trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.6289308176100629,
  "eval_steps": 500,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.006289308176100629,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.016325591132044792,
      "kl": 0.00010249206752632745,
      "learning_rate": 0.0,
      "loss": 0.0,
      "num_tokens": 68316.0,
      "reward": -0.375,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.375,
      "rewards/length_penalty/std": 0.22360680997371674,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 1
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.012578616352201259,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.07733656466007233,
      "kl": 8.590362267568707e-05,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.0,
      "num_tokens": 137112.0,
      "reward": -0.21875,
      "reward_std": 0.27900636196136475,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 2
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.018867924528301886,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.017757171764969826,
      "kl": 0.00010768742322397884,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0,
      "num_tokens": 205628.0,
      "reward": -0.15625,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 3
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.025157232704402517,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.018300458788871765,
      "kl": 0.00010197597111982759,
      "learning_rate": 1.5e-06,
      "loss": 0.0,
      "num_tokens": 276860.0,
      "reward": -0.25,
      "reward_std": 0.26933756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 4
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.031446540880503145,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.032985903322696686,
      "kl": 0.00010790089436341077,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0,
      "num_tokens": 344820.0,
      "reward": -0.21875,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 5
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.03773584905660377,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.017706185579299927,
      "kl": 0.00011222602188354358,
      "learning_rate": 2.5e-06,
      "loss": 0.0,
      "num_tokens": 413964.0,
      "reward": -0.28125,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 6
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0440251572327044,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.0198284350335598,
      "kl": 0.00010518240742385387,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "num_tokens": 482264.0,
      "reward": -0.21875,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 7
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.050314465408805034,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.021514201536774635,
      "kl": 0.0001032647633110173,
      "learning_rate": 3.5e-06,
      "loss": 0.0,
      "num_tokens": 550552.0,
      "reward": -0.1875,
      "reward_std": 0.25,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.1875,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 8
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.05660377358490566,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.017368245869874954,
      "kl": 0.00010589188059384469,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0,
      "num_tokens": 619280.0,
      "reward": -0.3125,
      "reward_std": 0.21650634706020355,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 9
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.06289308176100629,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.01433493010699749,
      "kl": 9.300672081735684e-05,
      "learning_rate": 4.5e-06,
      "loss": 0.0,
      "num_tokens": 688348.0,
      "reward": -0.125,
      "reward_std": 0.14433756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.125,
      "rewards/length_penalty/std": 0.22360680997371674,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 10
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.06918238993710692,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.018409231677651405,
      "kl": 0.00011178611384821124,
      "learning_rate": 5e-06,
      "loss": 0.0,
      "num_tokens": 757088.0,
      "reward": -0.28125,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 11
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.07547169811320754,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.01990220881998539,
      "kl": 0.00010754678442026488,
      "learning_rate": 4.99847706754774e-06,
      "loss": 0.0,
      "num_tokens": 825452.0,
      "reward": -0.21875,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 12
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.08176100628930817,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.017681574448943138,
      "kl": 0.00011142175571876578,
      "learning_rate": 4.993910125649561e-06,
      "loss": 0.0,
      "num_tokens": 893748.0,
      "reward": -0.28125,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 13
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0880503144654088,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.015456760302186012,
      "kl": 9.743376904225443e-05,
      "learning_rate": 4.986304738420684e-06,
      "loss": 0.0,
      "num_tokens": 962436.0,
      "reward": -0.25,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 14
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.09433962264150944,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.017371585592627525,
      "kl": 0.00010160906458622776,
      "learning_rate": 4.975670171853926e-06,
      "loss": 0.0,
      "num_tokens": 1030892.0,
      "reward": -0.25,
      "reward_std": 0.26933756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 15
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.10062893081761007,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.01290306355804205,
