{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.21333333333333335,
  "eval_steps": 500,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "calib/answer_extract_rate": 0.1015625,
      "calib/avg_num_step_conf": 0.42578125,
      "calib/ece": 0.8952941176470588,
      "calib/final_conf_rate": 0.06640625,
      "calib/format_rate": 0.0390625,
      "calib/frac_conf_gt_0.9": 0.9411764705882353,
      "calib/gap": 0.006249999999999978,
      "calib/mean_conf": 0.9541176470588235,
      "calib/mu_c": 0.96,
      "calib/mu_w": 0.95375,
      "calib/nonempty_final_conf_rate": 0.06640625,
      "calib/nonempty_reasoning_rate": 0.14453125,
      "calib/nonempty_step_conf_rate": 0.09765625,
      "calib/pce": 0.8952941176470588,
      "calib/std_conf": 0.044066637177242435,
      "calib/step_conf_rate": 0.09765625,
      "calib/step_q_c": 0.62,
      "calib/step_q_c_n": 4.0,
      "calib/step_q_gap": -0.12542857142857122,
      "calib/step_q_w": 0.7454285714285712,
      "calib/step_q_w_n": 105.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 3054.0,
      "completions/max_terminated_length": 3054.0,
      "completions/mean_length": 622.23828125,
      "completions/mean_terminated_length": 674.9703369140625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0010666666666666667,
      "grad_norm": 0.09340800344944,
      "learning_rate": 0.0,
      "loss": 0.0278,
      "num_tokens": 266877.0,
      "reward": 0.0234375,
      "reward_std": 0.04294103384017944,
      "rewards/accuracy_reward_step": 0.00390625,
      "rewards/format_reward_step": 0.0390625,
      "step": 1
    },
    {
      "calib/answer_extract_rate": 0.1015625,
      "calib/avg_num_step_conf": 0.4375,
      "calib/ece": 0.6367055555555554,
      "calib/final_conf_rate": 0.0703125,
      "calib/format_rate": 0.0546875,
      "calib/frac_conf_gt_0.9": 0.8888888888888888,
      "calib/gap": -0.05005833333333343,
      "calib/mean_conf": 0.8867055555555555,
      "calib/mu_c": 0.8533333333333334,
      "calib/mu_w": 0.9033916666666668,
      "calib/nonempty_final_conf_rate": 0.0703125,
      "calib/nonempty_reasoning_rate": 0.13671875,
      "calib/nonempty_step_conf_rate": 0.09375,
      "calib/pce": 0.5950388888888888,
      "calib/std_conf": 0.27071644959301405,
      "calib/step_conf_rate": 0.09375,
      "calib/step_q_c": 0.6966666666666667,
      "calib/step_q_c_n": 24.0,
      "calib/step_q_gap": -0.09584469696969677,
      "calib/step_q_w": 0.7925113636363634,
      "calib/step_q_w_n": 88.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.140625,
      "completions/max_length": 3049.0,
      "completions/max_terminated_length": 3049.0,
      "completions/mean_length": 608.22265625,
      "completions/mean_terminated_length": 707.75,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 7.0,
      "epoch": 0.0021333333333333334,
      "grad_norm": 0.1929834485054016,
      "learning_rate": 2.5000000000000004e-07,
      "loss": -0.0217,
      "num_tokens": 525870.0,
      "reward": 0.05078125,
      "reward_std": 0.10960911214351654,
      "rewards/accuracy_reward_step": 0.0234375,
      "rewards/format_reward_step": 0.0546875,
      "step": 2
    },
    {
      "calib/answer_extract_rate": 0.05859375,
      "calib/avg_num_step_conf": 0.30078125,
      "calib/ece": 0.851,
      "calib/final_conf_rate": 0.0390625,
      "calib/format_rate": 0.02734375,
      "calib/frac_conf_gt_0.9": 0.9,
      "calib/gap": 0.03222222222222204,
      "calib/mean_conf": 0.9509999999999998,
      "calib/mu_c": 0.98,
      "calib/mu_w": 0.9477777777777779,
      "calib/nonempty_final_conf_rate": 0.0390625,
      "calib/nonempty_reasoning_rate": 0.078125,
      "calib/nonempty_step_conf_rate": 0.05859375,
      "calib/pce": 0.851,
      "calib/std_conf": 0.05356304696336829,
      "calib/step_conf_rate": 0.05859375,
      "calib/step_q_c": 0.8320000000000001,
      "calib/step_q_c_n": 10.0,
      "calib/step_q_gap": 5.970149253731183e-05,
      "calib/step_q_w": 0.8319402985074628,
      "calib/step_q_w_n": 67.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1171875,
      "completions/max_length": 2989.0,
      "completions/max_terminated_length": 2989.0,
      "completions/mean_length": 603.53515625,
      "completions/mean_terminated_length": 683.6504516601562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0032,
      "grad_norm": 0.10258053988218307,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.0015,
      "num_tokens": 785631.0,
      "reward": 0.021484375,
      "reward_std": 0.03884732350707054,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/format_reward_step": 0.02734375,
      "step": 3
    },
    {
      "calib/answer_extract_rate": 0.04296875,
      "calib/avg_num_step_conf": 0.16015625,
      "calib/ece": 0.8185714285714286,
      "calib/final_conf_rate": 0.02734375,
      "calib/format_rate": 0.0234375,
      "calib/frac_conf_gt_0.9": 0.7142857142857143,
      "calib/gap": -0.08833333333333337,
      "calib/mean_conf": 0.9157142857142857,
      "calib/mu_c": 0.84,
      "calib/mu_w": 0.9283333333333333,
      "calib/nonempty_final_conf_rate": 0.02734375,
      "calib/nonempty_reasoning_rate": 0.05859375,
      "calib/nonempty_step_conf_rate": 0.0390625,
      "calib/pce": 0.7957142857142857,
      "calib/std_conf": 0.08086269541762801,
      "calib/step_conf_rate": 0.0390625,
      "calib/step_q_c": 0.8725,
      "calib/step_q_c_n": 4.0,
      "calib/step_q_gap": 0.13087837837837846,
      "calib/step_q_w": 0.7416216216216216,
      "calib/step_q_w_n": 37.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1328125,
      "completions/max_length": 3050.0,
      "completions/max_terminated_length": 3050.0,
      "completions/mean_length": 617.1484375,
      "completions/mean_terminated_length": 711.6666870117188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.004266666666666667,
      "grad_norm": 0.07202501595020294,
      "learning_rate": 7.5e-07,
      "loss": 0.0102,
      "num_tokens": 1049789.0,
      "reward": 0.015625,
      "reward_std": 0.04037860035896301,
      "rewards/accuracy_reward_step": 0.00390625,
      "rewards/format_reward_step": 0.0234375,
      "step": 4
    },
    {
      "calib/answer_extract_rate": 0.0546875,
      "calib/avg_num_step_conf": 0.11328125,
      "calib/ece": 0.712857142857143,
      "calib/final_conf_rate": 0.02734375,
      "calib/format_rate": 0.0234375,
      "calib/frac_conf_gt_0.9": 0.8571428571428571,
      "calib/gap": -0.07100000000000006,
      "calib/mean_conf": 0.9357142857142858,
      "calib/mu_c": 0.885,
      "calib/mu_w": 0.9560000000000001,
      "calib/nonempty_final_conf_rate": 0.02734375,
      "calib/nonempty_reasoning_rate": 0.05859375,
      "calib/nonempty_step_conf_rate": 0.02734375,
      "calib/pce": 0.6814285714285715,
      "calib/std_conf": 0.06477590885002646,
      "calib/step_conf_rate": 0.02734375,
      "calib/step_q_c": 0.95625,
      "calib/step_q_c_n": 8.0,
      "calib/step_q_gap": 0.24529761904761904,
      "calib/step_q_w": 0.710952380952381,
      "calib/step_q_w_n": 21.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.125,
      "completions/max_length": 2706.0,
      "completions/max_terminated_length": 2706.0,
      "completions/mean_length": 626.7421875,
      "completions/mean_terminated_length": 716.2767944335938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.005333333333333333,
      "grad_norm": 0.10153383016586304,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.009,
      "num_tokens": 1316923.0,
      "reward": 0.01953125,
      "reward_std": 0.04761157184839249,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/format_reward_step": 0.0234375,
      "step": 5
    },
    {
      "calib/answer_extract_rate": 0.1015625,
      "calib/avg_num_step_conf": 0.49609375,
      "calib/ece": 0.73524,
      "calib/final_conf_rate": 0.08203125,
      "calib/format_rate": 0.06640625,
      "calib/frac_conf_gt_0.9": 0.7619047619047619,
      "calib/gap": 0.11499777777777764,
      "calib/mean_conf": 0.878097142857143,
      "calib/mu_c": 0.9766666666666666,
      "calib/mu_w": 0.8616688888888889,
      "calib/nonempty_final_conf_rate": 0.08203125,
      "calib/nonempty_reasoning_rate": 0.12109375,
      "calib/nonempty_step_conf_rate": 0.09375,
      "calib/pce": 0.73524,
      "calib/std_conf": 0.24146915049226425,
      "calib/step_conf_rate": 0.09375,
      "calib/step_q_c": 0.7169999999999999,
      "calib/step_q_c_n": 10.0,
      "calib/step_q_gap": 0.008461025641025466,
      "calib/step_q_w": 0.7085389743589744,
      "calib/step_q_w_n": 117.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0859375,
      "completions/max_length": 3018.0,
      "completions/max_terminated_length": 3018.0,
      "completions/mean_length": 527.4765625,
      "completions/mean_terminated_length": 577.0684204101562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0064,
      "grad_norm": 0.23361903429031372,
      "learning_rate": 1.25e-06,
      "loss": 0.0122,
      "num_tokens": 1557909.0,
      "reward": 0.044921875,
      "reward_std": 0.09989244490861893,
      "rewards/accuracy_reward_step": 0.01171875,
      "rewards/format_reward_step": 0.06640625,
      "step": 6
    },
    {
      "calib/answer_extract_rate": 0.06640625,
      "calib/avg_num_step_conf": 0.25390625,
      "calib/ece": 0.77125,
      "calib/final_conf_rate": 0.03125,
      "calib/format_rate": 0.015625,
      "calib/frac_conf_gt_0.9": 0.625,
      "calib/gap": 0.004285714285714226,
      "calib/mean_conf": 0.89625,
      "calib/mu_c": 0.9,
      "calib/mu_w": 0.8957142857142858,
      "calib/nonempty_final_conf_rate": 0.03125,
      "calib/nonempty_reasoning_rate": 0.0859375,
      "calib/nonempty_step_conf_rate": 0.0546875,
      "calib/pce": 0.77125,
      "calib/std_conf": 0.13056966531319592,
      "calib/step_conf_rate": 0.0546875,
      "calib/step_q_c": 0.7416666666666667,
      "calib/step_q_c_n": 6.0,
      "calib/step_q_gap": -0.08680790960451967,
      "calib/step_q_w": 0.8284745762711864,
      "calib/step_q_w_n": 59.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.13671875,
      "completions/max_length": 2886.0,
      "completions/max_terminated_length": 2886.0,
      "completions/mean_length": 673.59765625,
      "completions/mean_terminated_length": 780.2760620117188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 4.0,
      "epoch": 0.007466666666666667,
      "grad_norm": 0.06771083176136017,
      "learning_rate": 1.5e-06,
      "loss": 0.0274,
      "num_tokens": 1837774.0,
      "reward": 0.01171875,
      "reward_std": 0.0293300561606884,
      "rewards/accuracy_reward_step": 0.00390625,
      "rewards/format_reward_step": 0.015625,
      "step": 7
    },
    {
      "calib/answer_extract_rate": 0.125,
      "calib/avg_num_step_conf": 0.3359375,
      "calib/ece": 0.5172222222222222,
      "calib/final_conf_rate": 0.0703125,
      "calib/format_rate": 0.046875,
      "calib/frac_conf_gt_0.9": 0.7777777777777778,
      "calib/gap": 0.024249999999999994,
      "calib/mean_conf": 0.9027777777777778,
      "calib/mu_c": 0.91625,
      "calib/mu_w": 0.892,
      "calib/nonempty_final_conf_rate": 0.0703125,
      "calib/nonempty_reasoning_rate": 0.13671875,
      "calib/nonempty_step_conf_rate": 0.07421875,
      "calib/pce": 0.4877777777777778,
      "calib/std_conf": 0.14417603266514767,
      "calib/step_conf_rate": 0.07421875,
      "calib/step_q_c": 0.5926666666666667,
      "calib/step_q_c_n": 15.0,
      "calib/step_q_gap": -0.22578403755868548,
      "calib/step_q_w": 0.8184507042253522,
      "calib/step_q_w_n": 71.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.06640625,
      "completions/max_length": 2851.0,
      "completions/max_terminated_length": 2851.0,
      "completions/mean_length": 626.35546875,
      "completions/mean_terminated_length": 670.907958984375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.008533333333333334,
      "grad_norm": 0.1453000009059906,
      "learning_rate": 1.75e-06,
      "loss": 0.0087,
      "num_tokens": 2104633.0,
      "reward": 0.0546875,
      "reward_std": 0.11970394104719162,
      "rewards/accuracy_reward_step": 0.03125,
      "rewards/format_reward_step": 0.046875,
      "step": 8
    },
    {
      "calib/answer_extract_rate": 0.09375,
      "calib/avg_num_step_conf": 0.29296875,
      "calib/ece": 0.9231578947368422,
      "calib/final_conf_rate": 0.07421875,
      "calib/format_rate": 0.0546875,
      "calib/frac_conf_gt_0.9": 0.8947368421052632,
      "calib/mean_conf": 0.9231578947368421,
      "calib/mu_c": NaN,
      "calib/mu_w": 0.9231578947368421,
      "calib/nonempty_final_conf_rate": 0.07421875,
      "calib/nonempty_reasoning_rate": 0.1015625,
      "calib/nonempty_step_conf_rate": 0.06640625,
      "calib/pce": 0.9231578947368422,
      "calib/std_conf": 0.1707892709430311,
      "calib/step_conf_rate": 0.06640625,
      "calib/step_q_w": 0.7977333333333333,
      "calib/step_q_w_n": 75.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.10546875,
      "completions/max_length": 3008.0,
      "completions/max_terminated_length": 3008.0,
      "completions/mean_length": 634.09375,
      "completions/mean_terminated_length": 708.8558959960938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0096,
      "grad_norm": 0.1522383689880371,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0123,
      "num_tokens": 2374497.0,
      "reward": 0.03125,
      "reward_std": 0.054752420634031296,
      "rewards/accuracy_reward_step": 0.00390625,
      "rewards/format_reward_step": 0.0546875,
      "step": 9
    },
    {
      "calib/answer_extract_rate": 0.0859375,
      "calib/avg_num_step_conf": 0.3125,
      "calib/ece": 0.6352941176470588,
      "calib/final_conf_rate": 0.06640625,
      "calib/format_rate": 0.0625,
      "calib/frac_conf_gt_0.9": 0.5294117647058824,
      "calib/gap": 0.034047619047619104,
      "calib/mean_conf": 0.7352941176470589,
      "calib/mu_c": 0.7633333333333333,
      "calib/mu_w": 0.7292857142857142,
      "calib/nonempty_final_conf_rate": 0.06640625,
      "calib/nonempty_reasoning_rate": 0.10546875,
      "calib/nonempty_step_conf_rate": 0.0859375,
      "calib/pce": 0.5970588235294118,
      "calib/std_conf": 0.32093828703201643,
      "calib/step_conf_rate": 0.0859375,
      "calib/step_q_c": 0.7142857142857144,
      "calib/step_q_c_n": 7.0,
      "calib/step_q_gap": 0.00017612524461851908,
      "calib/step_q_w": 0.7141095890410959,
      "calib/step_q_w_n": 73.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.08203125,
      "completions/max_length": 2912.0,
      "completions/max_terminated_length": 2912.0,
      "completions/mean_length": 633.73828125,
      "completions/mean_terminated_length": 690.3701782226562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.010666666666666666,
      "grad_norm": 0.2347153276205063,
      "learning_rate": 2.25e-06,
      "loss": 0.0364,
      "num_tokens": 2643534.0,
      "reward": 0.04296875,
      "reward_std": 0.10383903235197067,
      "rewards/accuracy_reward_step": 0.01171875,
      "rewards/format_reward_step": 0.0625,
      "step": 10
    },
    {
      "calib/answer_extract_rate": 0.18359375,
      "calib/avg_num_step_conf": 0.73046875,
      "calib/ece": 0.6881818181818182,
      "calib/final_conf_rate": 0.12890625,
      "calib/format_rate": 0.1015625,
      "calib/frac_conf_gt_0.9": 0.8181818181818182,
      "calib/gap": -0.043194444444444424,
      "calib/mean_conf": 0.9003030303030304,
      "calib/mu_c": 0.8688888888888889,
      "calib/mu_w": 0.9120833333333334,
      "calib/nonempty_final_conf_rate": 0.12890625,
      "calib/nonempty_reasoning_rate": 0.2265625,
      "calib/nonempty_step_conf_rate": 0.15234375,
      "calib/pce": 0.657878787878788,
      "calib/std_conf": 0.20900173768285668,
      "calib/step_conf_rate": 0.15234375,
      "calib/step_q_c": 0.7881264705882352,
      "calib/step_q_c_n": 34.0,
      "calib/step_q_gap": 0.04878006535947699,
      "calib/step_q_w": 0.7393464052287582,
      "calib/step_q_w_n": 153.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.06640625,
      "completions/max_length": 2999.0,
      "completions/max_terminated_length": 2999.0,
      "completions/mean_length": 675.6875,
      "completions/mean_terminated_length": 723.7489624023438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.011733333333333333,
      "grad_norm": 0.2045155018568039,
      "learning_rate": 2.5e-06,
      "loss": 0.0122,
      "num_tokens": 2920990.0,
      "reward": 0.0859375,
      "reward_std": 0.1808602213859558,
      "rewards/accuracy_reward_step": 0.03515625,
      "rewards/format_reward_step": 0.1015625,
      "step": 11
    },
    {
      "calib/answer_extract_rate": 0.20703125,
      "calib/avg_num_step_conf": 0.953125,
      "calib/ece": 0.5427272727272727,
      "calib/final_conf_rate": 0.171875,
      "calib/format_rate": 0.140625,
      "calib/frac_conf_gt_0.9": 0.6363636363636364,
      "calib/gap": -0.0040000000000000036,
      "calib/mean_conf": 0.8386363636363637,
      "calib/mu_c": 0.836,
      "calib/mu_w": 0.84,
      "calib/nonempty_final_conf_rate": 0.171875,
      "calib/nonempty_reasoning_rate": 0.2421875,
      "calib/nonempty_step_conf_rate": 0.1953125,
      "calib/pce": 0.5202272727272728,
      "calib/std_conf": 0.25541046206495027,
      "calib/step_conf_rate": 0.1953125,
      "calib/step_q_c": 0.8345333333333333,
      "calib/step_q_c_n": 75.0,
      "calib/step_q_gap": 0.07672268244575942,
      "calib/step_q_w": 0.7578106508875739,
      "calib/step_q_w_n": 169.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.09765625,
      "completions/max_length": 3049.0,
      "completions/max_terminated_length": 3049.0,
      "completions/mean_length": 581.84375,
      "completions/mean_terminated_length": 644.8138427734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 4.0,
      "epoch": 0.0128,
      "grad_norm": 0.2739364504814148,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.0439,
      "num_tokens": 3174118.0,
      "reward": 0.12890625,
      "reward_std": 0.2127167135477066,
      "rewards/accuracy_reward_step": 0.05859375,
      "rewards/format_reward_step": 0.140625,
      "step": 12
    },
    {
      "calib/answer_extract_rate": 0.23828125,
      "calib/avg_num_step_conf": 1.03125,
      "calib/ece": 0.6739102564102565,
      "calib/final_conf_rate": 0.203125,
      "calib/format_rate": 0.16015625,
      "calib/frac_conf_gt_0.9": 0.7884615384615384,
      "calib/gap": 0.09339985218033986,
      "calib/mean_conf": 0.8854487179487178,
      "calib/mu_c": 0.959090909090909,
      "calib/mu_w": 0.8656910569105691,
      "calib/nonempty_final_conf_rate": 0.203125,
      "calib/nonempty_reasoning_rate": 0.2890625,
      "calib/nonempty_step_conf_rate": 0.21484375,
      "calib/pce": 0.6739102564102565,
      "calib/std_conf": 0.19547775011226887,
      "calib/step_conf_rate": 0.21484375,
      "calib/step_q_c": 0.7627272727272728,
      "calib/step_q_c_n": 44.0,
      "calib/step_q_gap": -0.017076666666666518,
      "calib/step_q_w": 0.7798039393939393,
      "calib/step_q_w_n": 220.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 2994.0,
      "completions/max_terminated_length": 2994.0,
      "completions/mean_length": 591.5390625,
      "completions/mean_terminated_length": 615.5853271484375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.013866666666666666,
      "grad_norm": 0.2713400721549988,
      "learning_rate": 3e-06,
      "loss": 0.0977,
      "num_tokens": 3430144.0,
      "reward": 0.123046875,
      "reward_std": 0.24303366243839264,
      "rewards/accuracy_reward_step": 0.04296875,
      "rewards/format_reward_step": 0.16015625,
      "step": 13
