropy": 0.1012420654296875,
      "epoch": 0.9033361847733106,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03130873292684555,
      "kl": 0.0025815963745117188,
      "learning_rate": 5.815563636047539e-07,
      "loss": -0.0035,
      "num_tokens": 40569268.0,
      "reward": 1.0923125743865967,
      "reward_std": 0.24533653259277344,
      "rewards/cross_solution_unittest_reward/mean": 0.4748411625623703,
      "rewards/cross_solution_unittest_reward/std": 0.2197420746088028,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "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": -5.0,
      "completions/max_length": 914.0,
      "completions/max_terminated_length": 914.0,
      "completions/mean_length": 474.8125,
      "completions/mean_terminated_length": 474.8125,
      "completions/min_length": 261.0,
      "completions/min_terminated_length": 261.0,
      "entropy": 0.1113739013671875,
      "epoch": 0.9124607927003137,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03394750878214836,
      "kl": 0.003151416778564453,
      "learning_rate": 4.894348370484648e-07,
      "loss": 0.0001,
      "num_tokens": 40993476.0,
      "reward": 0.9635044932365417,
      "reward_std": 0.24135762453079224,
      "rewards/cross_solution_unittest_reward/mean": 0.4190831929445267,
      "rewards/cross_solution_unittest_reward/std": 0.24994324147701263,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "step": 100
    },
    {
      "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": -4.984375,
      "completions/max_length": 3584.0,
      "completions/max_terminated_length": 848.0,
      "completions/mean_length": 488.19012451171875,
      "completions/mean_terminated_length": 480.1070556640625,
      "completions/min_length": 248.0,
      "completions/min_terminated_length": 248.0,
      "entropy": 0.0947265625,
      "epoch": 0.9215854006273168,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.034001559019088745,
      "kl": 0.003081798553466797,
      "learning_rate": 4.0507026385502747e-07,
      "loss": 0.0004,
      "num_tokens": 41383509.0,
      "reward": 1.047755479812622,
      "reward_std": 0.23132114112377167,
      "rewards/cross_solution_unittest_reward/mean": 0.4617375209927559,
      "rewards/cross_solution_unittest_reward/std": 0.20788954198360443,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "step": 101
    },
    {
      "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": -5.0,
      "completions/max_length": 995.0,
      "completions/max_terminated_length": 995.0,
      "completions/mean_length": 502.48699951171875,
      "completions/mean_terminated_length": 502.48699951171875,
      "completions/min_length": 252.0,
      "completions/min_terminated_length": 252.0,
      "entropy": 0.094207763671875,
      "epoch": 0.93071000855432,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.028330322355031967,
      "kl": 0.002512693405151367,
      "learning_rate": 3.2853145298042954e-07,
      "loss": -0.0035,
      "num_tokens": 41811784.0,
      "reward": 1.1942728757858276,
      "reward_std": 0.28333157300949097,
      "rewards/cross_solution_unittest_reward/mean": 0.5126944035291672,
      "rewards/cross_solution_unittest_reward/std": 0.1964806318283081,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "step": 102
    },
    {
      "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": -5.0,
      "completions/max_length": 968.0,
      "completions/max_terminated_length": 968.0,
      "completions/mean_length": 471.36199951171875,
      "completions/mean_terminated_length": 471.36199951171875,
      "completions/min_length": 241.0,
      "completions/min_terminated_length": 241.0,
      "entropy": 0.103363037109375,
      "epoch": 0.939834616481323,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.0329383909702301,
      "kl": 0.0033469200134277344,
      "learning_rate": 2.5988083057666534e-07,
      "loss": -0.0076,
      "num_tokens": 42183707.0,
      "reward": 1.1951045989990234,
      "reward_std": 0.2977498471736908,
      "rewards/cross_solution_unittest_reward/mean": 0.514451876282692,
      "rewards/cross_solution_unittest_reward/std": 0.251402385532856,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "step": 103
    },
    {
      "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": -5.0,
      "completions/max_length": 866.0,
      "completions/max_terminated_length": 866.0,
      "completions/mean_length": 500.29168701171875,
      "completions/mean_terminated_length": 500.29168701171875,
      "completions/min_length": 287.0,
      "completions/min_terminated_length": 287.0,
      "entropy": 0.09759521484375,
      "epoch": 0.9489592244083263,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.0342080220580101,
      "kl": 0.0033740997314453125,
      "learning_rate": 1.9917438907606556e-07,
      "loss": 0.0017,
      "num_tokens": 42602043.0,
      "reward": 1.004610300064087,
      "reward_std": 0.2067084014415741,
      "rewards/cross_solution_unittest_reward/mean": 0.44767728447914124,
      "rewards/cross_solution_unittest_reward/std": 0.19676169008016586,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "step": 104