      "kl": 0.0001026637382892659,
      "learning_rate": 4.962019382530521e-06,
      "loss": 0.0,
      "num_tokens": 1099392.0,
      "reward": -0.28125,
      "reward_std": 0.13466878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 16
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.1069182389937107,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.017305204644799232,
      "kl": 0.0001403935566486325,
      "learning_rate": 4.9453690018345144e-06,
      "loss": 0.0,
      "num_tokens": 1168952.0,
      "reward": -0.3125,
      "reward_std": 0.21650634706020355,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 17
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.11320754716981132,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.026463210582733154,
      "kl": 0.00010177845251746476,
      "learning_rate": 4.925739315689991e-06,
      "loss": 0.0,
      "num_tokens": 1237180.0,
      "reward": -0.28125,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 18
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.11949685534591195,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.013850911520421505,
      "kl": 0.00010606551040837076,
      "learning_rate": 4.903154239845798e-06,
      "loss": 0.0,
      "num_tokens": 1305796.0,
      "reward": -0.28125,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 19
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.12578616352201258,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.01833195611834526,
      "kl": 0.00010958383609249722,
      "learning_rate": 4.8776412907378845e-06,
      "loss": 0.0,
      "num_tokens": 1374084.0,
      "reward": -0.21875,
      "reward_std": 0.27900636196136475,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 20
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.1320754716981132,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.017120856791734695,
      "kl": 9.574014075042214e-05,
      "learning_rate": 4.849231551964771e-06,
      "loss": 0.0,
      "num_tokens": 1442040.0,
      "reward": -0.21875,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 21
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.13836477987421383,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.013351180590689182,
      "kl": 0.00011904810162377544,
      "learning_rate": 4.817959636416969e-06,
      "loss": 0.0,
      "num_tokens": 1510516.0,
      "reward": -0.25,
      "reward_std": 0.125,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 22
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.14465408805031446,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.019939737394452095,
      "kl": 0.0001281359654967673,
      "learning_rate": 4.783863644106502e-06,
      "loss": 0.0,
      "num_tokens": 1578756.0,
      "reward": -0.3125,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 23
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.1509433962264151,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.01900831237435341,
      "kl": 9.51789043028839e-05,
      "learning_rate": 4.746985115747918e-06,
      "loss": 0.0,
      "num_tokens": 1646764.0,
      "reward": -0.28125,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 24
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.15723270440251572,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01629672758281231,
      "kl": 0.00011995839122391772,
      "learning_rate": 4.707368982147318e-06,
      "loss": 0.0,
      "num_tokens": 1715904.0,
      "reward": -0.25,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 25
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.16352201257861634,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.019083648920059204,
      "kl": 0.0001452796786907129,
      "learning_rate": 4.665063509461098e-06,
      "loss": 0.0,
      "num_tokens": 1784896.0,
      "reward": -0.28125,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 26
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.16981132075471697,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.012859661132097244,
      "kl": 0.00012349323333182838,
      "learning_rate": 4.620120240391065e-06,
      "loss": 0.0,
      "num_tokens": 1853088.0,
      "reward": -0.3125,
      "reward_std": 0.125,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 27
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.1761006289308176,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02429352141916752,
      "kl": 0.00016827746003400534,
      "learning_rate": 4.572593931387604e-06,
      "loss": 0.0,
      "num_tokens": 1922692.0,
      "reward": -0.25,
      "reward_std": 0.25,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 28
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.18238993710691823,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.017776580527424812,
      "kl": 0.00010630501310515683,
      "learning_rate": 4.522542485937369e-06,
      "loss": 0.0,
      "num_tokens": 1991076.0,
      "reward": -0.375,
      "reward_std": 0.25,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.375,
      "rewards/length_penalty/std": 0.22360680997371674,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 29
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.18867924528301888,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.013591092079877853,
      "kl": 0.00010891661986534018,
      "learning_rate": 4.470026884016805e-06,
      "loss": 0.0,
      "num_tokens": 2059424.0,
      "reward": -0.15625,
      "reward_std": 0.13466878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 30
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.1949685534591195,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.012544853612780571,
      "kl": 0.00012365094880806282,
      "learning_rate": 4.415111107797445e-06,
      "loss": 0.0,
      "num_tokens": 2128544.0,
      "reward": -0.40625,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.40625,