    },
    {
      "calib/answer_extract_rate": 0.3125,
      "calib/avg_num_step_conf": 1.22265625,
      "calib/ece": 0.6165846153846155,
      "calib/final_conf_rate": 0.25390625,
      "calib/format_rate": 0.20703125,
      "calib/frac_conf_gt_0.9": 0.7846153846153846,
      "calib/gap": 0.057099290780141754,
      "calib/mean_conf": 0.8830461538461538,
      "calib/mu_c": 0.9243333333333332,
      "calib/mu_w": 0.8672340425531915,
      "calib/nonempty_final_conf_rate": 0.25390625,
      "calib/nonempty_reasoning_rate": 0.37890625,
      "calib/nonempty_step_conf_rate": 0.28515625,
      "calib/pce": 0.6113538461538462,
      "calib/std_conf": 0.22442819175504194,
      "calib/step_conf_rate": 0.28515625,
      "calib/step_q_c": 0.7554666666666666,
      "calib/step_q_c_n": 75.0,
      "calib/step_q_gap": 0.022105322128851412,
      "calib/step_q_w": 0.7333613445378152,
      "calib/step_q_w_n": 238.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 3053.0,
      "completions/max_terminated_length": 3053.0,
      "completions/mean_length": 521.171875,
      "completions/mean_terminated_length": 565.3389892578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 4.0,
      "epoch": 0.014933333333333333,
      "grad_norm": 0.32171332836151123,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.0173,
      "num_tokens": 3668964.0,
      "reward": 0.177734375,
      "reward_std": 0.24229903519153595,
      "rewards/accuracy_reward_step": 0.07421875,
      "rewards/format_reward_step": 0.20703125,
      "step": 14
    },
    {
      "calib/answer_extract_rate": 0.51953125,
      "calib/avg_num_step_conf": 2.54296875,
      "calib/ece": 0.6969990909090908,
      "calib/final_conf_rate": 0.4296875,
      "calib/format_rate": 0.35546875,
      "calib/frac_conf_gt_0.9": 0.7181818181818181,
      "calib/gap": 0.0034599999999999076,
      "calib/mean_conf": 0.9077263636363636,
      "calib/mu_c": 0.9103999999999999,
      "calib/mu_w": 0.90694,
      "calib/nonempty_final_conf_rate": 0.4296875,
      "calib/nonempty_reasoning_rate": 0.59375,
      "calib/nonempty_step_conf_rate": 0.48046875,
      "calib/pce": 0.6887263636363635,
      "calib/std_conf": 0.14853568207812298,
      "calib/step_conf_rate": 0.48046875,
      "calib/step_q_c": 0.718747663551402,
      "calib/step_q_c_n": 107.0,
      "calib/step_q_gap": -0.039332912428990285,
      "calib/step_q_w": 0.7580805759803922,
      "calib/step_q_w_n": 544.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 3031.0,
      "completions/max_terminated_length": 3031.0,
      "completions/mean_length": 514.18359375,
      "completions/mean_terminated_length": 524.4263305664062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.016,
      "grad_norm": 0.39926886558532715,
      "learning_rate": 3.5e-06,
      "loss": 0.1415,
      "num_tokens": 3908475.0,
      "reward": 0.279296875,
      "reward_std": 0.3604772984981537,
      "rewards/accuracy_reward_step": 0.1015625,
      "rewards/format_reward_step": 0.35546875,
      "step": 15
    },
    {
      "calib/answer_extract_rate": 0.57421875,
      "calib/avg_num_step_conf": 3.03125,
      "calib/ece": 0.7074580152671756,
      "calib/final_conf_rate": 0.51171875,
      "calib/format_rate": 0.43359375,
      "calib/frac_conf_gt_0.9": 0.7786259541984732,
      "calib/gap": -0.07769666048237456,
      "calib/mean_conf": 0.9153969465648855,
      "calib/mu_c": 0.8572727272727273,
      "calib/mu_w": 0.9349693877551019,
      "calib/nonempty_final_conf_rate": 0.51171875,
      "calib/nonempty_reasoning_rate": 0.65625,
      "calib/nonempty_step_conf_rate": 0.55859375,
      "calib/pce": 0.6854732824427481,
      "calib/std_conf": 0.16302843284022775,
      "calib/step_conf_rate": 0.55859375,
      "calib/step_q_c": 0.756219512195122,
      "calib/step_q_c_n": 164.0,
      "calib/step_q_gap": -0.027582884319039325,
      "calib/step_q_w": 0.7838023965141613,
      "calib/step_q_w_n": 612.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 2917.0,
      "completions/max_terminated_length": 2917.0,
      "completions/mean_length": 515.140625,
      "completions/mean_terminated_length": 531.758056640625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.017066666666666667,
      "grad_norm": 0.3101026713848114,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.0761,
      "num_tokens": 4149199.0,
      "reward": 0.353515625,
      "reward_std": 0.3571109175682068,
      "rewards/accuracy_reward_step": 0.13671875,
      "rewards/format_reward_step": 0.43359375,
      "step": 16
    },
    {
      "calib/answer_extract_rate": 0.703125,
      "calib/avg_num_step_conf": 3.8828125,
      "calib/ece": 0.6893592814371259,
      "calib/final_conf_rate": 0.65234375,
      "calib/format_rate": 0.58984375,
      "calib/frac_conf_gt_0.9": 0.7964071856287425,
      "calib/gap": -0.00263681102362201,
      "calib/mean_conf": 0.928880239520958,
      "calib/mu_c": 0.9268750000000001,
      "calib/mu_w": 0.9295118110236221,
      "calib/nonempty_final_conf_rate": 0.65234375,
      "calib/nonempty_reasoning_rate": 0.7734375,
      "calib/nonempty_step_conf_rate": 0.70703125,
      "calib/pce": 0.6893592814371259,
      "calib/std_conf": 0.10589311085660263,
      "calib/step_conf_rate": 0.70703125,
      "calib/step_q_c": 0.7367592592592592,
      "calib/step_q_c_n": 216.0,
      "calib/step_q_gap": -0.0004886841854707669,
      "calib/step_q_w": 0.73724794344473,
      "calib/step_q_w_n": 778.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 3041.0,
      "completions/max_terminated_length": 3041.0,
      "completions/mean_length": 452.05859375,
      "completions/mean_terminated_length": 464.7670593261719,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 34.0,
      "epoch": 0.018133333333333335,
      "grad_norm": 0.3115542232990265,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.1478,
      "num_tokens": 4368454.0,
      "reward": 0.455078125,
      "reward_std": 0.4204365015029907,
      "rewards/accuracy_reward_step": 0.16015625,
      "rewards/format_reward_step": 0.58984375,
      "step": 17
    },
    {
      "calib/answer_extract_rate": 0.73828125,
      "calib/avg_num_step_conf": 4.1484375,
      "calib/ece": 0.637949438202247,
      "calib/final_conf_rate": 0.6953125,
      "calib/format_rate": 0.61328125,
      "calib/frac_conf_gt_0.9": 0.7078651685393258,
      "calib/gap": 0.030807852965747617,
      "calib/mean_conf": 0.8907584269662921,
      "calib/mu_c": 0.9137777777777777,
      "calib/mu_w": 0.8829699248120301,
      "calib/nonempty_final_conf_rate": 0.6953125,
      "calib/nonempty_reasoning_rate": 0.83203125,
      "calib/nonempty_step_conf_rate": 0.7421875,
      "calib/pce": 0.637949438202247,
      "calib/std_conf": 0.19176036386095893,
      "calib/step_conf_rate": 0.7421875,
      "calib/step_q_c": 0.7423841807909604,
      "calib/step_q_c_n": 177.0,
      "calib/step_q_gap": -0.010902937853107364,
      "calib/step_q_w": 0.7532871186440677,
      "calib/step_q_w_n": 885.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2624.0,
      "completions/max_terminated_length": 2624.0,
      "completions/mean_length": 407.09375,
      "completions/mean_terminated_length": 410.2992248535156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 11.0,
      "epoch": 0.0192,
      "grad_norm": 0.3702068030834198,
      "learning_rate": 4.25e-06,
      "loss": 0.1734,
      "num_tokens": 4583390.0,
      "reward": 0.482421875,
      "reward_std": 0.42351776361465454,
      "rewards/accuracy_reward_step": 0.17578125,
      "rewards/format_reward_step": 0.61328125,
      "step": 18
    },
    {
      "calib/answer_extract_rate": 0.91796875,
      "calib/avg_num_step_conf": 4.828125,
      "calib/ece": 0.6976190476190478,
      "calib/final_conf_rate": 0.90234375,
      "calib/format_rate": 0.8359375,
      "calib/frac_conf_gt_0.9": 0.7532467532467533,
      "calib/gap": 0.02320110497237582,
      "calib/mean_conf": 0.9092207792207792,
      "calib/mu_c": 0.9274,
      "calib/mu_w": 0.9041988950276242,
      "calib/nonempty_final_conf_rate": 0.90234375,
      "calib/nonempty_reasoning_rate": 0.98046875,
      "calib/nonempty_step_conf_rate": 0.92578125,
      "calib/pce": 0.6951948051948054,
      "calib/std_conf": 0.1566126793817019,
      "calib/step_conf_rate": 0.92578125,
      "calib/step_q_c": 0.8100411985018727,
      "calib/step_q_c_n": 267.0,
      "calib/step_q_gap": 0.08050163238617014,
      "calib/step_q_w": 0.7295395661157026,
      "calib/step_q_w_n": 968.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1001.0,
      "completions/max_terminated_length": 1001.0,
      "completions/mean_length": 291.68359375,
      "completions/mean_terminated_length": 291.68359375,
      "completions/min_length": 34.0,
      "completions/min_terminated_length": 34.0,
      "epoch": 0.020266666666666665,
      "grad_norm": 0.3535246253013611,
      "learning_rate": 4.5e-06,
      "loss": 0.0078,
      "num_tokens": 4762821.0,
      "reward": 0.6171875,
      "reward_std": 0.372276246547699,
      "rewards/accuracy_reward_step": 0.19921875,
      "rewards/format_reward_step": 0.8359375,
      "step": 19
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 5.06640625,
      "calib/ece": 0.5800403225806453,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 0.7903225806451613,
      "calib/gap": 0.023405923344947843,
      "calib/mean_conf": 0.9123790322580645,
      "calib/mu_c": 0.927857142857143,
      "calib/mu_w": 0.9044512195121952,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.97265625,
      "calib/pce": 0.5768548387096776,
      "calib/std_conf": 0.16562803705027718,
      "calib/step_conf_rate": 0.97265625,
      "calib/step_q_c": 0.760523560209424,
      "calib/step_q_c_n": 382.0,
      "calib/step_q_gap": 0.008751975509970533,
      "calib/step_q_w": 0.7517715846994535,
      "calib/step_q_w_n": 915.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2428.0,
      "completions/max_terminated_length": 2428.0,
      "completions/mean_length": 289.0859375,
      "completions/mean_terminated_length": 289.0859375,
      "completions/min_length": 92.0,
      "completions/min_terminated_length": 92.0,
      "epoch": 0.021333333333333333,
      "grad_norm": 0.27600613236427307,
      "learning_rate": 4.75e-06,
      "loss": 0.0353,
      "num_tokens": 4941699.0,
      "reward": 0.80078125,
      "reward_std": 0.3837167024612427,
      "rewards/accuracy_reward_step": 0.33203125,
      "rewards/format_reward_step": 0.9375,
      "step": 20
    },
    {
      "calib/answer_extract_rate": 0.9609375,
      "calib/avg_num_step_conf": 4.96875,
      "calib/ece": 0.6792874829931973,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.921875,
      "calib/frac_conf_gt_0.9": 0.8367346938775511,
      "calib/gap": 0.017045490334332225,
      "calib/mean_conf": 0.9323487074829931,
      "calib/mu_c": 0.9450806451612903,
      "calib/mu_w": 0.9280351548269581,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.6792874829931973,
      "calib/std_conf": 0.1318469984727218,
      "calib/step_conf_rate": 0.96875,
      "calib/step_q_c": 0.8108069620253165,
      "calib/step_q_c_n": 316.0,
      "calib/step_q_gap": 0.015945107074131104,
      "calib/step_q_w": 0.7948618549511854,
      "calib/step_q_w_n": 956.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2980.0,
      "completions/max_terminated_length": 2980.0,
      "completions/mean_length": 292.47265625,
      "completions/mean_terminated_length": 293.61962890625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 56.0,
      "epoch": 0.0224,
      "grad_norm": 0.262796014547348,
      "learning_rate": 5e-06,
      "loss": 0.0414,
      "num_tokens": 5119532.0,
      "reward": 0.703125,
      "reward_std": 0.37573766708374023,
      "rewards/accuracy_reward_step": 0.2421875,
      "rewards/format_reward_step": 0.921875,
      "step": 21
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 4.97265625,
      "calib/ece": 0.6300711462450593,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.7865612648221344,
      "calib/gap": -0.0007543589743589374,
      "calib/mean_conf": 0.9320474308300395,
      "calib/mu_c": 0.9315256410256411,
      "calib/mu_w": 0.93228,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.626909090909091,
      "calib/std_conf": 0.10416734466703273,
      "calib/step_conf_rate": 0.98046875,
      "calib/step_q_c": 0.8036150127226462,
      "calib/step_q_c_n": 393.0,
      "calib/step_q_gap": 0.004382058177191617,
      "calib/step_q_w": 0.7992329545454546,
      "calib/step_q_w_n": 880.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 753.0,
      "completions/max_terminated_length": 753.0,
      "completions/mean_length": 262.359375,
      "completions/mean_terminated_length": 263.38824462890625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 33.0,
      "epoch": 0.023466666666666667,
      "grad_norm": 0.23157238960266113,
      "learning_rate": 4.9722222222222224e-06,
      "loss": -0.0086,
      "num_tokens": 5288512.0,
      "reward": 0.783203125,
      "reward_std": 0.3809746205806732,
      "rewards/accuracy_reward_step": 0.3046875,
      "rewards/format_reward_step": 0.95703125,
      "step": 22
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 4.80859375,
      "calib/ece": 0.6420756097560976,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.8414634146341463,
      "calib/gap": 0.016442911877394772,
      "calib/mean_conf": 0.9347585365853659,
      "calib/mu_c": 0.946388888888889,
      "calib/mu_w": 0.9299459770114942,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.97265625,
      "calib/pce": 0.6420756097560976,
      "calib/std_conf": 0.12590850202592063,
      "calib/step_conf_rate": 0.97265625,
      "calib/step_q_c": 0.8159972899728999,
      "calib/step_q_c_n": 369.0,
      "calib/step_q_gap": 0.012226988348770007,
      "calib/step_q_w": 0.8037703016241299,
      "calib/step_q_w_n": 862.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1923.0,
      "completions/max_terminated_length": 1923.0,
      "completions/mean_length": 278.5,
      "completions/mean_terminated_length": 279.5921630859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 32.0,
      "epoch": 0.024533333333333334,
      "grad_norm": 0.22491341829299927,
      "learning_rate": 4.944444444444445e-06,
      "loss": 0.0016,
      "num_tokens": 5463744.0,
      "reward": 0.755859375,
      "reward_std": 0.3748607039451599,
      "rewards/accuracy_reward_step": 0.28515625,
      "rewards/format_reward_step": 0.94140625,
      "step": 23
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 5.41796875,
      "calib/ece": 0.7062290836653384,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.8326693227091634,
      "calib/gap": -0.008835951134380515,
      "calib/mean_conf": 0.9408904382470119,
      "calib/mu_c": 0.9341666666666666,
      "calib/mu_w": 0.9430026178010471,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.7040378486055775,
      "calib/std_conf": 0.11581519894633541,
      "calib/step_conf_rate": 0.9921875,
      "calib/step_q_c": 0.782109090909091,
      "calib/step_q_c_n": 275.0,
      "calib/step_q_gap": -0.019002779594506136,
      "calib/step_q_w": 0.8011118705035971,
      "calib/step_q_w_n": 1112.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2370.0,
      "completions/max_terminated_length": 2370.0,
      "completions/mean_length": 296.42578125,
      "completions/mean_terminated_length": 296.42578125,
      "completions/min_length": 40.0,
      "completions/min_terminated_length": 40.0,
      "epoch": 0.0256,
      "grad_norm": 0.17748171091079712,
      "learning_rate": 4.9166666666666665e-06,
      "loss": 0.0558,
      "num_tokens": 5644141.0,
      "reward": 0.71875,
      "reward_std": 0.30310535430908203,
      "rewards/accuracy_reward_step": 0.234375,
      "rewards/format_reward_step": 0.96875,
      "step": 24
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 5.37109375,
      "calib/ece": 0.66048031496063,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.8503937007874016,
      "calib/gap": -0.004199099099099213,
      "calib/mean_conf": 0.9465433070866143,
      "calib/mu_c": 0.9435675675675675,
      "calib/mu_w": 0.9477666666666668,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.6578425196850394,
      "calib/std_conf": 0.07255448736161858,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.8115143222506394,
      "calib/step_q_c_n": 391.0,
      "calib/step_q_gap": -0.008937913521718466,
      "calib/step_q_w": 0.8204522357723578,
      "calib/step_q_w_n": 984.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 772.0,
      "completions/max_terminated_length": 772.0,
      "completions/mean_length": 279.05859375,
      "completions/mean_terminated_length": 280.1529541015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 105.0,
      "epoch": 0.02666666666666667,
      "grad_norm": 0.2226199209690094,
      "learning_rate": 4.888888888888889e-06,
      "loss": 0.0067,
      "num_tokens": 5818804.0,
      "reward": 0.783203125,
      "reward_std": 0.3985111117362976,
      "rewards/accuracy_reward_step": 0.296875,
      "rewards/format_reward_step": 0.97265625,
      "step": 25
    },
    {
      "calib/answer_extract_rate": 0.9921875,
      "calib/avg_num_step_conf": 5.4765625,
      "calib/ece": 0.6369146825396825,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.9206349206349206,
      "calib/gap": 0.00606420373027261,
      "calib/mean_conf": 0.9623115079365079,
      "calib/mu_c": 0.9664024390243903,
      "calib/mu_w": 0.9603382352941177,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.6369146825396825,
      "calib/std_conf": 0.03689188390047501,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8371509846827133,
      "calib/step_q_c_n": 457.0,
      "calib/step_q_gap": 0.008553173963136707,
      "calib/step_q_w": 0.8285978107195766,
      "calib/step_q_w_n": 945.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2280.0,
      "completions/max_terminated_length": 2280.0,
      "completions/mean_length": 311.203125,
      "completions/mean_terminated_length": 311.203125,
      "completions/min_length": 99.0,
      "completions/min_terminated_length": 99.0,
      "epoch": 0.027733333333333332,
      "grad_norm": 0.18909384310245514,
      "learning_rate": 4.861111111111111e-06,
      "loss": 0.0547,
      "num_tokens": 6003712.0,
      "reward": 0.8125,
      "reward_std": 0.3554888665676117,
      "rewards/accuracy_reward_step": 0.3203125,
      "rewards/format_reward_step": 0.984375,
      "step": 26
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 5.875,
      "calib/ece": 0.6326521739130435,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.8893280632411067,
      "calib/gap": -0.0024731396172925635,
      "calib/mean_conf": 0.9575533596837943,
      "calib/mu_c": 0.9558915662650602,
      "calib/mu_w": 0.9583647058823528,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.6310711462450593,
      "calib/std_conf": 0.043072717704501245,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.8300023752969121,
      "calib/step_q_c_n": 421.0,
      "calib/step_q_gap": 0.001743649535139169,
      "calib/step_q_w": 0.828258725761773,
      "calib/step_q_w_n": 1083.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2382.0,
      "completions/max_terminated_length": 2382.0,
      "completions/mean_length": 310.98046875,
      "completions/mean_terminated_length": 310.98046875,
      "completions/min_length": 86.0,
      "completions/min_terminated_length": 86.0,
      "epoch": 0.0288,
      "grad_norm": 0.2512206733226776,
      "learning_rate": 4.833333333333333e-06,
      "loss": 0.0554,