    },
    {
      "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": -5.0,
      "completions/max_length": 942.0,
      "completions/max_terminated_length": 942.0,
      "completions/mean_length": 461.2161560058594,
      "completions/mean_terminated_length": 461.2161560058594,
      "completions/min_length": 275.0,
      "completions/min_terminated_length": 275.0,
      "entropy": 0.1075592041015625,
      "epoch": 0.9580838323353293,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.0386245995759964,
      "kl": 0.0034677982330322266,
      "learning_rate": 1.464616415230702e-07,
      "loss": -0.0062,
      "num_tokens": 42963262.0,
      "reward": 0.8114880323410034,
      "reward_std": 0.26597505807876587,
      "rewards/cross_solution_unittest_reward/mean": 0.3618558570742607,
      "rewards/cross_solution_unittest_reward/std": 0.23664288967847824,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "step": 105
    },
    {
      "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": -5.0,
      "completions/max_length": 811.0,
      "completions/max_terminated_length": 811.0,
      "completions/mean_length": 524.5104370117188,
      "completions/mean_terminated_length": 524.5104370117188,
      "completions/min_length": 274.0,
      "completions/min_terminated_length": 274.0,
      "entropy": 0.093170166015625,
      "epoch": 0.9672084402623324,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.028961867094039917,
      "kl": 0.0027861595153808594,
      "learning_rate": 1.0178558119067316e-07,
      "loss": -0.0063,
      "num_tokens": 43390042.0,
      "reward": 1.000401496887207,
      "reward_std": 0.30445563793182373,
      "rewards/cross_solution_unittest_reward/mean": 0.43068091571331024,
      "rewards/cross_solution_unittest_reward/std": 0.25178908556699753,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "step": 106
    },
    {
      "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": -5.0,
      "completions/max_length": 885.0,
      "completions/max_terminated_length": 885.0,
      "completions/mean_length": 494.6015625,
      "completions/mean_terminated_length": 494.6015625,
      "completions/min_length": 282.0,
      "completions/min_terminated_length": 282.0,
      "entropy": 0.1114959716796875,
      "epoch": 0.9763330481893356,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.034926630556583405,
      "kl": 0.0031135082244873047,
      "learning_rate": 6.51826465144978e-08,
      "loss": -0.0033,
      "num_tokens": 43794513.0,
      "reward": 1.0083611011505127,
      "reward_std": 0.23411409556865692,
      "rewards/cross_solution_unittest_reward/mean": 0.4444109722971916,
      "rewards/cross_solution_unittest_reward/std": 0.2296873927116394,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "step": 107
    },
    {
      "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": -5.0,
      "completions/max_length": 1069.0,
      "completions/max_terminated_length": 1069.0,
      "completions/mean_length": 530.9869995117188,
      "completions/mean_terminated_length": 530.9869995117188,
      "completions/min_length": 252.0,
      "completions/min_terminated_length": 252.0,
      "entropy": 0.0896759033203125,
      "epoch": 0.9854576561163387,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.029170462861657143,
      "kl": 0.002541780471801758,
      "learning_rate": 3.668269137308666e-08,
      "loss": 0.0042,
      "num_tokens": 44236428.0,
      "reward": 1.2101051807403564,
      "reward_std": 0.225676029920578,
      "rewards/cross_solution_unittest_reward/mean": 0.525607168674469,
      "rewards/cross_solution_unittest_reward/std": 0.20895672589540482,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "step": 108
    },
    {
      "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": -5.0,
      "completions/max_length": 841.0,
      "completions/max_terminated_length": 841.0,
      "completions/mean_length": 468.3489685058594,
      "completions/mean_terminated_length": 468.3489685058594,
      "completions/min_length": 216.0,
      "completions/min_terminated_length": 216.0,
      "entropy": 0.101654052734375,
      "epoch": 0.9945822640433419,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.03142247721552849,
      "kl": 0.002844095230102539,
      "learning_rate": 1.630896073864352e-08,
      "loss": 0.0032,
      "num_tokens": 44609178.0,
      "reward": 1.2266151905059814,
      "reward_std": 0.24263408780097961,
      "rewards/cross_solution_unittest_reward/mean": 0.5332132428884506,
      "rewards/cross_solution_unittest_reward/std": 0.22311152517795563,
      "rewards/format_reward/mean": 0.0,
      "rewards/format_reward/std": 0.0,
      "step": 109
    },
    {
      "epoch": 0.9945822640433419,
      "step": 109,
      "total_flos": 0.0,
      "train_loss": 0.0030027573193699003,
      "train_runtime": 43116.833,
      "train_samples_per_second": 0.03,
      "train_steps_per_second": 0.003
    }
  ],
  "logging_steps": 1,
  "max_steps": 110,
  "num_input_tokens_seen": 44609178,
  "num_train_epochs": 1,
  "save_steps": 25,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}