      "rewards/length_penalty/std": 0.20155644416809082,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 31
    },
    {
      "completion_length": 4021.875,
      "completions/clipped_ratio": 0.9375,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 2910.0,
      "completions/mean_length": 4021.875,
      "completions/mean_terminated_length": 2910.0,
      "completions/min_length": 2910.0,
      "completions/min_terminated_length": 2910.0,
      "epoch": 0.20125786163522014,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.02055083028972149,
      "kl": 0.00012053796672262251,
      "learning_rate": 4.357862063693486e-06,
      "loss": 0.0,
      "num_tokens": 2196138.0,
      "reward": -0.1875,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.1875,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 32
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.20754716981132076,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01593278907239437,
      "kl": 0.0001260052995348815,
      "learning_rate": 4.2983495008466285e-06,
      "loss": 0.0,
      "num_tokens": 2264882.0,
      "reward": -0.21875,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 33
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.2138364779874214,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01661592908203602,
      "kl": 0.0001022304204525426,
      "learning_rate": 4.236645926147493e-06,
      "loss": 0.0,
      "num_tokens": 2333358.0,
      "reward": -0.125,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.125,
      "rewards/length_penalty/std": 0.22360680997371674,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 34
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.22012578616352202,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.025219203904271126,
      "kl": 0.00012560350478452165,
      "learning_rate": 4.172826515897146e-06,
      "loss": 0.0,
      "num_tokens": 2402414.0,
      "reward": -0.15625,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 35
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.22641509433962265,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01724395528435707,
      "kl": 0.00011793708654295187,
      "learning_rate": 4.106969024216348e-06,
      "loss": 0.0,
      "num_tokens": 2471462.0,
      "reward": -0.3125,
      "reward_std": 0.21650634706020355,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 36
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.23270440251572327,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.013114077970385551,
      "kl": 0.00012667544069699943,
      "learning_rate": 4.039153688314146e-06,
      "loss": 0.0,
      "num_tokens": 2539970.0,
      "reward": -0.21875,
      "reward_std": 0.13466878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 37
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.2389937106918239,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.012886935845017433,
      "kl": 0.00011692461885104422,
      "learning_rate": 3.969463130731183e-06,
      "loss": 0.0,
      "num_tokens": 2607850.0,
      "reward": -0.15625,
      "reward_std": 0.13466878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 38
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.24528301886792453,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.1053331270813942,
      "kl": 0.00012466574844438583,
      "learning_rate": 3.897982258676867e-06,
      "loss": 0.0,
      "num_tokens": 2676374.0,
      "reward": -0.1875,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.1875,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 39
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.25157232704402516,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 0.0093563012778759,
      "kl": 0.00012254506509634666,
      "learning_rate": 3.824798160583012e-06,
      "loss": 0.0,
      "num_tokens": 2745118.0,
      "reward": -0.28125,
      "reward_std": 0.0625,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 40
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.2578616352201258,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.04284551739692688,
      "kl": 0.00014869343431200832,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.0,
      "num_tokens": 2814710.0,
      "reward": -0.25,
      "reward_std": 0.26933756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 41
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.2641509433962264,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 0.00853828527033329,
      "kl": 0.0001323007199971471,
      "learning_rate": 3.6736789069647273e-06,
      "loss": 0.0,
      "num_tokens": 2883202.0,
      "reward": -0.3125,
      "reward_std": 0.07216878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 42
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.27044025157232704,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.016302255913615227,
      "kl": 0.00010226479753328022,
      "learning_rate": 3.595927866972694e-06,
      "loss": 0.0,
      "num_tokens": 2951082.0,
      "reward": -0.125,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.125,
      "rewards/length_penalty/std": 0.22360680997371674,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 43
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.27672955974842767,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.013052207417786121,
      "kl": 0.00010622412810334936,
      "learning_rate": 3.516841607689501e-06,
      "loss": 0.0,
      "num_tokens": 3019634.0,
      "reward": -0.125,
      "reward_std": 0.125,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.125,
      "rewards/length_penalty/std": 0.22360680997371674,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 44
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.2830188679245283,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.018321262672543526,
      "kl": 0.00015826924209250137,