      "num_tokens": 6188539.0,
      "reward": 0.81640625,
      "reward_std": 0.3243214786052704,
      "rewards/accuracy_reward_step": 0.32421875,
      "rewards/format_reward_step": 0.984375,
      "step": 27
    },
    {
      "calib/answer_extract_rate": 0.99609375,
      "calib/avg_num_step_conf": 5.72265625,
      "calib/ece": 0.5212460317460319,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.9404761904761905,
      "calib/gap": -0.005501785714285834,
      "calib/mean_conf": 0.9656904761904762,
      "calib/mu_c": 0.9626339285714286,
      "calib/mu_w": 0.9681357142857144,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.5212460317460319,
      "calib/std_conf": 0.03520787797072999,
      "calib/step_conf_rate": 0.9921875,
      "calib/step_q_c": 0.8592661157024793,
      "calib/step_q_c_n": 605.0,
      "calib/step_q_gap": 0.007533557562944493,
      "calib/step_q_w": 0.8517325581395349,
      "calib/step_q_w_n": 860.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2449.0,
      "completions/max_terminated_length": 2449.0,
      "completions/mean_length": 343.22265625,
      "completions/mean_terminated_length": 343.22265625,
      "completions/min_length": 115.0,
      "completions/min_terminated_length": 115.0,
      "epoch": 0.029866666666666666,
      "grad_norm": 0.18749113380908966,
      "learning_rate": 4.805555555555556e-06,
      "loss": 0.0048,
      "num_tokens": 6383348.0,
      "reward": 0.931640625,
      "reward_std": 0.3386801481246948,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/format_reward_step": 0.98046875,
      "step": 28
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 6.50390625,
      "calib/ece": 0.6446484,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.924,
      "calib/gap": 0.009045384558442726,
      "calib/mean_conf": 0.9606484000000001,
      "calib/mu_c": 0.9668354430379748,
      "calib/mu_w": 0.957790058479532,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.6446484,
      "calib/std_conf": 0.06880537491678976,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8503131524008349,
      "calib/step_q_c_n": 479.0,
      "calib/step_q_gap": 0.004799661675708444,
      "calib/step_q_w": 0.8455134907251265,
      "calib/step_q_w_n": 1186.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2479.0,
      "completions/max_terminated_length": 2479.0,
      "completions/mean_length": 402.20703125,
      "completions/mean_terminated_length": 403.7843322753906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 153.0,
      "epoch": 0.030933333333333334,
      "grad_norm": 0.18767628073692322,
      "learning_rate": 4.777777777777778e-06,
      "loss": 0.0569,
      "num_tokens": 6593441.0,
      "reward": 0.796875,
      "reward_std": 0.32658594846725464,
      "rewards/accuracy_reward_step": 0.30859375,
      "rewards/format_reward_step": 0.9765625,
      "step": 29
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 6.53125,
      "calib/ece": 0.6063754940711462,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.9051383399209486,
      "calib/gap": -0.000441581458759166,
      "calib/mean_conf": 0.962106719367589,
      "calib/mu_c": 0.9618222222222224,
      "calib/mu_w": 0.9622638036809815,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.6063754940711462,
      "calib/std_conf": 0.0342245617471847,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8520050966608084,
      "calib/step_q_c_n": 569.0,
      "calib/step_q_gap": 0.007608723134063022,
      "calib/step_q_w": 0.8443963735267453,
      "calib/step_q_w_n": 1103.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2346.0,
      "completions/max_terminated_length": 2346.0,
      "completions/mean_length": 406.7890625,
      "completions/mean_terminated_length": 406.7890625,
      "completions/min_length": 161.0,
      "completions/min_terminated_length": 161.0,
      "epoch": 0.032,
      "grad_norm": 0.18368232250213623,
      "learning_rate": 4.75e-06,
      "loss": 0.0472,
      "num_tokens": 6804563.0,
      "reward": 0.841796875,
      "reward_std": 0.37309718132019043,
      "rewards/accuracy_reward_step": 0.3515625,
      "rewards/format_reward_step": 0.98046875,
      "step": 30
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 7.34765625,
      "calib/ece": 0.5997330677290836,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.9322709163346613,
      "calib/gap": -0.001713043478260956,
      "calib/mean_conf": 0.9582988047808766,
      "calib/mu_c": 0.9571999999999998,
      "calib/mu_w": 0.9589130434782608,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5997330677290836,
      "calib/std_conf": 0.067515457567313,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8475205128205129,
      "calib/step_q_c_n": 585.0,
      "calib/step_q_gap": -0.014571693969610666,
      "calib/step_q_w": 0.8620922067901235,
      "calib/step_q_w_n": 1296.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2306.0,
      "completions/max_terminated_length": 2306.0,
      "completions/mean_length": 436.0078125,
      "completions/mean_terminated_length": 436.0078125,
      "completions/min_length": 136.0,
      "completions/min_terminated_length": 136.0,
      "epoch": 0.03306666666666667,
      "grad_norm": 0.14242278039455414,
      "learning_rate": 4.722222222222222e-06,
      "loss": 0.0464,
      "num_tokens": 7022093.0,
      "reward": 0.845703125,
      "reward_std": 0.2595009207725525,
      "rewards/accuracy_reward_step": 0.35546875,
      "rewards/format_reward_step": 0.98046875,
      "step": 31
    },
    {
      "calib/answer_extract_rate": 0.9921875,
      "calib/avg_num_step_conf": 6.703125,
      "calib/ece": 0.5951372549019607,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.9019607843137255,
      "calib/gap": 0.0041697786076286825,
      "calib/mean_conf": 0.955921568627451,
      "calib/mu_c": 0.9585869565217391,
      "calib/mu_w": 0.9544171779141104,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5951372549019607,
      "calib/std_conf": 0.06854677611081697,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.848257042253521,
      "calib/step_q_c_n": 568.0,
      "calib/step_q_gap": 0.009546240859792943,
      "calib/step_q_w": 0.8387108013937281,
      "calib/step_q_w_n": 1148.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2281.0,
      "completions/max_terminated_length": 2281.0,
      "completions/mean_length": 395.75,
      "completions/mean_terminated_length": 395.75,
      "completions/min_length": 126.0,
      "completions/min_terminated_length": 126.0,
      "epoch": 0.034133333333333335,
      "grad_norm": 0.16663441061973572,
      "learning_rate": 4.694444444444445e-06,
      "loss": -0.0072,
      "num_tokens": 7230109.0,
      "reward": 0.85546875,
      "reward_std": 0.19465388357639313,
      "rewards/accuracy_reward_step": 0.359375,
      "rewards/format_reward_step": 0.9921875,
      "step": 32
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 7.01953125,
      "calib/ece": 0.594716,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.956,
      "calib/gap": 0.0006919389082938876,
      "calib/mean_conf": 0.966716,
      "calib/mu_c": 0.9671505376344084,
      "calib/mu_w": 0.9664585987261145,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.594716,
      "calib/std_conf": 0.03542817161525556,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8711590909090909,
      "calib/step_q_c_n": 660.0,
      "calib/step_q_gap": 0.006741382026065557,
      "calib/step_q_w": 0.8644177088830254,
      "calib/step_q_w_n": 1137.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2602.0,
      "completions/max_terminated_length": 2602.0,
      "completions/mean_length": 441.78515625,
      "completions/mean_terminated_length": 443.5176696777344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 164.0,
      "epoch": 0.0352,
      "grad_norm": 0.16815558075904846,
      "learning_rate": 4.666666666666667e-06,
      "loss": 0.038,
      "num_tokens": 7450078.0,
      "reward": 0.849609375,
      "reward_std": 0.2716839909553528,
      "rewards/accuracy_reward_step": 0.36328125,
      "rewards/format_reward_step": 0.97265625,
      "step": 33
    },
    {
      "calib/answer_extract_rate": 0.99609375,
      "calib/avg_num_step_conf": 6.50390625,
      "calib/ece": 0.5414941176470589,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.9372549019607843,
      "calib/gap": 8.767951625077153e-05,
      "calib/mean_conf": 0.9650235294117647,
      "calib/mu_c": 0.965074074074074,
      "calib/mu_w": 0.9649863945578232,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5414941176470589,
      "calib/std_conf": 0.025525677869167495,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8557174556213019,
      "calib/step_q_c_n": 676.0,
      "calib/step_q_gap": 0.005662855014628487,
      "calib/step_q_w": 0.8500546006066734,
      "calib/step_q_w_n": 989.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2688.0,
      "completions/max_terminated_length": 2688.0,
      "completions/mean_length": 379.390625,
      "completions/mean_terminated_length": 379.390625,
      "completions/min_length": 159.0,
      "completions/min_terminated_length": 159.0,
      "epoch": 0.03626666666666667,
      "grad_norm": 0.2050214260816574,
      "learning_rate": 4.638888888888889e-06,
      "loss": 0.0256,
      "num_tokens": 7652314.0,
      "reward": 0.919921875,
      "reward_std": 0.3066937327384949,
      "rewards/accuracy_reward_step": 0.421875,
      "rewards/format_reward_step": 0.99609375,
      "step": 34
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 7.3359375,
      "calib/ece": 0.5662908366533868,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.9442231075697212,
      "calib/gap": 0.00823278145695372,
      "calib/mean_conf": 0.9646972111553785,
      "calib/mu_c": 0.9696499999999999,
      "calib/mu_w": 0.9614172185430462,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5662908366533868,
      "calib/std_conf": 0.04091254826783785,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8770158959537572,
      "calib/step_q_c_n": 692.0,
      "calib/step_q_gap": 0.016350634570957978,
      "calib/step_q_w": 0.8606652613827992,
      "calib/step_q_w_n": 1186.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2759.0,
      "completions/max_terminated_length": 2759.0,
      "completions/mean_length": 460.00390625,
      "completions/mean_terminated_length": 461.807861328125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 188.0,
      "epoch": 0.037333333333333336,
      "grad_norm": 0.19532760977745056,
      "learning_rate": 4.611111111111112e-06,
      "loss": 0.0054,
      "num_tokens": 7879331.0,
      "reward": 0.884765625,
      "reward_std": 0.39114731550216675,
      "rewards/accuracy_reward_step": 0.39453125,
      "rewards/format_reward_step": 0.98046875,
      "step": 35
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 6.765625,
      "calib/ece": 0.369897233201581,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.9565217391304348,
      "calib/gap": -0.004632450331125804,
      "calib/mean_conf": 0.9667351778656127,
      "calib/mu_c": 0.9648675496688741,
      "calib/mu_w": 0.9694999999999999,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.369897233201581,
      "calib/std_conf": 0.02525124791789609,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8713951120162933,
      "calib/step_q_c_n": 982.0,
      "calib/step_q_gap": -0.010074221317040055,
      "calib/step_q_w": 0.8814693333333333,
      "calib/step_q_w_n": 750.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2772.0,
      "completions/max_terminated_length": 2772.0,
      "completions/mean_length": 447.75390625,
      "completions/mean_terminated_length": 447.75390625,
      "completions/min_length": 163.0,
      "completions/min_terminated_length": 163.0,
      "epoch": 0.0384,
      "grad_norm": 0.18701352179050446,
      "learning_rate": 4.583333333333333e-06,
      "loss": 0.0439,
      "num_tokens": 8096668.0,
      "reward": 1.083984375,
      "reward_std": 0.2993289530277252,
      "rewards/accuracy_reward_step": 0.58984375,
      "rewards/format_reward_step": 0.98828125,
      "step": 36
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 7.5390625,
      "calib/ece": 0.545100806451613,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.9717741935483871,
      "calib/gap": -0.006544455544455374,
      "calib/mean_conf": 0.9684879032258066,
      "calib/mu_c": 0.9647142857142856,
      "calib/mu_w": 0.971258741258741,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.545100806451613,
      "calib/std_conf": 0.022842020886606838,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8704848046309698,
      "calib/step_q_c_n": 691.0,
      "calib/step_q_gap": -0.013976050897682502,
      "calib/step_q_w": 0.8844608555286523,
      "calib/step_q_w_n": 1239.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2587.0,
      "completions/max_terminated_length": 2587.0,
      "completions/mean_length": 471.2265625,
      "completions/mean_terminated_length": 476.8142395019531,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 149.0,
      "epoch": 0.039466666666666664,
      "grad_norm": 0.1921999752521515,
      "learning_rate": 4.555555555555556e-06,
      "loss": -0.0012,
      "num_tokens": 8324398.0,
      "reward": 0.89453125,
      "reward_std": 0.276373028755188,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/format_reward_step": 0.96875,
      "step": 37
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 7.30859375,
      "calib/ece": 0.5370607287449393,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.97165991902834,
      "calib/gap": -0.004610778044231156,
      "calib/mean_conf": 0.9743076923076923,
      "calib/mu_c": 0.9717129629629631,
      "calib/mu_w": 0.9763237410071942,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5370607287449393,
      "calib/std_conf": 0.020714232172895357,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8914092140921409,
      "calib/step_q_c_n": 738.0,
      "calib/step_q_gap": -0.009175075404770028,
      "calib/step_q_w": 0.9005842894969109,
      "calib/step_q_w_n": 1133.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2440.0,
      "completions/max_terminated_length": 2440.0,
      "completions/mean_length": 495.8125,
      "completions/mean_terminated_length": 499.7165222167969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 184.0,
      "epoch": 0.04053333333333333,
      "grad_norm": 0.19463488459587097,
      "learning_rate": 4.527777777777778e-06,
      "loss": 0.0497,
      "num_tokens": 8558214.0,
      "reward": 0.904296875,
      "reward_std": 0.34605079889297485,
      "rewards/accuracy_reward_step": 0.421875,
      "rewards/format_reward_step": 0.96484375,
      "step": 38
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 7.4609375,
      "calib/ece": 0.532296442687747,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.9841897233201581,
      "calib/gap": -0.0008047112462005179,
      "calib/mean_conf": 0.9749841897233201,
      "calib/mu_c": 0.9745357142857144,
      "calib/mu_w": 0.9753404255319149,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.532296442687747,
      "calib/std_conf": 0.020270496557005274,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8955542307692309,
      "calib/step_q_c_n": 780.0,
      "calib/step_q_gap": -0.006108601089176213,
      "calib/step_q_w": 0.9016628318584071,
      "calib/step_q_w_n": 1130.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2264.0,
      "completions/max_terminated_length": 2264.0,
      "completions/mean_length": 521.78125,
      "completions/mean_terminated_length": 521.78125,
      "completions/min_length": 195.0,
      "completions/min_terminated_length": 195.0,
      "epoch": 0.0416,
      "grad_norm": 0.1901213526725769,
      "learning_rate": 4.5e-06,
      "loss": -0.0001,
      "num_tokens": 8797878.0,
      "reward": 0.931640625,
      "reward_std": 0.25382548570632935,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/format_reward_step": 0.98828125,
      "step": 39
    },
    {
      "calib/answer_extract_rate": 0.94921875,
      "calib/avg_num_step_conf": 6.87109375,
      "calib/ece": 0.5636244897959185,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 0.9918367346938776,
      "calib/gap": -0.0031860002742353233,
      "calib/mean_conf": 0.9758693877551021,
      "calib/mu_c": 0.9740098039215687,
      "calib/mu_w": 0.977195804195804,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5615836734693879,
      "calib/std_conf": 0.03475816277102392,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9029054913294798,
      "calib/step_q_c_n": 692.0,
      "calib/step_q_gap": 0.0015829046378210698,
      "calib/step_q_w": 0.9013225866916588,
      "calib/step_q_w_n": 1067.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2483.0,
      "completions/max_terminated_length": 2483.0,
      "completions/mean_length": 533.328125,
      "completions/mean_terminated_length": 533.328125,
      "completions/min_length": 158.0,
      "completions/min_terminated_length": 158.0,
      "epoch": 0.042666666666666665,
      "grad_norm": 0.20792528986930847,
      "learning_rate": 4.472222222222223e-06,
      "loss": 0.0438,
      "num_tokens": 9041170.0,
      "reward": 0.873046875,
      "reward_std": 0.35490143299102783,
      "rewards/accuracy_reward_step": 0.3984375,
      "rewards/format_reward_step": 0.94921875,
      "step": 40
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 6.98828125,
      "calib/ece": 0.27429435483870956,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.9959677419354839,
      "calib/gap": -0.003156104380242053,
      "calib/mean_conf": 0.9759072580645163,
      "calib/mu_c": 0.9749655172413794,
      "calib/mu_w": 0.9781216216216214,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.27429435483870956,
      "calib/std_conf": 0.01858425310443043,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8983594827586207,
      "calib/step_q_c_n": 1160.0,
      "calib/step_q_gap": -0.009166749355846826,
      "calib/step_q_w": 0.9075262321144675,
      "calib/step_q_w_n": 629.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2332.0,
      "completions/max_terminated_length": 2332.0,
      "completions/mean_length": 488.15625,
      "completions/mean_terminated_length": 488.15625,
      "completions/min_length": 200.0,
      "completions/min_terminated_length": 200.0,
      "epoch": 0.04373333333333333,
      "grad_norm": 0.2168932557106018,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.0576,
      "num_tokens": 9273386.0,
      "reward": 1.1640625,
      "reward_std": 0.3478030860424042,
      "rewards/accuracy_reward_step": 0.6796875,
      "rewards/format_reward_step": 0.96875,
      "step": 41
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 7.26171875,
      "calib/ece": 0.484244,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.992,
      "calib/gap": -0.0009284296779974444,
      "calib/mean_conf": 0.976244,
      "calib/mu_c": 0.9757723577235772,
      "calib/mu_w": 0.9767007874015746,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.484244,
      "calib/std_conf": 0.016944865416992853,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8981546894031669,
      "calib/step_q_c_n": 821.0,
      "calib/step_q_gap": -0.00843201579914532,
      "calib/step_q_w": 0.9065867052023122,
      "calib/step_q_w_n": 1038.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2621.0,
      "completions/max_terminated_length": 2621.0,
      "completions/mean_length": 448.59765625,
      "completions/mean_terminated_length": 452.1299133300781,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 190.0,