      "learning_rate": 3.436516483539781e-06,
      "loss": 0.0,
      "num_tokens": 3088074.0,
      "reward": -0.34375,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.34375,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 45
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.2893081761006289,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.0323839895427227,
      "kl": 0.00015421856733155437,
      "learning_rate": 3.3550503583141726e-06,
      "loss": 0.0,
      "num_tokens": 3156594.0,
      "reward": -0.25,
      "reward_std": 0.26933756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 46
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.29559748427672955,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.019736355170607567,
      "kl": 0.00012661038999794982,
      "learning_rate": 3.272542485937369e-06,
      "loss": 0.0,
      "num_tokens": 3225158.0,
      "reward": -0.28125,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 47
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.3018867924528302,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.012744598090648651,
      "kl": 0.00013378711992118042,
      "learning_rate": 3.189093389542498e-06,
      "loss": 0.0,
      "num_tokens": 3294386.0,
      "reward": -0.21875,
      "reward_std": 0.13466878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 48
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.3081761006289308,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01501748338341713,
      "kl": 0.00011261255713179708,
      "learning_rate": 3.1048047389991693e-06,
      "loss": 0.0,
      "num_tokens": 3363586.0,
      "reward": -0.21875,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 49
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.31446540880503143,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01936727948486805,
      "kl": 0.00010851330989680719,
      "learning_rate": 3.019779227044398e-06,
      "loss": 0.0,
      "num_tokens": 3432534.0,
      "reward": -0.21875,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 50
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.32075471698113206,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.013259466737508774,
      "kl": 0.00011628576248767786,
      "learning_rate": 2.9341204441673267e-06,
      "loss": 0.0,
      "num_tokens": 3501158.0,
      "reward": -0.15625,
      "reward_std": 0.13466878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 51
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.3270440251572327,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.04822860658168793,
      "kl": 0.0001678316984907724,
      "learning_rate": 2.847932752400164e-06,
      "loss": 0.0,
      "num_tokens": 3569454.0,
      "reward": -0.21875,
      "reward_std": 0.27900636196136475,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 52
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.3333333333333333,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.015680057927966118,
      "kl": 0.00011564494707272388,
      "learning_rate": 2.761321158169134e-06,
      "loss": 0.0,
      "num_tokens": 3637870.0,
      "reward": -0.21875,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 53
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.33962264150943394,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.018732121214270592,
      "kl": 0.00016443646381958388,
      "learning_rate": 2.6743911843603134e-06,
      "loss": 0.0,
      "num_tokens": 3706402.0,
      "reward": -0.28125,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 54
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.34591194968553457,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 0.008841070346534252,
      "kl": 0.0001238682962139137,
      "learning_rate": 2.587248741756253e-06,
      "loss": 0.0,
      "num_tokens": 3776738.0,
      "reward": -0.3125,
      "reward_std": 0.07216878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 55
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.3522012578616352,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.018190307542681694,
      "kl": 0.00014203431055648252,
      "learning_rate": 2.5e-06,
      "loss": 0.0,
      "num_tokens": 3845338.0,
      "reward": -0.3125,
      "reward_std": 0.25,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 56
    },
    {
      "completion_length": 4032.375,
      "completions/clipped_ratio": 0.9375,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 3078.0,
      "completions/mean_length": 4032.375,
      "completions/mean_terminated_length": 3078.0,
      "completions/min_length": 3078.0,
      "completions/min_terminated_length": 3078.0,
      "epoch": 0.3584905660377358,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.011392134241759777,
      "kl": 9.039508040586952e-05,
      "learning_rate": 2.4127512582437486e-06,
      "loss": 0.0,
      "num_tokens": 3912380.0,
      "reward": -0.125,
      "reward_std": 0.14433756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.125,
      "rewards/length_penalty/std": 0.22360680997371674,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 57
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.36477987421383645,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.018618181347846985,
      "kl": 0.00017752719577401876,
      "learning_rate": 2.325608815639687e-06,
      "loss": 0.0,
      "num_tokens": 3982232.0,
      "reward": -0.21875,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 58
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.3710691823899371,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 0.008148152381181717,
      "kl": 0.0001503387375123566,
      "learning_rate": 2.238678841830867e-06,
      "loss": 0.0,
      "num_tokens": 4051828.0,
      "reward": -0.28125,
      "reward_std": 0.0625,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 59
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.37735849056603776,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.015091936103999615,
      "kl": 0.00011921779878321104,
      "learning_rate": 2.1520672475998374e-06,
      "loss": 0.0,
      "num_tokens": 4120096.0,
      "reward": -0.1875,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.1875,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 60
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.3836477987421384,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01511855237185955,
      "kl": 0.00013280290295369923,
      "learning_rate": 2.0658795558326745e-06,
      "loss": 0.0,
      "num_tokens": 4188892.0,
      "reward": -0.1875,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.1875,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 61
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.389937106918239,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.0146819232031703,
      "kl": 0.00016415627032984048,
      "learning_rate": 1.9802207729556023e-06,
      "loss": 0.0,
      "num_tokens": 4257508.0,
      "reward": -0.25,
      "reward_std": 0.14433756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 62
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.39622641509433965,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.01902368664741516,
      "kl": 0.00015298420112230815,
      "learning_rate": 1.895195261000831e-06,
      "loss": 0.0,
      "num_tokens": 4326240.0,
      "reward": -0.25,
      "reward_std": 0.26933756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 63
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.4025157232704403,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01879284530878067,
      "kl": 0.000160826002684189,
      "learning_rate": 1.8109066104575023e-06,
      "loss": 0.0,
      "num_tokens": 4396008.0,
      "reward": -0.3125,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 64
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.4088050314465409,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.0197551678866148,
      "kl": 0.0001399103202857077,
      "learning_rate": 1.7274575140626318e-06,
      "loss": 0.0,
      "num_tokens": 4463884.0,
      "reward": -0.25,
      "reward_std": 0.26933756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 65
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.41509433962264153,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.018731985241174698,
      "kl": 0.0001391547757521039,
      "learning_rate": 1.6449496416858285e-06,
      "loss": 0.0,
      "num_tokens": 4533192.0,
      "reward": -0.15625,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 66
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.42138364779874216,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01650979183614254,
      "kl": 0.00014604274110752158,
      "learning_rate": 1.56348351646022e-06,
      "loss": 0.0,
      "num_tokens": 4601596.0,
      "reward": -0.3125,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 67
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.4276729559748428,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01574656181037426,
      "kl": 0.00016481637067045085,
      "learning_rate": 1.4831583923105e-06,
      "loss": 0.0,
      "num_tokens": 4670308.0,
      "reward": -0.25,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 68
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.4339622641509434,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.017636191099882126,
      "kl": 0.00013569381735578645,
      "learning_rate": 1.4040721330273063e-06,
      "loss": 0.0,
      "num_tokens": 4738980.0,
      "reward": -0.15625,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 69
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.44025157232704404,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.016505496576428413,
      "kl": 0.0001501081460446585,
      "learning_rate": 1.3263210930352737e-06,
      "loss": 0.0,
      "num_tokens": 4807088.0,
      "reward": -0.375,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.375,
      "rewards/length_penalty/std": 0.22360680997371674,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 70
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.44654088050314467,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.12390376627445221,
      "kl": 0.00020521143596852198,
      "learning_rate": 1.2500000000000007e-06,
      "loss": 0.0,
      "num_tokens": 4876132.0,
      "reward": -0.21875,
      "reward_std": 0.13466878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 71
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.4528301886792453,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.018132826313376427,
      "kl": 0.00014764678235223982,
      "learning_rate": 1.1752018394169882e-06,
      "loss": 0.0,
      "num_tokens": 4944576.0,
      "reward": -0.15625,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 72
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.4591194968553459,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.022564060986042023,
      "kl": 0.000160812556714518,
      "learning_rate": 1.1020177413231334e-06,
      "loss": 0.0,
      "num_tokens": 5013532.0,
      "reward": -0.3125,
      "reward_std": 0.19716878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 73
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.46540880503144655,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.019128454849123955,
      "kl": 0.00011911594629054889,
      "learning_rate": 1.0305368692688175e-06,
      "loss": 0.0,
      "num_tokens": 5082368.0,
      "reward": -0.28125,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 74
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.4716981132075472,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.019733579829335213,
      "kl": 0.0001243672304553911,
      "learning_rate": 9.608463116858544e-07,