      "epoch": 0.0448,
      "grad_norm": 0.2360077202320099,
      "learning_rate": 4.416666666666667e-06,
      "loss": 0.0183,
      "num_tokens": 9492595.0,
      "reward": 0.96875,
      "reward_std": 0.3092150092124939,
      "rewards/accuracy_reward_step": 0.48046875,
      "rewards/format_reward_step": 0.9765625,
      "step": 42
    },
    {
      "calib/answer_extract_rate": 0.95703125,
      "calib/avg_num_step_conf": 7.18359375,
      "calib/ece": 0.5074163265306122,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 0.9918367346938776,
      "calib/gap": -0.0011056856187290354,
      "calib/mean_conf": 0.9768040816326531,
      "calib/mu_c": 0.9762173913043478,
      "calib/mu_w": 0.9773230769230768,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.5074163265306122,
      "calib/std_conf": 0.01736364909432598,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.8979627163781623,
      "calib/step_q_c_n": 751.0,
      "calib/step_q_gap": -0.013234893915955448,
      "calib/step_q_w": 0.9111976102941177,
      "calib/step_q_w_n": 1088.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2490.0,
      "completions/max_terminated_length": 2490.0,
      "completions/mean_length": 525.7578125,
      "completions/mean_terminated_length": 534.1032104492188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 176.0,
      "epoch": 0.04586666666666667,
      "grad_norm": 0.2095789611339569,
      "learning_rate": 4.388888888888889e-06,
      "loss": 0.0273,
      "num_tokens": 9732413.0,
      "reward": 0.923828125,
      "reward_std": 0.3529389798641205,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/format_reward_step": 0.94921875,
      "step": 43
    },
    {
      "calib/answer_extract_rate": 0.95703125,
      "calib/avg_num_step_conf": 7.16796875,
      "calib/ece": 0.603582995951417,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.953125,
      "calib/frac_conf_gt_0.9": 0.9959514170040485,
      "calib/gap": -0.00047276916631766586,
      "calib/mean_conf": 0.9801012145748987,
      "calib/mu_c": 0.979806451612903,
      "calib/mu_w": 0.9802792207792207,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.603582995951417,
      "calib/std_conf": 0.015421010647003998,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9029935897435899,
      "calib/step_q_c_n": 624.0,
      "calib/step_q_gap": -0.0022454688856423655,
      "calib/step_q_w": 0.9052390586292323,
      "calib/step_q_w_n": 1211.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2350.0,
      "completions/max_terminated_length": 2350.0,
      "completions/mean_length": 543.63671875,
      "completions/mean_terminated_length": 545.7686767578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 148.0,
      "epoch": 0.046933333333333334,
      "grad_norm": 0.19954414665699005,
      "learning_rate": 4.361111111111112e-06,
      "loss": 0.0314,
      "num_tokens": 9977904.0,
      "reward": 0.83984375,
      "reward_std": 0.2857242524623871,
      "rewards/accuracy_reward_step": 0.36328125,
      "rewards/format_reward_step": 0.953125,
      "step": 44
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 7.359375,
      "calib/ece": 0.5497738095238096,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.9920634920634921,
      "calib/gap": 0.0020046296296295507,
      "calib/mean_conf": 0.9783452380952381,
      "calib/mu_c": 0.9794907407407407,
      "calib/mu_w": 0.9774861111111112,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5497738095238096,
      "calib/std_conf": 0.016030596277331763,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9021262002743485,
      "calib/step_q_c_n": 729.0,
      "calib/step_q_gap": -0.005504102755954587,
      "calib/step_q_w": 0.9076303030303031,
      "calib/step_q_w_n": 1155.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2429.0,
      "completions/max_terminated_length": 2429.0,
      "completions/mean_length": 503.96484375,
      "completions/mean_terminated_length": 505.9411926269531,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 177.0,
      "epoch": 0.048,
      "grad_norm": 0.21172115206718445,
      "learning_rate": 4.333333333333334e-06,
      "loss": 0.0142,
      "num_tokens": 10211967.0,
      "reward": 0.9140625,
      "reward_std": 0.3250906765460968,
      "rewards/accuracy_reward_step": 0.421875,
      "rewards/format_reward_step": 0.984375,
      "step": 45
    },
    {
      "calib/answer_extract_rate": 0.953125,
      "calib/avg_num_step_conf": 7.8984375,
      "calib/ece": 0.507512396694215,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.9453125,
      "calib/frac_conf_gt_0.9": 0.987603305785124,
      "calib/gap": 0.002683405364615532,
      "calib/mean_conf": 0.9744545454545456,
      "calib/mu_c": 0.9758849557522125,
      "calib/mu_w": 0.9732015503875969,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.507512396694215,
      "calib/std_conf": 0.034341623541091014,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9051211734693876,
      "calib/step_q_c_n": 784.0,
      "calib/step_q_gap": -0.009437106013649554,
      "calib/step_q_w": 0.9145582794830371,
      "calib/step_q_w_n": 1238.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 3026.0,
      "completions/max_terminated_length": 3026.0,
      "completions/mean_length": 591.01171875,
      "completions/mean_terminated_length": 593.3294677734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 182.0,
      "epoch": 0.04906666666666667,
      "grad_norm": 0.1701515018939972,
      "learning_rate": 4.305555555555556e-06,
      "loss": 0.0433,
      "num_tokens": 10468034.0,
      "reward": 0.9140625,
      "reward_std": 0.2716418504714966,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/format_reward_step": 0.9453125,
      "step": 46
    },
    {
      "calib/answer_extract_rate": 0.95703125,
      "calib/avg_num_step_conf": 8.078125,
      "calib/ece": 0.38388114754098346,
      "calib/final_conf_rate": 0.953125,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0013972135144549869,
      "calib/mean_conf": 0.9781434426229507,
      "calib/mu_c": 0.9787103448275862,
      "calib/mu_w": 0.9773131313131312,
      "calib/nonempty_final_conf_rate": 0.953125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.38388114754098346,
      "calib/std_conf": 0.014596482215464805,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9026735653809972,
      "calib/step_q_c_n": 1063.0,
      "calib/step_q_gap": -0.015431907255818667,
      "calib/step_q_w": 0.9181054726368159,
      "calib/step_q_w_n": 1005.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2569.0,
      "completions/max_terminated_length": 2569.0,
      "completions/mean_length": 592.296875,
      "completions/mean_terminated_length": 592.296875,
      "completions/min_length": 209.0,
      "completions/min_terminated_length": 209.0,
      "epoch": 0.050133333333333335,
      "grad_norm": 0.18556036055088043,
      "learning_rate": 4.277777777777778e-06,
      "loss": 0.0751,
      "num_tokens": 10725638.0,
      "reward": 1.044921875,
      "reward_std": 0.3660392761230469,
      "rewards/accuracy_reward_step": 0.5703125,
      "rewards/format_reward_step": 0.94921875,
      "step": 47
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 7.3125,
      "calib/ece": 0.506847389558233,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.9919678714859438,
      "calib/gap": -0.001242229992230004,
      "calib/mean_conf": 0.9767269076305222,
      "calib/mu_c": 0.976068376068376,
      "calib/mu_w": 0.977310606060606,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.506847389558233,
      "calib/std_conf": 0.017063480546151713,
      "calib/step_conf_rate": 0.9921875,
      "calib/step_q_c": 0.9031273644388399,
      "calib/step_q_c_n": 793.0,
      "calib/step_q_gap": -0.0056863519652379235,
      "calib/step_q_w": 0.9088137164040778,
      "calib/step_q_w_n": 1079.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2872.0,
      "completions/max_terminated_length": 2872.0,
      "completions/mean_length": 531.93359375,
      "completions/mean_terminated_length": 531.93359375,
      "completions/min_length": 52.0,
      "completions/min_terminated_length": 52.0,
      "epoch": 0.0512,
      "grad_norm": 0.2239966094493866,
      "learning_rate": 4.25e-06,
      "loss": 0.0354,
      "num_tokens": 10965501.0,
      "reward": 0.943359375,
      "reward_std": 0.3430543541908264,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/format_reward_step": 0.97265625,
      "step": 48
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 7.421875,
      "calib/ece": 0.4573067729083665,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.9920318725099602,
      "calib/gap": 0.0014112523839797664,
      "calib/mean_conf": 0.9752350597609561,
      "calib/mu_c": 0.9759153846153847,
      "calib/mu_w": 0.974504132231405,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4573067729083665,
      "calib/std_conf": 0.018632227090474466,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9006391752577318,
      "calib/step_q_c_n": 970.0,
      "calib/step_q_gap": 0.004569282784613549,
      "calib/step_q_w": 0.8960698924731183,
      "calib/step_q_w_n": 930.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2752.0,
      "completions/max_terminated_length": 2752.0,
      "completions/mean_length": 535.91015625,
      "completions/mean_terminated_length": 535.91015625,
      "completions/min_length": 203.0,
      "completions/min_terminated_length": 203.0,
      "epoch": 0.05226666666666667,
      "grad_norm": 0.15744777023792267,
      "learning_rate": 4.222222222222223e-06,
      "loss": 0.0511,
      "num_tokens": 11207230.0,
      "reward": 0.998046875,
      "reward_std": 0.24411951005458832,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/format_reward_step": 0.98046875,
      "step": 49
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 8.3203125,
      "calib/ece": 0.44008000000000014,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.996,
      "calib/gap": 0.0008572713835869328,
      "calib/mean_conf": 0.9720800000000002,
      "calib/mu_c": 0.9724812030075186,
      "calib/mu_w": 0.9716239316239317,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.44008000000000014,
      "calib/std_conf": 0.01713690753899315,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8968477546549836,
      "calib/step_q_c_n": 913.0,
      "calib/step_q_gap": -0.0060774713105054,
      "calib/step_q_w": 0.902925225965489,
      "calib/step_q_w_n": 1217.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2766.0,
      "completions/max_terminated_length": 2766.0,
      "completions/mean_length": 562.51953125,
      "completions/mean_terminated_length": 569.1897583007812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 157.0,
      "epoch": 0.05333333333333334,
      "grad_norm": 0.1873403787612915,
      "learning_rate": 4.194444444444445e-06,
      "loss": 0.0303,
      "num_tokens": 11456595.0,
      "reward": 1.0078125,
      "reward_std": 0.3558337688446045,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/format_reward_step": 0.9765625,
      "step": 50
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 7.66796875,
      "calib/ece": 0.41886746987951806,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.9959839357429718,
      "calib/gap": -0.0027575401488444395,
      "calib/mean_conf": 0.9730843373493976,
      "calib/mu_c": 0.971855072463768,
      "calib/mu_w": 0.9746126126126125,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.41886746987951806,
      "calib/std_conf": 0.015309029400566207,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8945123456790123,
      "calib/step_q_c_n": 972.0,
      "calib/step_q_gap": -0.005328219406759649,
      "calib/step_q_w": 0.8998405650857719,
      "calib/step_q_w_n": 991.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2979.0,
      "completions/max_terminated_length": 2979.0,
      "completions/mean_length": 591.3359375,
      "completions/mean_terminated_length": 600.7222290039062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 201.0,
      "epoch": 0.0544,
      "grad_norm": 0.16895967721939087,
      "learning_rate": 4.166666666666667e-06,
      "loss": 0.0259,
      "num_tokens": 11717273.0,
      "reward": 1.025390625,
      "reward_std": 0.27671748399734497,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/format_reward_step": 0.97265625,
      "step": 51
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 7.03125,
      "calib/ece": 0.3673769841269843,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.9880952380952381,
      "calib/gap": -0.0009092105263159844,
      "calib/mean_conf": 0.9705515873015874,
      "calib/mu_c": 0.9701907894736842,
      "calib/mu_w": 0.9711000000000002,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3673769841269843,
      "calib/std_conf": 0.017846827267394613,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8935311890838206,
      "calib/step_q_c_n": 1026.0,
      "calib/step_q_gap": -0.0045415930221225365,
      "calib/step_q_w": 0.8980727821059431,
      "calib/step_q_w_n": 774.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2482.0,
      "completions/max_terminated_length": 2482.0,
      "completions/mean_length": 581.5078125,
      "completions/mean_terminated_length": 583.7882690429688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 233.0,
      "epoch": 0.055466666666666664,
      "grad_norm": 0.1962101310491562,
      "learning_rate": 4.138888888888889e-06,
      "loss": 0.0641,
      "num_tokens": 11974091.0,
      "reward": 1.08203125,
      "reward_std": 0.3331637382507324,
      "rewards/accuracy_reward_step": 0.59375,
      "rewards/format_reward_step": 0.9765625,
      "step": 52
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 7.30859375,
      "calib/ece": 0.39697333333333334,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.992,
      "calib/gap": 0.005176350129621143,
      "calib/mean_conf": 0.9689733333333334,
      "calib/mu_c": 0.9711888111888112,
      "calib/mu_w": 0.96601246105919,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.39697333333333334,
      "calib/std_conf": 0.043638557874124725,
      "calib/step_conf_rate": 0.9921875,
      "calib/step_q_c": 0.8942307692307693,
      "calib/step_q_c_n": 988.0,
      "calib/step_q_gap": -0.0021237041554141856,
      "calib/step_q_w": 0.8963544733861835,
      "calib/step_q_w_n": 883.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2854.0,
      "completions/max_terminated_length": 2854.0,
      "completions/mean_length": 591.51953125,
      "completions/mean_terminated_length": 596.1771850585938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 152.0,
      "epoch": 0.05653333333333333,
      "grad_norm": 0.1859111487865448,
      "learning_rate": 4.111111111111111e-06,
      "loss": 0.0109,
      "num_tokens": 12231344.0,
      "reward": 1.04296875,
      "reward_std": 0.34486693143844604,
      "rewards/accuracy_reward_step": 0.55859375,
      "rewards/format_reward_step": 0.96875,
      "step": 53
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 7.765625,
      "calib/ece": 0.33599206349206356,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.9880952380952381,
      "calib/gap": -0.006206127003449025,
      "calib/mean_conf": 0.9669444444444445,
      "calib/mu_c": 0.9646540880503144,
      "calib/mu_w": 0.9708602150537634,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.33599206349206356,
      "calib/std_conf": 0.01889268169204031,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.890474060822898,
      "calib/step_q_c_n": 1118.0,
      "calib/step_q_gap": -0.012905249521929507,
      "calib/step_q_w": 0.9033793103448275,
      "calib/step_q_w_n": 870.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 3032.0,
      "completions/max_terminated_length": 3032.0,
      "completions/mean_length": 556.73046875,
      "completions/mean_terminated_length": 558.9137573242188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 173.0,
      "epoch": 0.0576,
      "grad_norm": 0.18028347194194794,
      "learning_rate": 4.083333333333334e-06,
      "loss": 0.0789,
      "num_tokens": 12480099.0,
      "reward": 1.109375,
      "reward_std": 0.26777738332748413,
      "rewards/accuracy_reward_step": 0.62109375,
      "rewards/format_reward_step": 0.9765625,
      "step": 54
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 7.46875,
      "calib/ece": 0.5246586345381526,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.9839357429718876,
      "calib/gap": 0.002404325032765553,
      "calib/mean_conf": 0.9624096385542168,
      "calib/mu_c": 0.9637614678899082,
      "calib/mu_w": 0.9613571428571427,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5246586345381526,
      "calib/std_conf": 0.021505039744383402,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8906507304116865,
      "calib/step_q_c_n": 753.0,
      "calib/step_q_gap": -0.005639174678908732,
      "calib/step_q_w": 0.8962899050905953,
      "calib/step_q_w_n": 1159.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2682.0,
      "completions/max_terminated_length": 2682.0,
      "completions/mean_length": 562.30078125,
      "completions/mean_terminated_length": 568.9683837890625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 185.0,
      "epoch": 0.058666666666666666,
      "grad_norm": 0.182333305478096,
      "learning_rate": 4.055555555555556e-06,
      "loss": 0.0163,
      "num_tokens": 12731872.0,
      "reward": 0.912109375,
      "reward_std": 0.27818602323532104,
      "rewards/accuracy_reward_step": 0.42578125,
      "rewards/format_reward_step": 0.97265625,
      "step": 55
    },
    {
      "calib/answer_extract_rate": 0.95703125,
      "calib/avg_num_step_conf": 8.05078125,
      "calib/ece": 0.5173673469387754,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 0.9959183673469387,
      "calib/gap": -0.005001683501683396,
      "calib/mean_conf": 0.9663469387755101,
      "calib/mu_c": 0.9635909090909092,
      "calib/mu_w": 0.9685925925925926,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.5173673469387754,
      "calib/std_conf": 0.018311980042526377,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.894064039408867,
      "calib/step_q_c_n": 812.0,
      "calib/step_q_gap": -0.0012281943781625726,
      "calib/step_q_w": 0.8952922337870296,
      "calib/step_q_w_n": 1249.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2483.0,
      "completions/max_terminated_length": 2483.0,
      "completions/mean_length": 634.90234375,
      "completions/mean_terminated_length": 637.3922119140625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 255.0,
      "epoch": 0.05973333333333333,
      "grad_norm": 0.15238536894321442,
      "learning_rate": 4.027777777777779e-06,
      "loss": 0.0984,
      "num_tokens": 13001247.0,
      "reward": 0.904296875,
      "reward_std": 0.2974989414215088,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/format_reward_step": 0.94921875,
      "step": 56
    },
    {
      "calib/answer_extract_rate": 0.9453125,
      "calib/avg_num_step_conf": 8.68359375,
      "calib/ece": 0.3995247933884297,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.9453125,
      "calib/frac_conf_gt_0.9": 0.9917355371900827,
      "calib/gap": 0.003234271810914402,
      "calib/mean_conf": 0.9656404958677686,
      "calib/mu_c": 0.9670437956204381,
      "calib/mu_w": 0.9638095238095237,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3995247933884297,
      "calib/std_conf": 0.022542496428050295,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8948071359691417,
      "calib/step_q_c_n": 1037.0,
      "calib/step_q_gap": -0.018253572293927434,
      "calib/step_q_w": 0.9130607082630692,
      "calib/step_q_w_n": 1186.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 2637.0,
      "completions/max_terminated_length": 2637.0,
      "completions/mean_length": 637.07421875,
      "completions/mean_terminated_length": 649.7649536132812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 305.0,
      "epoch": 0.0608,
      "grad_norm": 0.16103595495224,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0348,
      "num_tokens": 13271130.0,
      "reward": 1.0078125,
      "reward_std": 0.33539533615112305,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/format_reward_step": 0.9453125,
      "step": 57
    },
    {
      "calib/answer_extract_rate": 0.95703125,
      "calib/avg_num_step_conf": 7.85546875,
      "calib/ece": 0.45600000000000007,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.953125,
      "calib/frac_conf_gt_0.9": 0.9877551020408163,
      "calib/gap": 0.0002366666666666628,
      "calib/mean_conf": 0.9662040816326531,
      "calib/mu_c": 0.9663200000000001,
      "calib/mu_w": 0.9660833333333334,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.45600000000000007,
      "calib/std_conf": 0.019812364302592516,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8964467005076142,
      "calib/step_q_c_n": 985.0,
      "calib/step_q_gap": 0.006100696608978828,
      "calib/step_q_w": 0.8903460038986354,
      "calib/step_q_w_n": 1026.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2510.0,
      "completions/max_terminated_length": 2510.0,
      "completions/mean_length": 682.671875,
      "completions/mean_terminated_length": 693.5079956054688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 281.0,
      "epoch": 0.06186666666666667,
      "grad_norm": 0.1501203328371048,
      "learning_rate": 3.972222222222223e-06,
      "loss": 0.0261,
      "num_tokens": 13552214.0,
      "reward": 0.96484375,
      "reward_std": 0.35837113857269287,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/format_reward_step": 0.953125,
      "step": 58
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 7.6328125,
      "calib/ece": 0.4594758064516129,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.9959677419354839,
      "calib/gap": -0.00013723577235780215,
      "calib/mean_conf": 0.963508064516129,
      "calib/mu_c": 0.9634400000000001,
      "calib/mu_w": 0.9635772357723579,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4594758064516129,
      "calib/std_conf": 0.019842929986457408,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.895260180995475,
      "calib/step_q_c_n": 884.0,
      "calib/step_q_gap": -0.007169725546581129,
      "calib/step_q_w": 0.9024299065420561,
      "calib/step_q_w_n": 1070.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2862.0,
      "completions/max_terminated_length": 2862.0,
      "completions/mean_length": 660.53515625,
      "completions/mean_terminated_length": 663.1255493164062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 209.0,
      "epoch": 0.06293333333333333,
      "grad_norm": 0.1628417819738388,
      "learning_rate": 3.944444444444445e-06,
      "loss": 0.0698,
      "num_tokens": 13827559.0,
      "reward": 0.970703125,
      "reward_std": 0.313901424407959,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/format_reward_step": 0.96484375,
      "step": 59
    },
    {
      "calib/answer_extract_rate": 0.9609375,
      "calib/avg_num_step_conf": 7.6640625,
      "calib/ece": 0.5115447154471544,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00614472281449896,
      "calib/mean_conf": 0.9668292682926828,
      "calib/mu_c": 0.9634821428571428,
      "calib/mu_w": 0.9696268656716418,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.5115447154471544,
      "calib/std_conf": 0.01945673011117428,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.894878640776699,
      "calib/step_q_c_n": 824.0,
      "calib/step_q_gap": -0.005226807377958442,
      "calib/step_q_w": 0.9001054481546574,
      "calib/step_q_w_n": 1138.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 3001.0,
      "completions/max_terminated_length": 3001.0,
      "completions/mean_length": 682.29296875,
      "completions/mean_terminated_length": 687.6653442382812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 287.0,
      "epoch": 0.064,
      "grad_norm": 0.16230860352516174,
      "learning_rate": 3.916666666666667e-06,
      "loss": 0.0715,
      "num_tokens": 14111082.0,
      "reward": 0.91796875,
      "reward_std": 0.332300066947937,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/format_reward_step": 0.9609375,
      "step": 60
    },
    {
      "calib/answer_extract_rate": 0.953125,
      "calib/avg_num_step_conf": 8.92578125,
      "calib/ece": 0.3466803278688526,
      "calib/final_conf_rate": 0.953125,
      "calib/format_rate": 0.953125,
      "calib/frac_conf_gt_0.9": 0.9959016393442623,
      "calib/gap": 0.001121555223243087,
      "calib/mean_conf": 0.965532786885246,
      "calib/mu_c": 0.9659602649006623,
      "calib/mu_w": 0.9648387096774192,
      "calib/nonempty_final_conf_rate": 0.953125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3466803278688526,
      "calib/std_conf": 0.019900442546454743,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8966260869565217,
      "calib/step_q_c_n": 1150.0,
      "calib/step_q_gap": -0.001743957096341786,
      "calib/step_q_w": 0.8983700440528635,
      "calib/step_q_w_n": 1135.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2936.0,
      "completions/max_terminated_length": 2936.0,
      "completions/mean_length": 639.625,
      "completions/mean_terminated_length": 642.1333618164062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 204.0,
      "epoch": 0.06506666666666666,
      "grad_norm": 0.1348682940006256,
      "learning_rate": 3.88888888888889e-06,
      "loss": 0.1511,
      "num_tokens": 14378890.0,
      "reward": 1.06640625,
      "reward_std": 0.28018099069595337,
      "rewards/accuracy_reward_step": 0.58984375,
      "rewards/format_reward_step": 0.953125,
      "step": 61
    },
    {
      "calib/answer_extract_rate": 0.9375,
      "calib/avg_num_step_conf": 8.50390625,
      "calib/ece": 0.4697095435684647,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00022933884297526852,
      "calib/mean_conf": 0.9676348547717842,
      "calib/mu_c": 0.9677500000000001,
      "calib/mu_w": 0.9675206611570248,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4697095435684647,
      "calib/std_conf": 0.017326719732256125,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8983193779904305,
      "calib/step_q_c_n": 836.0,
      "calib/step_q_gap": 0.006620645701094108,
      "calib/step_q_w": 0.8916987322893364,
      "calib/step_q_w_n": 1341.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 3011.0,
      "completions/max_terminated_length": 3011.0,
      "completions/mean_length": 709.62109375,
      "completions/mean_terminated_length": 720.8849487304688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 248.0,
      "epoch": 0.06613333333333334,
      "grad_norm": 0.14342306554317474,
      "learning_rate": 3.861111111111112e-06,
      "loss": 0.0776,
      "num_tokens": 14667633.0,
      "reward": 0.9375,
      "reward_std": 0.33649760484695435,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/format_reward_step": 0.9375,
      "step": 62
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 7.171875,
      "calib/ece": 0.4264940239043825,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.9920318725099602,
      "calib/gap": -0.009075989782886151,
      "calib/mean_conf": 0.9643426294820717,
      "calib/mu_c": 0.9601481481481482,
      "calib/mu_w": 0.9692241379310343,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.4264940239043825,
      "calib/std_conf": 0.019876765136435336,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.8941853360488797,
      "calib/step_q_c_n": 982.0,
      "calib/step_q_gap": -0.002231525777818155,
      "calib/step_q_w": 0.8964168618266979,
      "calib/step_q_w_n": 854.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2880.0,
      "completions/max_terminated_length": 2880.0,
      "completions/mean_length": 699.48828125,
      "completions/mean_terminated_length": 702.2313842773438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 240.0,
      "epoch": 0.0672,
      "grad_norm": 0.1324063539505005,
      "learning_rate": 3.833333333333334e-06,
      "loss": 0.0506,
      "num_tokens": 14955342.0,
      "reward": 1.015625,
      "reward_std": 0.2985878586769104,
      "rewards/accuracy_reward_step": 0.52734375,
      "rewards/format_reward_step": 0.9765625,
      "step": 63
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 7.59375,
      "calib/ece": 0.3233919999999999,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0009152777777776677,
      "calib/mean_conf": 0.9633919999999999,
      "calib/mu_c": 0.9630625,
      "calib/mu_w": 0.9639777777777777,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3233919999999999,
      "calib/std_conf": 0.017738949687058702,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8929639401934916,
      "calib/step_q_c_n": 1137.0,
      "calib/step_q_gap": -0.005779554230300299,
      "calib/step_q_w": 0.8987434944237919,
      "calib/step_q_w_n": 807.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2398.0,
      "completions/max_terminated_length": 2398.0,
      "completions/mean_length": 655.84765625,
      "completions/mean_terminated_length": 661.0117797851562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 231.0,
      "epoch": 0.06826666666666667,
      "grad_norm": 0.1513136774301529,
      "learning_rate": 3.8055555555555556e-06,
      "loss": 0.0724,
      "num_tokens": 15227015.0,
      "reward": 1.11328125,
      "reward_std": 0.3160572648048401,
      "rewards/accuracy_reward_step": 0.625,
      "rewards/format_reward_step": 0.9765625,
      "step": 64
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 7.1171875,
      "calib/ece": 0.3814624505928854,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.9920948616600791,
      "calib/gap": 0.0010383776151970947,
      "calib/mean_conf": 0.9624901185770751,
      "calib/mu_c": 0.9629251700680272,
      "calib/mu_w": 0.9618867924528302,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.3814624505928854,
      "calib/std_conf": 0.020287579223183028,
      "calib/step_conf_rate": 0.9921875,
      "calib/step_q_c": 0.895662188099808,
      "calib/step_q_c_n": 1042.0,
      "calib/step_q_gap": -0.0013249913873714103,
      "calib/step_q_w": 0.8969871794871794,
      "calib/step_q_w_n": 780.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2471.0,
      "completions/max_terminated_length": 2471.0,
      "completions/mean_length": 601.99609375,
      "completions/mean_terminated_length": 601.99609375,
      "completions/min_length": 255.0,
      "completions/min_terminated_length": 255.0,
      "epoch": 0.06933333333333333,
      "grad_norm": 0.13157853484153748,
      "learning_rate": 3.777777777777778e-06,
      "loss": 0.0275,
      "num_tokens": 15486150.0,
      "reward": 1.064453125,
      "reward_std": 0.22381740808486938,
      "rewards/accuracy_reward_step": 0.57421875,
      "rewards/format_reward_step": 0.98046875,
      "step": 65
    },
    {
      "calib/answer_extract_rate": 0.9453125,
      "calib/avg_num_step_conf": 7.77734375,
      "calib/ece": 0.4534710743801653,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.9453125,
      "calib/frac_conf_gt_0.9": 0.9958677685950413,
      "calib/gap": -0.0015453986472636627,
      "calib/mean_conf": 0.9617355371900826,
      "calib/mu_c": 0.9609756097560977,
      "calib/mu_w": 0.9625210084033614,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.4534710743801653,
      "calib/std_conf": 0.01922797918110774,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.8947882736156351,
      "calib/step_q_c_n": 921.0,
      "calib/step_q_gap": -0.0035201375993182404,
      "calib/step_q_w": 0.8983084112149533,
      "calib/step_q_w_n": 1070.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2815.0,
      "completions/max_terminated_length": 2815.0,
      "completions/mean_length": 745.53125,
      "completions/mean_terminated_length": 748.4549560546875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 252.0,
      "epoch": 0.0704,
      "grad_norm": 0.13143111765384674,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.0906,
      "num_tokens": 15783358.0,
      "reward": 0.953125,
      "reward_std": 0.28545472025871277,
      "rewards/accuracy_reward_step": 0.48046875,
      "rewards/format_reward_step": 0.9453125,
      "step": 66
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 7.4296875,
      "calib/ece": 0.38402390438247,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0017703318152244574,
      "calib/mean_conf": 0.9617131474103585,
      "calib/mu_c": 0.9609655172413791,
      "calib/mu_w": 0.9627358490566036,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.38402390438247,
      "calib/std_conf": 0.018289760117441916,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.897305447470817,
      "calib/step_q_c_n": 1028.0,
      "calib/step_q_gap": 0.0025800470131510655,
      "calib/step_q_w": 0.8947254004576659,
      "calib/step_q_w_n": 874.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2388.0,
      "completions/max_terminated_length": 2388.0,
      "completions/mean_length": 692.1015625,
      "completions/mean_terminated_length": 694.8157348632812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 253.0,
      "epoch": 0.07146666666666666,
      "grad_norm": 0.1354527324438095,
      "learning_rate": 3.7222222222222225e-06,
      "loss": 0.0546,
      "num_tokens": 16065544.0,
      "reward": 1.056640625,
      "reward_std": 0.22899171710014343,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/format_reward_step": 0.98046875,
      "step": 67
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 7.85546875,
      "calib/ece": 0.41004048582995933,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.9635627530364372,
      "calib/gap": -0.0029550264550266547,
      "calib/mean_conf": 0.9565991902834007,
      "calib/mu_c": 0.9552592592592591,
      "calib/mu_w": 0.9582142857142858,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.41004048582995933,
      "calib/std_conf": 0.021665783004325333,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.892875,
      "calib/step_q_c_n": 960.0,
      "calib/step_q_gap": -0.0020060656517602027,
      "calib/step_q_w": 0.8948810656517602,
      "calib/step_q_w_n": 1051.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2717.0,
      "completions/max_terminated_length": 2717.0,
      "completions/mean_length": 662.22265625,
      "completions/mean_terminated_length": 664.8196411132812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 212.0,
      "epoch": 0.07253333333333334,
      "grad_norm": 0.14325231313705444,
      "learning_rate": 3.694444444444445e-06,
      "loss": 0.0253,
      "num_tokens": 16339161.0,
      "reward": 1.013671875,
      "reward_std": 0.22165964543819427,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/format_reward_step": 0.96484375,
      "step": 68
    },
    {
      "calib/answer_extract_rate": 0.92578125,
      "calib/avg_num_step_conf": 7.5625,
      "calib/ece": 0.4172573839662447,
      "calib/final_conf_rate": 0.92578125,
      "calib/format_rate": 0.92578125,
      "calib/frac_conf_gt_0.9": 0.9831223628691983,
      "calib/gap": -0.0003354357798164287,
      "calib/mean_conf": 0.9573417721518986,
      "calib/mu_c": 0.9571875000000001,
      "calib/mu_w": 0.9575229357798165,
      "calib/nonempty_final_conf_rate": 0.92578125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4172573839662447,
      "calib/std_conf": 0.020869831299338597,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8965642151481887,
      "calib/step_q_c_n": 911.0,
      "calib/step_q_gap": 0.0017203127091643955,
      "calib/step_q_w": 0.8948439024390243,
      "calib/step_q_w_n": 1025.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2957.0,
      "completions/max_terminated_length": 2957.0,
      "completions/mean_length": 777.046875,
      "completions/mean_terminated_length": 789.3809814453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 290.0,
      "epoch": 0.0736,
      "grad_norm": 0.13327306509017944,
      "learning_rate": 3.6666666666666666e-06,
      "loss": 0.0825,
      "num_tokens": 16642581.0,
      "reward": 0.962890625,
      "reward_std": 0.2655287981033325,
      "rewards/accuracy_reward_step": 0.5,
      "rewards/format_reward_step": 0.92578125,
      "step": 69
    },
    {
      "calib/answer_extract_rate": 0.953125,
      "calib/avg_num_step_conf": 7.8203125,
      "calib/ece": 0.47163934426229515,
      "calib/final_conf_rate": 0.953125,
      "calib/format_rate": 0.953125,
      "calib/frac_conf_gt_0.9": 0.9672131147540983,
      "calib/gap": -0.008681732580037904,
      "calib/mean_conf": 0.9552459016393443,
      "calib/mu_c": 0.9507627118644066,
      "calib/mu_w": 0.9594444444444445,
      "calib/nonempty_final_conf_rate": 0.953125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.47163934426229515,
      "calib/std_conf": 0.019823529435219485,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8922645739910313,
      "calib/step_q_c_n": 892.0,
      "calib/step_q_gap": 0.0011564658829230634,
      "calib/step_q_w": 0.8911081081081083,
      "calib/step_q_w_n": 1110.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2907.0,
      "completions/max_terminated_length": 2907.0,
      "completions/mean_length": 734.734375,
      "completions/mean_terminated_length": 743.4466552734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 234.0,
      "epoch": 0.07466666666666667,
      "grad_norm": 0.13847662508487701,
      "learning_rate": 3.638888888888889e-06,
      "loss": 0.0989,
      "num_tokens": 16937665.0,
      "reward": 0.9375,
      "reward_std": 0.29290473461151123,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/format_reward_step": 0.953125,
      "step": 70
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 8.8984375,
      "calib/ece": 0.4378137651821862,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.97165991902834,
      "calib/gap": -0.001003937007874267,
      "calib/mean_conf": 0.9519838056680161,
      "calib/mu_c": 0.9514960629921257,
      "calib/mu_w": 0.9525,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4378137651821862,
      "calib/std_conf": 0.018846056001520246,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8933234421364985,
      "calib/step_q_c_n": 1011.0,
      "calib/step_q_gap": 0.003220837558755796,
      "calib/step_q_w": 0.8901026045777427,
      "calib/step_q_w_n": 1267.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 3036.0,
      "completions/max_terminated_length": 3036.0,
      "completions/mean_length": 706.60546875,
      "completions/mean_terminated_length": 712.1693115234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 276.0,
      "epoch": 0.07573333333333333,
      "grad_norm": 0.14184655249118805,
      "learning_rate": 3.6111111111111115e-06,
      "loss": 0.044,
      "num_tokens": 17222964.0,
      "reward": 0.978515625,
      "reward_std": 0.3507859408855438,
      "rewards/accuracy_reward_step": 0.49609375,
      "rewards/format_reward_step": 0.96484375,
      "step": 71
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 7.8046875,
      "calib/ece": 0.4324999999999999,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.9761904761904762,