      "loss": 0.0,
      "num_tokens": 5150384.0,
      "reward": -0.15625,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 75
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.4779874213836478,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.012829476967453957,
      "kl": 0.00010366842616349459,
      "learning_rate": 8.930309757836517e-07,
      "loss": 0.0,
      "num_tokens": 5218736.0,
      "reward": -0.125,
      "reward_std": 0.125,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.125,
      "rewards/length_penalty/std": 0.22360680997371674,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 76
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.48427672955974843,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.019120682030916214,
      "kl": 0.00011917242591152899,
      "learning_rate": 8.271734841028553e-07,
      "loss": 0.0,
      "num_tokens": 5287084.0,
      "reward": -0.25,
      "reward_std": 0.26933756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 77
    },
    {
      "completion_length": 4004.8125,
      "completions/clipped_ratio": 0.9375,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 2637.0,
      "completions/mean_length": 4004.8125,
      "completions/mean_terminated_length": 2637.0,
      "completions/min_length": 2637.0,
      "completions/min_terminated_length": 2637.0,
      "epoch": 0.49056603773584906,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.014471229165792465,
      "kl": 0.00014815009490121156,
      "learning_rate": 7.633540738525066e-07,
      "loss": 0.0,
      "num_tokens": 5355013.0,
      "reward": -0.125,
      "reward_std": 0.125,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.125,
      "rewards/length_penalty/std": 0.22360680997371674,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 78
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.4968553459119497,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.014715207740664482,
      "kl": 0.00013525429676519707,
      "learning_rate": 7.016504991533727e-07,
      "loss": 0.0,
      "num_tokens": 5423537.0,
      "reward": -0.34375,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.34375,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 79
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5031446540880503,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.018443701788783073,
      "kl": 0.00017349646441289224,
      "learning_rate": 6.421379363065142e-07,
      "loss": 0.0,
      "num_tokens": 5493113.0,
      "reward": -0.34375,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.34375,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 80
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5094339622641509,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.022510401904582977,
      "kl": 0.00013557132479036227,
      "learning_rate": 5.848888922025553e-07,
      "loss": 0.0,
      "num_tokens": 5561473.0,
      "reward": -0.21875,
      "reward_std": 0.13466878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 81
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5157232704402516,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.023799598217010498,
      "kl": 0.00015203715156530961,
      "learning_rate": 5.299731159831953e-07,
      "loss": 0.0,
      "num_tokens": 5629365.0,
      "reward": -0.25,
      "reward_std": 0.26933756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 82
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5220125786163522,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.33711668848991394,
      "kl": 0.00021842254864168353,
      "learning_rate": 4.774575140626317e-07,
      "loss": 0.0,
      "num_tokens": 5698401.0,
      "reward": -0.21875,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 83
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5283018867924528,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.016040142625570297,
      "kl": 0.00012902289199701045,
      "learning_rate": 4.27406068612396e-07,
      "loss": 0.0,
      "num_tokens": 5766933.0,
      "reward": -0.28125,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.28125,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 84
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5345911949685535,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.013614697381854057,
      "kl": 0.0001550722481624689,
      "learning_rate": 3.798797596089351e-07,
      "loss": 0.0,
      "num_tokens": 5835701.0,
      "reward": -0.3125,
      "reward_std": 0.125,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 85
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5408805031446541,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.018001973628997803,
      "kl": 0.0001305843543377705,
      "learning_rate": 3.3493649053890325e-07,
      "loss": 0.0,
      "num_tokens": 5903725.0,
      "reward": -0.21875,
      "reward_std": 0.27900636196136475,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 86
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5471698113207547,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.018992334604263306,
      "kl": 0.0001532331825728761,
      "learning_rate": 2.9263101785268253e-07,
      "loss": 0.0,
      "num_tokens": 5973033.0,
      "reward": -0.21875,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 87
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5534591194968553,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.03358004614710808,
      "kl": 0.00011775902385124937,
      "learning_rate": 2.53014884252083e-07,
      "loss": 0.0,
      "num_tokens": 6042289.0,
      "reward": -0.15625,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 88
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.559748427672956,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.07937372475862503,
      "kl": 0.00018618855028762482,
      "learning_rate": 2.1613635589349756e-07,