      "calib/gap": -0.006306857611506955,
      "calib/mean_conf": 0.9523412698412698,
      "calib/mu_c": 0.9493129770992368,
      "calib/mu_w": 0.9556198347107437,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4324999999999999,
      "calib/std_conf": 0.020248324523440216,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8914902363823227,
      "calib/step_q_c_n": 973.0,
      "calib/step_q_gap": 0.0035487729676886293,
      "calib/step_q_w": 0.8879414634146341,
      "calib/step_q_w_n": 1025.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2827.0,
      "completions/max_terminated_length": 2827.0,
      "completions/mean_length": 668.3984375,
      "completions/mean_terminated_length": 671.0196533203125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 303.0,
      "epoch": 0.0768,
      "grad_norm": 0.15166209638118744,
      "learning_rate": 3.5833333333333335e-06,
      "loss": 0.0593,
      "num_tokens": 17498482.0,
      "reward": 1.001953125,
      "reward_std": 0.29228222370147705,
      "rewards/accuracy_reward_step": 0.51171875,
      "rewards/format_reward_step": 0.98046875,
      "step": 72
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 7.65234375,
      "calib/ece": 0.3369169960474308,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.9762845849802372,
      "calib/gap": 0.00045622119815658113,
      "calib/mean_conf": 0.9495652173913044,
      "calib/mu_c": 0.9497419354838709,
      "calib/mu_w": 0.9492857142857143,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.3369169960474308,
      "calib/std_conf": 0.02027987708730161,
      "calib/step_conf_rate": 0.9921875,
      "calib/step_q_c": 0.8940547476475621,
      "calib/step_q_c_n": 1169.0,
      "calib/step_q_gap": 0.00519398815389116,
      "calib/step_q_w": 0.8888607594936709,
      "calib/step_q_w_n": 790.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1846.0,
      "completions/max_terminated_length": 1846.0,
      "completions/mean_length": 652.015625,
      "completions/mean_terminated_length": 652.015625,
      "completions/min_length": 280.0,
      "completions/min_terminated_length": 280.0,
      "epoch": 0.07786666666666667,
      "grad_norm": 0.1515604704618454,
      "learning_rate": 3.555555555555556e-06,
      "loss": 0.0284,
      "num_tokens": 17772430.0,
      "reward": 1.095703125,
      "reward_std": 0.30325573682785034,
      "rewards/accuracy_reward_step": 0.60546875,
      "rewards/format_reward_step": 0.98046875,
      "step": 73
    },
    {
      "calib/answer_extract_rate": 0.9375,
      "calib/avg_num_step_conf": 7.64453125,
      "calib/ece": 0.43337499999999995,
      "calib/final_conf_rate": 0.9375,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 0.9416666666666667,
      "calib/gap": -0.003423248053392558,
      "calib/mean_conf": 0.9500416666666667,
      "calib/mu_c": 0.9483870967741936,
      "calib/mu_w": 0.9518103448275862,
      "calib/nonempty_final_conf_rate": 0.9375,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.43337499999999995,
      "calib/std_conf": 0.02096620130644132,
      "calib/step_conf_rate": 0.98828125,
      "calib/step_q_c": 0.8894485683987274,
      "calib/step_q_c_n": 943.0,
      "calib/step_q_gap": 0.0018252942369917369,
      "calib/step_q_w": 0.8876232741617357,
      "calib/step_q_w_n": 1014.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2468.0,
      "completions/max_terminated_length": 2468.0,
      "completions/mean_length": 676.2421875,
      "completions/mean_terminated_length": 686.9762573242188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 232.0,
      "epoch": 0.07893333333333333,
      "grad_norm": 0.1617853343486786,
      "learning_rate": 3.5277777777777784e-06,
      "loss": -0.0034,
      "num_tokens": 18049476.0,
      "reward": 0.953125,
      "reward_std": 0.31296655535697937,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/format_reward_step": 0.9375,
      "step": 74
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 8.03515625,
      "calib/ece": 0.28699604743083007,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.9565217391304348,
      "calib/gap": -0.005362044817927347,
      "calib/mean_conf": 0.9510276679841898,
      "calib/mu_c": 0.9492261904761905,
      "calib/mu_w": 0.9545882352941178,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.28699604743083007,
      "calib/std_conf": 0.020052579730554358,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8908532934131735,
      "calib/step_q_c_n": 1336.0,
      "calib/step_q_gap": 0.0002985083923691123,
      "calib/step_q_w": 0.8905547850208044,
      "calib/step_q_w_n": 721.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2663.0,
      "completions/max_terminated_length": 2663.0,
      "completions/mean_length": 631.2265625,
      "completions/mean_terminated_length": 631.2265625,
      "completions/min_length": 213.0,
      "completions/min_terminated_length": 213.0,
      "epoch": 0.08,
      "grad_norm": 0.16196872293949127,
      "learning_rate": 3.5e-06,
      "loss": 0.0076,
      "num_tokens": 18315822.0,
      "reward": 1.150390625,
      "reward_std": 0.298104852437973,
      "rewards/accuracy_reward_step": 0.65625,
      "rewards/format_reward_step": 0.98828125,
      "step": 75
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 7.78125,
      "calib/ece": 0.35212,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.948,
      "calib/gap": -0.0011333333333335416,
      "calib/mean_conf": 0.95212,
      "calib/mu_c": 0.9516666666666664,
      "calib/mu_w": 0.9528,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.35212,
      "calib/std_conf": 0.02106906737375909,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.895199291408326,
      "calib/step_q_c_n": 1129.0,
      "calib/step_q_gap": 0.007169163945985324,
      "calib/step_q_w": 0.8880301274623407,
      "calib/step_q_w_n": 863.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2138.0,
      "completions/max_terminated_length": 2138.0,
      "completions/mean_length": 682.734375,
      "completions/mean_terminated_length": 693.5714721679688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 297.0,
      "epoch": 0.08106666666666666,
      "grad_norm": 0.13606837391853333,
      "learning_rate": 3.4722222222222224e-06,
      "loss": -0.0013,
      "num_tokens": 18593658.0,
      "reward": 1.07421875,
      "reward_std": 0.26280534267425537,
      "rewards/accuracy_reward_step": 0.5859375,
      "rewards/format_reward_step": 0.9765625,
      "step": 76
    },
    {
      "calib/answer_extract_rate": 0.94921875,
      "calib/avg_num_step_conf": 8.21484375,
      "calib/ece": 0.3220576131687243,
      "calib/final_conf_rate": 0.94921875,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.934156378600823,
      "calib/gap": -0.0020849673202613106,
      "calib/mean_conf": 0.9516872427983539,
      "calib/mu_c": 0.9509150326797386,
      "calib/mu_w": 0.953,
      "calib/nonempty_final_conf_rate": 0.94921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.3220576131687243,
      "calib/std_conf": 0.021373441314687876,
      "calib/step_conf_rate": 0.98828125,
      "calib/step_q_c": 0.8877159468438538,
      "calib/step_q_c_n": 1204.0,
      "calib/step_q_gap": 0.006236525264320947,
      "calib/step_q_w": 0.8814794215795329,
      "calib/step_q_w_n": 899.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2677.0,
      "completions/max_terminated_length": 2677.0,
      "completions/mean_length": 689.83203125,
      "completions/mean_terminated_length": 698.0119018554688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 194.0,
      "epoch": 0.08213333333333334,
      "grad_norm": 0.15443819761276245,
      "learning_rate": 3.444444444444445e-06,
      "loss": 0.0233,
      "num_tokens": 18874919.0,
      "reward": 1.068359375,
      "reward_std": 0.29310113191604614,
      "rewards/accuracy_reward_step": 0.59765625,
      "rewards/format_reward_step": 0.94140625,
      "step": 77
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 8.1640625,
      "calib/ece": 0.361195219123506,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.9123505976095617,
      "calib/gap": -0.004065861978483354,
      "calib/mean_conf": 0.9496414342629482,
      "calib/mu_c": 0.947972972972973,
      "calib/mu_w": 0.9520388349514564,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.360597609561753,
      "calib/std_conf": 0.023231703942294266,
      "calib/step_conf_rate": 0.984375,
      "calib/step_q_c": 0.8854640980735552,
      "calib/step_q_c_n": 1142.0,
      "calib/step_q_gap": -0.00010552217960946297,
      "calib/step_q_w": 0.8855696202531647,
      "calib/step_q_w_n": 948.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 3060.0,
      "completions/max_terminated_length": 3060.0,
      "completions/mean_length": 711.49609375,
      "completions/mean_terminated_length": 714.2863159179688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 292.0,
      "epoch": 0.0832,
      "grad_norm": 0.14217936992645264,
      "learning_rate": 3.416666666666667e-06,
      "loss": 0.0421,
      "num_tokens": 19165086.0,
      "reward": 1.0625,
      "reward_std": 0.28406330943107605,
      "rewards/accuracy_reward_step": 0.578125,
      "rewards/format_reward_step": 0.96875,
      "step": 78
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 8.00390625,
      "calib/ece": 0.30577689243027895,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.9561752988047809,
      "calib/gap": 0.003775794757389761,
      "calib/mean_conf": 0.9551792828685259,
      "calib/mu_c": 0.9565030674846625,
      "calib/mu_w": 0.9527272727272728,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.30577689243027895,
      "calib/std_conf": 0.0209216840057658,
      "calib/step_conf_rate": 0.98828125,
      "calib/step_q_c": 0.8896686746987952,
      "calib/step_q_c_n": 1328.0,
      "calib/step_q_gap": -0.0016073308490550664,
      "calib/step_q_w": 0.8912760055478502,
      "calib/step_q_w_n": 721.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2850.0,
      "completions/max_terminated_length": 2850.0,
      "completions/mean_length": 692.79296875,
      "completions/mean_terminated_length": 698.248046875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 358.0,
      "epoch": 0.08426666666666667,
      "grad_norm": 0.14957371354103088,
      "learning_rate": 3.3888888888888893e-06,
      "loss": 0.0032,
      "num_tokens": 19448817.0,
      "reward": 1.123046875,
      "reward_std": 0.2835806608200073,
      "rewards/accuracy_reward_step": 0.63671875,
      "rewards/format_reward_step": 0.97265625,
      "step": 79
    },
    {
      "calib/answer_extract_rate": 0.99609375,
      "calib/avg_num_step_conf": 8.76171875,
      "calib/ece": 0.32219607843137255,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.9333333333333333,
      "calib/gap": -0.00024514338575398753,
      "calib/mean_conf": 0.9535686274509803,
      "calib/mu_c": 0.9534782608695651,
      "calib/mu_w": 0.9537234042553191,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.32219607843137255,
      "calib/std_conf": 0.02371832658016091,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.8882774049217003,
      "calib/step_q_c_n": 1341.0,
      "calib/step_q_gap": -0.006212617251248664,
      "calib/step_q_w": 0.8944900221729489,
      "calib/step_q_w_n": 902.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1933.0,
      "completions/max_terminated_length": 1933.0,
      "completions/mean_length": 629.71484375,
      "completions/mean_terminated_length": 632.184326171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 318.0,
      "epoch": 0.08533333333333333,
      "grad_norm": 0.17289277911186218,
      "learning_rate": 3.3611111111111117e-06,
      "loss": 0.0163,
      "num_tokens": 19712184.0,
      "reward": 1.125,
      "reward_std": 0.3098638653755188,
      "rewards/accuracy_reward_step": 0.62890625,
      "rewards/format_reward_step": 0.9921875,
      "step": 80
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 9.24609375,
      "calib/ece": 0.32431451612903217,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.9556451612903226,
      "calib/gap": -0.0031465997770345755,
      "calib/mean_conf": 0.9533467741935483,
      "calib/mu_c": 0.9521794871794873,
      "calib/mu_w": 0.9553260869565219,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.32431451612903217,
      "calib/std_conf": 0.01985046694617709,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.8929888712241655,
      "calib/step_q_c_n": 1258.0,
      "calib/step_q_gap": 0.003538916309828166,
      "calib/step_q_w": 0.8894499549143373,
      "calib/step_q_w_n": 1109.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 2736.0,
      "completions/max_terminated_length": 2736.0,
      "completions/mean_length": 659.609375,
      "completions/mean_terminated_length": 675.4400024414062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 272.0,
      "epoch": 0.0864,
      "grad_norm": 0.15319500863552094,
      "learning_rate": 3.3333333333333333e-06,
      "loss": -0.0001,
      "num_tokens": 19987292.0,
      "reward": 1.091796875,
      "reward_std": 0.2939770221710205,
      "rewards/accuracy_reward_step": 0.609375,
      "rewards/format_reward_step": 0.96484375,
      "step": 81
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 8.80859375,
      "calib/ece": 0.3080158730158729,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.9682539682539683,
      "calib/gap": 0.003663059212793973,
      "calib/mean_conf": 0.9548412698412697,
      "calib/mu_c": 0.9561349693251535,
      "calib/mu_w": 0.9524719101123595,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.3080158730158729,
      "calib/std_conf": 0.024192814326865104,
      "calib/step_conf_rate": 0.9921875,
      "calib/step_q_c": 0.8965398420674804,
      "calib/step_q_c_n": 1393.0,
      "calib/step_q_gap": 0.011180213297178665,
      "calib/step_q_w": 0.8853596287703017,
      "calib/step_q_w_n": 862.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2553.0,
      "completions/max_terminated_length": 2553.0,
      "completions/mean_length": 653.3515625,
      "completions/mean_terminated_length": 655.9137573242188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 283.0,
      "epoch": 0.08746666666666666,
      "grad_norm": 0.16056913137435913,
      "learning_rate": 3.3055555555555558e-06,
      "loss": 0.0332,
      "num_tokens": 20260102.0,
      "reward": 1.125,
      "reward_std": 0.28356438875198364,
      "rewards/accuracy_reward_step": 0.63671875,
      "rewards/format_reward_step": 0.9765625,
      "step": 82
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 9.03515625,
      "calib/ece": 0.4024899598393574,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.9558232931726908,
      "calib/gap": -0.006429690560125434,
      "calib/mean_conf": 0.956706827309237,
      "calib/mu_c": 0.9538405797101449,
      "calib/mu_w": 0.9602702702702703,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4024899598393574,
      "calib/std_conf": 0.02036805893731143,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8947774979009234,
      "calib/step_q_c_n": 1191.0,
      "calib/step_q_gap": 0.005731152089871738,
      "calib/step_q_w": 0.8890463458110517,
      "calib/step_q_w_n": 1122.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 3031.0,
      "completions/max_terminated_length": 3031.0,
      "completions/mean_length": 754.16796875,
      "completions/mean_terminated_length": 760.1063232421875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 326.0,
      "epoch": 0.08853333333333334,
      "grad_norm": 0.13364803791046143,
      "learning_rate": 3.277777777777778e-06,
      "loss": 0.0606,
      "num_tokens": 20560433.0,
      "reward": 1.025390625,
      "reward_std": 0.22642236948013306,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/format_reward_step": 0.97265625,
      "step": 83
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 8.7109375,
      "calib/ece": 0.4169565217391304,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.9604743083003953,
      "calib/gap": -4.4620412267382115e-05,
      "calib/mean_conf": 0.954505928853755,
      "calib/mu_c": 0.954485294117647,
      "calib/mu_w": 0.9545299145299144,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4169565217391304,
      "calib/std_conf": 0.02147369390827028,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.89727868852459,
      "calib/step_q_c_n": 1220.0,
      "calib/step_q_gap": 0.0010806687226097011,
      "calib/step_q_w": 0.8961980198019803,
      "calib/step_q_w_n": 1010.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2600.0,
      "completions/max_terminated_length": 2600.0,
      "completions/mean_length": 652.03515625,
      "completions/mean_terminated_length": 652.03515625,
      "completions/min_length": 215.0,
      "completions/min_terminated_length": 215.0,
      "epoch": 0.0896,
      "grad_norm": 0.19131937623023987,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.0183,
      "num_tokens": 20833274.0,
      "reward": 1.025390625,
      "reward_std": 0.23556412756443024,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/format_reward_step": 0.98828125,
      "step": 84
    },
    {
      "calib/answer_extract_rate": 0.94140625,
      "calib/avg_num_step_conf": 9.48828125,
      "calib/ece": 0.3969709543568466,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 0.991701244813278,
      "calib/gap": 0.0021106442577029094,
      "calib/mean_conf": 0.961286307053942,
      "calib/mu_c": 0.962205882352941,
      "calib/mu_w": 0.9600952380952381,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3969709543568466,
      "calib/std_conf": 0.016155525860297164,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9003043110735418,
      "calib/step_q_c_n": 1183.0,
      "calib/step_q_gap": 0.002639784588790439,
      "calib/step_q_w": 0.8976645264847514,
      "calib/step_q_w_n": 1246.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 2915.0,
      "completions/max_terminated_length": 2915.0,
      "completions/mean_length": 733.83984375,
      "completions/mean_terminated_length": 751.4520263671875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 349.0,
      "epoch": 0.09066666666666667,
      "grad_norm": 0.1620008498430252,
      "learning_rate": 3.2222222222222227e-06,
      "loss": 0.0209,
      "num_tokens": 21128961.0,
      "reward": 1.0,
      "reward_std": 0.29893848299980164,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/format_reward_step": 0.9375,
      "step": 85
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 9.078125,
      "calib/ece": 0.4341666666666667,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.996031746031746,
      "calib/gap": 0.0021945357955981404,
      "calib/mean_conf": 0.9659126984126984,
      "calib/mu_c": 0.9669402985074627,
      "calib/mu_w": 0.9647457627118645,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4341666666666667,
      "calib/std_conf": 0.014892482305730816,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9049746621621619,
      "calib/step_q_c_n": 1184.0,
      "calib/step_q_gap": 0.0014571183025128276,
      "calib/step_q_w": 0.9035175438596491,
      "calib/step_q_w_n": 1140.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2919.0,
      "completions/max_terminated_length": 2919.0,
      "completions/mean_length": 697.66796875,
      "completions/mean_terminated_length": 703.1614379882812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 212.0,
      "epoch": 0.09173333333333333,
      "grad_norm": 0.16130724549293518,
      "learning_rate": 3.1944444444444443e-06,
      "loss": 0.0167,
      "num_tokens": 21413076.0,
      "reward": 1.015625,
      "reward_std": 0.2880152463912964,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/format_reward_step": 0.984375,
      "step": 86
    },
    {
      "calib/answer_extract_rate": 0.94140625,
      "calib/avg_num_step_conf": 10.81640625,
      "calib/ece": 0.28087136929460577,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.995850622406639,
      "calib/gap": -0.0016433734939759193,
      "calib/mean_conf": 0.9696680497925311,
      "calib/mu_c": 0.9691566265060241,
      "calib/mu_w": 0.9708,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.28087136929460577,
      "calib/std_conf": 0.014371110367996593,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9044444444444444,
      "calib/step_q_c_n": 1530.0,
      "calib/step_q_gap": -0.007928436911487924,
      "calib/step_q_w": 0.9123728813559323,
      "calib/step_q_w_n": 1239.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 2897.0,
      "completions/max_terminated_length": 2897.0,
      "completions/mean_length": 626.60546875,
      "completions/mean_terminated_length": 657.422119140625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 230.0,
      "epoch": 0.0928,
      "grad_norm": 0.18359941244125366,
      "learning_rate": 3.1666666666666667e-06,
      "loss": -0.0573,
      "num_tokens": 21678983.0,
      "reward": 1.119140625,
      "reward_std": 0.32765060663223267,
      "rewards/accuracy_reward_step": 0.6484375,
      "rewards/format_reward_step": 0.94140625,
      "step": 87
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 9.68359375,
      "calib/ece": 0.35642857142857143,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0031220485533753894,
      "calib/mean_conf": 0.9715079365079365,
      "calib/mu_c": 0.9727096774193548,
      "calib/mu_w": 0.9695876288659794,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.35642857142857143,
      "calib/std_conf": 0.010470477037499024,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9039452843772497,
      "calib/step_q_c_n": 1389.0,
      "calib/step_q_gap": -0.002485908283300753,
      "calib/step_q_w": 0.9064311926605505,
      "calib/step_q_w_n": 1090.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2887.0,
      "completions/max_terminated_length": 2887.0,
      "completions/mean_length": 726.24609375,
      "completions/mean_terminated_length": 734.8577270507812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 367.0,
      "epoch": 0.09386666666666667,
      "grad_norm": 0.14980705082416534,
      "learning_rate": 3.138888888888889e-06,
      "loss": 0.0058,
      "num_tokens": 21974750.0,
      "reward": 1.095703125,
      "reward_std": 0.23912689089775085,
      "rewards/accuracy_reward_step": 0.60546875,
      "rewards/format_reward_step": 0.98046875,
      "step": 88
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 9.45703125,
      "calib/ece": 0.4387,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00040530108080294447,
      "calib/mean_conf": 0.9747,
      "calib/mu_c": 0.9748880597014925,
      "calib/mu_w": 0.9744827586206896,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.4387,
      "calib/std_conf": 0.011764777940955802,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9084339774557166,
      "calib/step_q_c_n": 1242.0,
      "calib/step_q_gap": -0.0014726439364733945,
      "calib/step_q_w": 0.90990662139219,
      "calib/step_q_w_n": 1178.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2393.0,
      "completions/max_terminated_length": 2393.0,
      "completions/mean_length": 757.140625,
      "completions/mean_terminated_length": 763.1023559570312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 370.0,
      "epoch": 0.09493333333333333,
      "grad_norm": 0.14294065535068512,
      "learning_rate": 3.1111111111111116e-06,
      "loss": -0.0063,
      "num_tokens": 22277466.0,
      "reward": 1.0078125,
      "reward_std": 0.24918082356452942,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/format_reward_step": 0.96875,
      "step": 89
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 11.67578125,
      "calib/ece": 0.3420647773279353,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.9959514170040485,
      "calib/gap": -0.0025824175824175066,
      "calib/mean_conf": 0.9736437246963563,
      "calib/mu_c": 0.9726923076923077,
      "calib/mu_w": 0.9752747252747253,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3420647773279353,
      "calib/std_conf": 0.0131513530920179,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9066065879428217,
      "calib/step_q_c_n": 1609.0,
      "calib/step_q_gap": -8.906423109134032e-05,
      "calib/step_q_w": 0.906695652173913,
      "calib/step_q_w_n": 1380.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 2940.0,
      "completions/max_terminated_length": 2940.0,
      "completions/mean_length": 720.05859375,
      "completions/mean_terminated_length": 737.3400268554688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 311.0,
      "epoch": 0.096,
      "grad_norm": 0.12898504734039307,
      "learning_rate": 3.0833333333333336e-06,
      "loss": -0.0053,
      "num_tokens": 22565121.0,
      "reward": 1.091796875,
      "reward_std": 0.24940425157546997,
      "rewards/accuracy_reward_step": 0.609375,
      "rewards/format_reward_step": 0.96484375,
      "step": 90
    },
    {
      "calib/answer_extract_rate": 0.9609375,
      "calib/avg_num_step_conf": 10.08203125,
      "calib/ece": 0.3303252032520325,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0016005204944699036,
      "calib/mean_conf": 0.9766666666666667,
      "calib/mu_c": 0.9772327044025155,
      "calib/mu_w": 0.9756321839080456,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3303252032520325,
      "calib/std_conf": 0.013624861700896784,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.909741935483871,
      "calib/step_q_c_n": 1550.0,
      "calib/step_q_gap": -0.0014413816839273785,
      "calib/step_q_w": 0.9111833171677983,
      "calib/step_q_w_n": 1031.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2841.0,
      "completions/max_terminated_length": 2841.0,
      "completions/mean_length": 778.375,
      "completions/mean_terminated_length": 790.730224609375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 439.0,
      "epoch": 0.09706666666666666,
      "grad_norm": 0.14882391691207886,
      "learning_rate": 3.055555555555556e-06,
      "loss": -0.0026,
      "num_tokens": 22872097.0,
      "reward": 1.1015625,
      "reward_std": 0.2827257513999939,
      "rewards/accuracy_reward_step": 0.62109375,
      "rewards/format_reward_step": 0.9609375,
      "step": 91
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 10.078125,
      "calib/ece": 0.3375396825396826,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0014620162446249418,
      "calib/mean_conf": 0.9764285714285714,
      "calib/mu_c": 0.9769565217391304,
      "calib/mu_w": 0.9754945054945054,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3375396825396826,
      "calib/std_conf": 0.013999109139132488,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9123418491484184,
      "calib/step_q_c_n": 1644.0,
      "calib/step_q_gap": 0.002165567097136445,
      "calib/step_q_w": 0.910176282051282,
      "calib/step_q_w_n": 936.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2523.0,
      "completions/max_terminated_length": 2523.0,
      "completions/mean_length": 685.27734375,
      "completions/mean_terminated_length": 690.6732177734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 331.0,
      "epoch": 0.09813333333333334,
      "grad_norm": 0.1833995282649994,
      "learning_rate": 3.0277777777777776e-06,
      "loss": -0.0212,
      "num_tokens": 23154248.0,
      "reward": 1.115234375,
      "reward_std": 0.2714332938194275,
      "rewards/accuracy_reward_step": 0.62890625,
      "rewards/format_reward_step": 0.97265625,
      "step": 92
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 11.328125,
      "calib/ece": 0.4156275303643724,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0010711430855315784,
      "calib/mean_conf": 0.9783805668016193,
      "calib/mu_c": 0.9788489208633092,
      "calib/mu_w": 0.9777777777777776,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4156275303643724,
      "calib/std_conf": 0.013845917084890089,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.913076923076923,
      "calib/step_q_c_n": 1443.0,
      "calib/step_q_gap": 0.0008188585607941112,
      "calib/step_q_w": 0.9122580645161289,
      "calib/step_q_w_n": 1457.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 2794.0,
      "completions/max_terminated_length": 2794.0,
      "completions/mean_length": 710.171875,
      "completions/mean_terminated_length": 724.3187255859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 317.0,
      "epoch": 0.0992,
      "grad_norm": 0.1591537445783615,
      "learning_rate": 3e-06,
      "loss": 0.0032,
      "num_tokens": 23441828.0,
      "reward": 1.025390625,
      "reward_std": 0.36088210344314575,
      "rewards/accuracy_reward_step": 0.54296875,
      "rewards/format_reward_step": 0.96484375,
      "step": 93
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 9.3515625,
      "calib/ece": 0.3926360000000001,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.002385773726966245,
      "calib/mean_conf": 0.9806360000000001,
      "calib/mu_c": 0.9796530612244897,
      "calib/mu_w": 0.9820388349514559,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3926360000000001,
      "calib/std_conf": 0.013215124062981788,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9103254310344827,
      "calib/step_q_c_n": 1392.0,
      "calib/step_q_gap": 0.0039132553857802455,
      "calib/step_q_w": 0.9064121756487025,
      "calib/step_q_w_n": 1002.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2695.0,
      "completions/max_terminated_length": 2695.0,
      "completions/mean_length": 676.484375,
      "completions/mean_terminated_length": 681.81103515625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 361.0,
      "epoch": 0.10026666666666667,
      "grad_norm": 0.15960992872714996,
      "learning_rate": 2.9722222222222225e-06,
      "loss": 0.0035,
      "num_tokens": 23723688.0,
      "reward": 1.0625,
      "reward_std": 0.27309247851371765,
      "rewards/accuracy_reward_step": 0.57421875,
      "rewards/format_reward_step": 0.9765625,
      "step": 94
    },
    {
      "calib/answer_extract_rate": 0.9921875,
      "calib/avg_num_step_conf": 8.98046875,
      "calib/ece": 0.4093110236220473,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.002019930401771508,
      "calib/mean_conf": 0.9801771653543307,
      "calib/mu_c": 0.9793103448275862,
      "calib/mu_w": 0.9813302752293577,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.4093110236220473,
      "calib/std_conf": 0.01225018520815402,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9075780089153045,
      "calib/step_q_c_n": 1346.0,
      "calib/step_q_gap": 0.003385773868085251,
      "calib/step_q_w": 0.9041922350472192,
      "calib/step_q_w_n": 953.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1588.0,
      "completions/max_terminated_length": 1588.0,
      "completions/mean_length": 711.0,
      "completions/mean_terminated_length": 713.7882690429688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 267.0,
      "epoch": 0.10133333333333333,
      "grad_norm": 0.15578389167785645,
      "learning_rate": 2.944444444444445e-06,
      "loss": 0.0051,
      "num_tokens": 24011832.0,
      "reward": 1.060546875,
      "reward_std": 0.2807765603065491,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/format_reward_step": 0.98828125,
      "step": 95
    },
    {
      "calib/answer_extract_rate": 0.99609375,
      "calib/avg_num_step_conf": 8.54296875,
      "calib/ece": 0.23486274509803928,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0027437827225134104,
      "calib/mean_conf": 0.9838823529411765,
      "calib/mu_c": 0.9831937172774867,
      "calib/mu_w": 0.9859375000000001,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.23486274509803928,
      "calib/std_conf": 0.010674593363837137,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9067102803738317,
      "calib/step_q_c_n": 1605.0,
      "calib/step_q_gap": -0.002155699007611589,
      "calib/step_q_w": 0.9088659793814433,
      "calib/step_q_w_n": 582.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2198.0,
      "completions/max_terminated_length": 2198.0,
      "completions/mean_length": 633.14453125,
      "completions/mean_terminated_length": 633.14453125,
      "completions/min_length": 260.0,
      "completions/min_terminated_length": 260.0,
      "epoch": 0.1024,
      "grad_norm": 0.16016392409801483,
      "learning_rate": 2.916666666666667e-06,
      "loss": 0.0198,
      "num_tokens": 24279733.0,
      "reward": 1.244140625,
      "reward_std": 0.17649205029010773,
      "rewards/accuracy_reward_step": 0.74609375,
      "rewards/format_reward_step": 0.99609375,
      "step": 96
    },
    {
      "calib/answer_extract_rate": 0.9921875,
      "calib/avg_num_step_conf": 8.6796875,
      "calib/ece": 0.3653937007874015,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.002666622890537651,
      "calib/mean_conf": 0.983503937007874,
      "calib/mu_c": 0.9845222929936305,
      "calib/mu_w": 0.9818556701030928,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.3653937007874015,
      "calib/std_conf": 0.011769019798592162,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.908495575221239,
      "calib/step_q_c_n": 1356.0,
      "calib/step_q_gap": 0.0027219031658117787,
      "calib/step_q_w": 0.9057736720554272,
      "calib/step_q_w_n": 866.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2498.0,
      "completions/max_terminated_length": 2498.0,
      "completions/mean_length": 687.04296875,
      "completions/mean_terminated_length": 689.7373046875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 294.0,
      "epoch": 0.10346666666666667,
      "grad_norm": 0.17087474465370178,
      "learning_rate": 2.888888888888889e-06,
      "loss": 0.0099,
      "num_tokens": 24560688.0,
      "reward": 1.10546875,
      "reward_std": 0.2656993269920349,
      "rewards/accuracy_reward_step": 0.61328125,
      "rewards/format_reward_step": 0.984375,
      "step": 97
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 8.8984375,
      "calib/ece": 0.37048192771084343,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.000945669934640736,
      "calib/mean_conf": 0.9849397590361446,
      "calib/mu_c": 0.9845751633986928,
      "calib/mu_w": 0.9855208333333335,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.37048192771084343,
      "calib/std_conf": 0.011269623026294539,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9078645066273933,
      "calib/step_q_c_n": 1358.0,
      "calib/step_q_gap": 0.0008373327143498255,
      "calib/step_q_w": 0.9070271739130434,
      "calib/step_q_w_n": 920.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2856.0,
      "completions/max_terminated_length": 2856.0,
      "completions/mean_length": 717.7890625,
      "completions/mean_terminated_length": 723.44091796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 283.0,
      "epoch": 0.10453333333333334,
      "grad_norm": 0.16385309398174286,
      "learning_rate": 2.861111111111111e-06,
      "loss": 0.0191,
      "num_tokens": 24850626.0,
      "reward": 1.0859375,
      "reward_std": 0.29347917437553406,
      "rewards/accuracy_reward_step": 0.6015625,
      "rewards/format_reward_step": 0.96875,
      "step": 98
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 8.84375,
      "calib/ece": 0.566008064516129,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.003771367521367641,
      "calib/mean_conf": 0.9853629032258064,
      "calib/mu_c": 0.9831730769230769,
      "calib/mu_w": 0.9869444444444445,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.566008064516129,
      "calib/std_conf": 0.011317866746887967,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.906843729903537,
      "calib/step_q_c_n": 933.0,
      "calib/step_q_gap": 0.004226900451996918,
      "calib/step_q_w": 0.9026168294515401,
      "calib/step_q_w_n": 1331.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 2636.0,
      "completions/max_terminated_length": 2636.0,
      "completions/mean_length": 738.0546875,
      "completions/mean_terminated_length": 755.76806640625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 310.0,
      "epoch": 0.1056,
      "grad_norm": 0.14243488013744354,
      "learning_rate": 2.8333333333333335e-06,
      "loss": -0.0466,
      "num_tokens": 25145368.0,
      "reward": 0.890625,
      "reward_std": 0.2674334943294525,
      "rewards/accuracy_reward_step": 0.40625,
      "rewards/format_reward_step": 0.96875,
      "step": 99
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 7.9921875,
      "calib/ece": 0.42079600000000017,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0005894332747736764,
      "calib/mean_conf": 0.9847960000000001,
      "calib/mu_c": 0.9845390070921985,
      "calib/mu_w": 0.9851284403669722,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.42079600000000017,
      "calib/std_conf": 0.011173288862282235,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9026158038147138,
      "calib/step_q_c_n": 1101.0,
      "calib/step_q_gap": -0.0034847252858152267,
      "calib/step_q_w": 0.9061005291005291,
      "calib/step_q_w_n": 945.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2747.0,
      "completions/max_terminated_length": 2747.0,
      "completions/mean_length": 711.9609375,
      "completions/mean_terminated_length": 711.9609375,
      "completions/min_length": 345.0,
      "completions/min_terminated_length": 345.0,
      "epoch": 0.10666666666666667,
      "grad_norm": 0.16134221851825714,
      "learning_rate": 2.805555555555556e-06,
      "loss": 0.0355,
      "num_tokens": 25435038.0,
      "reward": 1.0390625,
      "reward_std": 0.2604951858520508,
      "rewards/accuracy_reward_step": 0.55078125,
      "rewards/format_reward_step": 0.9765625,
      "step": 100
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 8.37109375,
      "calib/ece": 0.5147389558232933,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0022630147630147146,
      "calib/mean_conf": 0.9846184738955824,
      "calib/mu_c": 0.9834188034188034,
      "calib/mu_w": 0.9856818181818181,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.5147389558232933,
      "calib/std_conf": 0.01144021565431309,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9019877049180327,
      "calib/step_q_c_n": 976.0,
      "calib/step_q_gap": -0.001585559863458319,
      "calib/step_q_w": 0.903573264781491,
      "calib/step_q_w_n": 1167.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2612.0,
      "completions/max_terminated_length": 2612.0,
      "completions/mean_length": 743.0703125,
      "completions/mean_terminated_length": 751.8814697265625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 289.0,
      "epoch": 0.10773333333333333,
      "grad_norm": 0.16494955122470856,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 0.0291,
      "num_tokens": 25732256.0,
      "reward": 0.939453125,
      "reward_std": 0.3241780698299408,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/format_reward_step": 0.96484375,
      "step": 101
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 8.24609375,
      "calib/ece": 0.3281200000000001,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.996,
      "calib/gap": -0.00382302892796349,