      "loss": 0.0,
      "num_tokens": 6110789.0,
      "reward": -0.25,
      "reward_std": 0.25,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.25,
      "rewards/length_penalty/std": 0.25819888710975647,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 89
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5660377358490566,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.020508408546447754,
      "kl": 0.00020692015095846727,
      "learning_rate": 1.8204036358303173e-07,
      "loss": 0.0,
      "num_tokens": 6180189.0,
      "reward": -0.34375,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.34375,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 90
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5723270440251572,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.016169408336281776,
      "kl": 0.00010522551565372851,
      "learning_rate": 1.507684480352292e-07,
      "loss": 0.0,
      "num_tokens": 6249033.0,
      "reward": -0.15625,
      "reward_std": 0.20683756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 91
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5786163522012578,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01694478653371334,
      "kl": 0.0001696805666142609,
      "learning_rate": 1.223587092621162e-07,
      "loss": 0.0,
      "num_tokens": 6318677.0,
      "reward": -0.15625,
      "reward_std": 0.1875,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 92
    },
    {
      "completion_length": 4018.125,
      "completions/clipped_ratio": 0.9375,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 2850.0,
      "completions/mean_length": 4018.125,
      "completions/mean_terminated_length": 2850.0,
      "completions/min_length": 2850.0,
      "completions/min_terminated_length": 2850.0,
      "epoch": 0.5849056603773585,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.017274906858801842,
      "kl": 0.0001594936038600281,
      "learning_rate": 9.684576015420277e-08,
      "loss": 0.0,
      "num_tokens": 6386959.0,
      "reward": -0.1875,
      "reward_std": 0.21650634706020355,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.1875,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 93
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5911949685534591,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 0.008834270760416985,
      "kl": 0.0001470650859118905,
      "learning_rate": 7.426068431000883e-08,
      "loss": 0.0,
      "num_tokens": 6455375.0,
      "reward": -0.1875,
      "reward_std": 0.07216878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.1875,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 94
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.5974842767295597,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.01272174995392561,
      "kl": 0.00012684663852269296,
      "learning_rate": 5.463099816548578e-08,
      "loss": 0.0,
      "num_tokens": 6523575.0,
      "reward": -0.21875,
      "reward_std": 0.13466878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 95
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.6037735849056604,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 0.01726069673895836,
      "kl": 0.00014613767780247144,
      "learning_rate": 3.798061746947995e-08,
      "loss": 0.0,
      "num_tokens": 6591987.0,
      "reward": -0.3125,
      "reward_std": 0.21650634706020355,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 96
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.610062893081761,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 0.0105120949447155,
      "kl": 0.00014528684914694168,
      "learning_rate": 2.4329828146074096e-08,
      "loss": 0.0,
      "num_tokens": 6660639.0,
      "reward": -0.1875,
      "reward_std": 0.07216878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.1875,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 97
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.6163522012578616,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 0.01425850484520197,
      "kl": 0.00016166625209734775,
      "learning_rate": 1.3695261579316776e-08,
      "loss": 0.0,
      "num_tokens": 6730011.0,
      "reward": -0.21875,
      "reward_std": 0.13466878235340118,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.21875,
      "rewards/length_penalty/std": 0.2561737895011902,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 98
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.6226415094339622,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.021140672266483307,
      "kl": 0.00014877854118822142,
      "learning_rate": 6.089874350439507e-09,
      "loss": 0.0,
      "num_tokens": 6800287.0,
      "reward": -0.3125,
      "reward_std": 0.26933756470680237,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.3125,
      "rewards/length_penalty/std": 0.25,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 99
    },
    {
      "completion_length": 4096.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 4096.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 4096.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.6289308176100629,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.02212378941476345,
      "kl": 0.00014287161320680752,
      "learning_rate": 1.5229324522605949e-09,
      "loss": 0.0,
      "num_tokens": 6868131.0,
      "reward": -0.15625,
      "reward_std": 0.25966876745224,
      "rewards/correctness_reward/mean": 0.0,
      "rewards/correctness_reward/std": 0.0,
      "rewards/length_penalty/mean": -0.15625,
      "rewards/length_penalty/std": 0.23935678601264954,
      "rewards/xml_format_reward/mean": 0.0,
      "rewards/xml_format_reward/std": 0.0,
      "step": 100
    }
  ],
  "logging_steps": 1,
  "max_steps": 100,
  "num_input_tokens_seen": 6868131,
  "num_train_epochs": 1,
  "save_steps": 10,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}