      "calib/mean_conf": 0.9841200000000001,
      "calib/mu_c": 0.9828048780487806,
      "calib/mu_w": 0.9866279069767441,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3281200000000001,
      "calib/std_conf": 0.012009396321214493,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9031399564902102,
      "calib/step_q_c_n": 1379.0,
      "calib/step_q_gap": -0.0012452894114289847,
      "calib/step_q_w": 0.9043852459016392,
      "calib/step_q_w_n": 732.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2417.0,
      "completions/max_terminated_length": 2417.0,
      "completions/mean_length": 638.1484375,
      "completions/mean_terminated_length": 640.6510009765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 345.0,
      "epoch": 0.1088,
      "grad_norm": 0.16382494568824768,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.0055,
      "num_tokens": 26002318.0,
      "reward": 1.125,
      "reward_std": 0.22881178557872772,
      "rewards/accuracy_reward_step": 0.640625,
      "rewards/format_reward_step": 0.96875,
      "step": 102
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 7.57421875,
      "calib/ece": 0.3984337349397592,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0006250831227556652,
      "calib/mean_conf": 0.9847791164658636,
      "calib/mu_c": 0.9845205479452055,
      "calib/mu_w": 0.9851456310679612,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3984337349397592,
      "calib/std_conf": 0.009696778624122087,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9016399286987522,
      "calib/step_q_c_n": 1122.0,
      "calib/step_q_gap": 0.001933686348691066,
      "calib/step_q_w": 0.8997062423500611,
      "calib/step_q_w_n": 817.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2663.0,
      "completions/max_terminated_length": 2663.0,
      "completions/mean_length": 734.54296875,
      "completions/mean_terminated_length": 743.2529907226562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 235.0,
      "epoch": 0.10986666666666667,
      "grad_norm": 0.12834714353084564,
      "learning_rate": 2.7222222222222224e-06,
      "loss": -0.0067,
      "num_tokens": 26294913.0,
      "reward": 1.056640625,
      "reward_std": 0.22446666657924652,
      "rewards/accuracy_reward_step": 0.5703125,
      "rewards/format_reward_step": 0.97265625,
      "step": 103
    },
    {
      "calib/answer_extract_rate": 0.99609375,
      "calib/avg_num_step_conf": 7.81640625,
      "calib/ece": 0.4489803921568628,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0008554991958429348,
      "calib/mean_conf": 0.9862352941176471,
      "calib/mu_c": 0.9858394160583942,
      "calib/mu_w": 0.9866949152542371,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4489803921568628,
      "calib/std_conf": 0.009071801246426833,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9019178082191781,
      "calib/step_q_c_n": 1095.0,
      "calib/step_q_gap": -0.0005656354894311155,
      "calib/step_q_w": 0.9024834437086092,
      "calib/step_q_w_n": 906.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2795.0,
      "completions/max_terminated_length": 2795.0,
      "completions/mean_length": 672.34765625,
      "completions/mean_terminated_length": 672.34765625,
      "completions/min_length": 270.0,
      "completions/min_terminated_length": 270.0,
      "epoch": 0.11093333333333333,
      "grad_norm": 0.16741102933883667,
      "learning_rate": 2.6944444444444444e-06,
      "loss": 0.0343,
      "num_tokens": 26573714.0,
      "reward": 1.033203125,
      "reward_std": 0.24684098362922668,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/format_reward_step": 0.99609375,
      "step": 104
    },
    {
      "calib/answer_extract_rate": 0.99609375,
      "calib/avg_num_step_conf": 8.171875,
      "calib/ece": 0.38137254901960793,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 7.715057220025745e-05,
      "calib/mean_conf": 0.9852941176470589,
      "calib/mu_c": 0.9853246753246754,
      "calib/mu_w": 0.9852475247524751,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.38137254901960793,
      "calib/std_conf": 0.009892152004507259,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.902717041800643,
      "calib/step_q_c_n": 1244.0,
      "calib/step_q_gap": -0.0008843732936965054,
      "calib/step_q_w": 0.9036014150943396,
      "calib/step_q_w_n": 848.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2202.0,
      "completions/max_terminated_length": 2202.0,
      "completions/mean_length": 737.6484375,
      "completions/mean_terminated_length": 740.5411987304688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 296.0,
      "epoch": 0.112,
      "grad_norm": 0.14760896563529968,
      "learning_rate": 2.666666666666667e-06,
      "loss": 0.0258,
      "num_tokens": 26868312.0,
      "reward": 1.09765625,
      "reward_std": 0.20331215858459473,
      "rewards/accuracy_reward_step": 0.6015625,
      "rewards/format_reward_step": 0.9921875,
      "step": 105
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 7.6015625,
      "calib/ece": 0.43822834645669284,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00318861432593065,
      "calib/mean_conf": 0.9854724409448818,
      "calib/mu_c": 0.9840287769784173,
      "calib/mu_w": 0.9872173913043479,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.43822834645669284,
      "calib/std_conf": 0.009368507113614807,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.8987144259077526,
      "calib/step_q_c_n": 1019.0,
      "calib/step_q_gap": -0.004867019615440338,
      "calib/step_q_w": 0.903581445523193,
      "calib/step_q_w_n": 927.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2742.0,
      "completions/max_terminated_length": 2742.0,
      "completions/mean_length": 672.83203125,
      "completions/mean_terminated_length": 672.83203125,
      "completions/min_length": 321.0,
      "completions/min_terminated_length": 321.0,
      "epoch": 0.11306666666666666,
      "grad_norm": 0.18181079626083374,
      "learning_rate": 2.6388888888888893e-06,
      "loss": 0.0294,
      "num_tokens": 27145141.0,
      "reward": 1.037109375,
      "reward_std": 0.27421265840530396,
      "rewards/accuracy_reward_step": 0.54296875,
      "rewards/format_reward_step": 0.98828125,
      "step": 106
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 8.1171875,
      "calib/ece": 0.28341269841269856,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.996031746031746,
      "calib/gap": -0.0014327683615816822,
      "calib/mean_conf": 0.9857936507936509,
      "calib/mu_c": 0.9853672316384182,
      "calib/mu_w": 0.9867999999999999,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.28341269841269856,
      "calib/std_conf": 0.01160549778058649,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9035870331219169,
      "calib/step_q_c_n": 1419.0,
      "calib/step_q_gap": -0.0038939987445474333,
      "calib/step_q_w": 0.9074810318664643,
      "calib/step_q_w_n": 659.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2731.0,
      "completions/max_terminated_length": 2731.0,
      "completions/mean_length": 662.66796875,
      "completions/mean_terminated_length": 667.8858032226562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 329.0,
      "epoch": 0.11413333333333334,
      "grad_norm": 0.20197457075119019,
      "learning_rate": 2.6111111111111113e-06,
      "loss": -0.0018,
      "num_tokens": 27419400.0,
      "reward": 1.177734375,
      "reward_std": 0.3254070281982422,
      "rewards/accuracy_reward_step": 0.69140625,
      "rewards/format_reward_step": 0.97265625,
      "step": 107
    },
    {
      "calib/answer_extract_rate": 0.9609375,
      "calib/avg_num_step_conf": 8.56640625,
      "calib/ece": 0.2770564516129034,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0004924242424242387,
      "calib/mean_conf": 0.9867338709677421,
      "calib/mu_c": 0.986590909090909,
      "calib/mu_w": 0.9870833333333332,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.2770564516129034,
      "calib/std_conf": 0.009170085368237865,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.906232782369146,
      "calib/step_q_c_n": 1452.0,
      "calib/step_q_gap": -0.00043388429752055924,
      "calib/step_q_w": 0.9066666666666665,
      "calib/step_q_w_n": 741.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2815.0,
      "completions/max_terminated_length": 2815.0,
      "completions/mean_length": 717.0625,
      "completions/mean_terminated_length": 728.4444580078125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 261.0,
      "epoch": 0.1152,
      "grad_norm": 0.16753298044204712,
      "learning_rate": 2.5833333333333337e-06,
      "loss": -0.0362,
      "num_tokens": 27706200.0,
      "reward": 1.16796875,
      "reward_std": 0.30900561809539795,
      "rewards/accuracy_reward_step": 0.6875,
      "rewards/format_reward_step": 0.9609375,
      "step": 108
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 9.51171875,
      "calib/ece": 0.48186234817813767,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0010026229508198758,
      "calib/mean_conf": 0.9879352226720648,
      "calib/mu_c": 0.98744,
      "calib/mu_w": 0.9884426229508199,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.48186234817813767,
      "calib/std_conf": 0.007805054824974812,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9067155963302752,
      "calib/step_q_c_n": 1090.0,
      "calib/step_q_gap": -0.00024351147641632487,
      "calib/step_q_w": 0.9069591078066915,
      "calib/step_q_w_n": 1345.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 3028.0,
      "completions/max_terminated_length": 3028.0,
      "completions/mean_length": 727.7890625,
      "completions/mean_terminated_length": 739.34130859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 371.0,
      "epoch": 0.11626666666666667,
      "grad_norm": 0.14441752433776855,
      "learning_rate": 2.5555555555555557e-06,
      "loss": -0.0015,
      "num_tokens": 27997114.0,
      "reward": 0.96875,
      "reward_std": 0.226576566696167,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/format_reward_step": 0.9609375,
      "step": 109
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 8.71875,
      "calib/ece": 0.39039370078740154,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0006475748194016973,
      "calib/mean_conf": 0.9888188976377953,
      "calib/mu_c": 0.9890789473684211,
      "calib/mu_w": 0.9884313725490194,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.39039370078740154,
      "calib/std_conf": 0.007275692810407256,
      "calib/step_conf_rate": 0.9921875,
      "calib/step_q_c": 0.9081130507065669,
      "calib/step_q_c_n": 1203.0,
      "calib/step_q_gap": -0.00981697844795204,
      "calib/step_q_w": 0.9179300291545189,
      "calib/step_q_w_n": 1029.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1968.0,
      "completions/max_terminated_length": 1968.0,
      "completions/mean_length": 663.86328125,
      "completions/mean_terminated_length": 666.4667358398438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 261.0,
      "epoch": 0.11733333333333333,
      "grad_norm": 0.15150228142738342,
      "learning_rate": 2.5277777777777778e-06,
      "loss": 0.0132,
      "num_tokens": 28271983.0,
      "reward": 1.0859375,
      "reward_std": 0.25357189774513245,
      "rewards/accuracy_reward_step": 0.59375,
      "rewards/format_reward_step": 0.984375,
      "step": 110
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 8.94921875,
      "calib/ece": 0.3785943775100403,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0012920510037981447,
      "calib/mean_conf": 0.9890361445783135,
      "calib/mu_c": 0.9895394736842105,
      "calib/mu_w": 0.9882474226804123,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.3785943775100403,
      "calib/std_conf": 0.007436120112295035,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9105534495830174,
      "calib/step_q_c_n": 1319.0,
      "calib/step_q_gap": -0.007152311733854999,
      "calib/step_q_w": 0.9177057613168724,
      "calib/step_q_w_n": 972.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2780.0,
      "completions/max_terminated_length": 2780.0,
      "completions/mean_length": 717.40234375,
      "completions/mean_terminated_length": 725.9091186523438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 338.0,
      "epoch": 0.1184,
      "grad_norm": 0.1705169975757599,
      "learning_rate": 2.5e-06,
      "loss": -0.0176,
      "num_tokens": 28563046.0,
      "reward": 1.080078125,
      "reward_std": 0.33822542428970337,
      "rewards/accuracy_reward_step": 0.59375,
      "rewards/format_reward_step": 0.97265625,
      "step": 111
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 8.578125,
      "calib/ece": 0.39368000000000014,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0010352847365274753,
      "calib/mean_conf": 0.9896800000000001,
      "calib/mu_c": 0.9892617449664428,
      "calib/mu_w": 0.9902970297029703,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.39368000000000014,
      "calib/std_conf": 0.0075297808733056845,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9101587301587303,
      "calib/step_q_c_n": 1323.0,
      "calib/step_q_gap": -0.001021109474717674,
      "calib/step_q_w": 0.911179839633448,
      "calib/step_q_w_n": 873.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 2876.0,
      "completions/max_terminated_length": 2876.0,
      "completions/mean_length": 707.21875,
      "completions/mean_terminated_length": 721.3068237304688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 360.0,
      "epoch": 0.11946666666666667,
      "grad_norm": 0.16528332233428955,
      "learning_rate": 2.4722222222222226e-06,
      "loss": -0.028,
      "num_tokens": 28852014.0,
      "reward": 1.0703125,
      "reward_std": 0.23935158550739288,
      "rewards/accuracy_reward_step": 0.58203125,
      "rewards/format_reward_step": 0.9765625,
      "step": 112
    },
    {
      "calib/answer_extract_rate": 0.9921875,
      "calib/avg_num_step_conf": 9.5703125,
      "calib/ece": 0.40708661417322833,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0005673125956144887,
      "calib/mean_conf": 0.989763779527559,
      "calib/mu_c": 0.989527027027027,
      "calib/mu_w": 0.9900943396226415,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.40708661417322833,
      "calib/std_conf": 0.007629673415280169,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9150035486160397,
      "calib/step_q_c_n": 1409.0,
      "calib/step_q_gap": -0.00504448212363362,
      "calib/step_q_w": 0.9200480307396733,
      "calib/step_q_w_n": 1041.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2690.0,
      "completions/max_terminated_length": 2690.0,
      "completions/mean_length": 689.4140625,
      "completions/mean_terminated_length": 692.11767578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 378.0,
      "epoch": 0.12053333333333334,
      "grad_norm": 0.17403432726860046,
      "learning_rate": 2.4444444444444447e-06,
      "loss": 0.0031,
      "num_tokens": 29133704.0,
      "reward": 1.07421875,
      "reward_std": 0.268023818731308,
      "rewards/accuracy_reward_step": 0.578125,
      "rewards/format_reward_step": 0.9921875,
      "step": 113
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 10.5546875,
      "calib/ece": 0.3272936507936508,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0003670306445929805,
      "calib/mean_conf": 0.9899920634920635,
      "calib/mu_c": 0.9898682634730539,
      "calib/mu_w": 0.9902352941176469,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3272936507936508,
      "calib/std_conf": 0.0062246259934130655,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9154449938195304,
      "calib/step_q_c_n": 1618.0,
      "calib/step_q_gap": -0.017676777398181853,
      "calib/step_q_w": 0.9331217712177122,
      "calib/step_q_w_n": 1084.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 3060.0,
      "completions/max_terminated_length": 3060.0,
      "completions/mean_length": 676.90625,
      "completions/mean_terminated_length": 682.2362060546875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 282.0,
      "epoch": 0.1216,
      "grad_norm": 0.1270322948694229,
      "learning_rate": 2.4166666666666667e-06,
      "loss": 0.0156,
      "num_tokens": 29412016.0,
      "reward": 1.14453125,
      "reward_std": 0.1828223466873169,
      "rewards/accuracy_reward_step": 0.65234375,
      "rewards/format_reward_step": 0.984375,
      "step": 114
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 9.7109375,
      "calib/ece": 0.46885375494071146,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0001308539944905318,
      "calib/mean_conf": 0.9905928853754941,
      "calib/mu_c": 0.990530303030303,
      "calib/mu_w": 0.9906611570247935,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.46885375494071146,
      "calib/std_conf": 0.005696912413070543,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9160295395308427,
      "calib/step_q_c_n": 1151.0,
      "calib/step_q_gap": -0.00796296983245337,
      "calib/step_q_w": 0.923992509363296,
      "calib/step_q_w_n": 1335.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2682.0,
      "completions/max_terminated_length": 2682.0,
      "completions/mean_length": 703.52734375,
      "completions/mean_terminated_length": 706.2863159179688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 368.0,
      "epoch": 0.12266666666666666,
      "grad_norm": 0.16146261990070343,
      "learning_rate": 2.388888888888889e-06,
      "loss": 0.0307,
      "num_tokens": 29697383.0,
      "reward": 1.0078125,
      "reward_std": 0.23582719266414642,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/format_reward_step": 0.984375,
      "step": 115
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 10.7890625,
      "calib/ece": 0.399436,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00020747217806049711,
      "calib/mean_conf": 0.991436,
      "calib/mu_c": 0.9913513513513511,
      "calib/mu_w": 0.9915588235294116,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.399436,
      "calib/std_conf": 0.005609091191984675,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9153267045454547,
      "calib/step_q_c_n": 1408.0,
      "calib/step_q_gap": -0.020064728246273478,
      "calib/step_q_w": 0.9353914327917282,
      "calib/step_q_w_n": 1354.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2712.0,
      "completions/max_terminated_length": 2712.0,
      "completions/mean_length": 726.54296875,
      "completions/mean_terminated_length": 738.075439453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 292.0,
      "epoch": 0.12373333333333333,
      "grad_norm": 0.15040576457977295,
      "learning_rate": 2.361111111111111e-06,
      "loss": 0.0051,
      "num_tokens": 29987898.0,
      "reward": 1.06640625,
      "reward_std": 0.30778148770332336,
      "rewards/accuracy_reward_step": 0.578125,
      "rewards/format_reward_step": 0.9765625,
      "step": 116
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 10.1953125,
      "calib/ece": 0.490796,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0008080000000000309,
      "calib/mean_conf": 0.990796,
      "calib/mu_c": 0.9911999999999999,
      "calib/mu_w": 0.9903919999999998,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.490796,
      "calib/std_conf": 0.00514688099726427,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9176000000000002,
      "calib/step_q_c_n": 1150.0,
      "calib/step_q_gap": -0.008115068493150357,
      "calib/step_q_w": 0.9257150684931506,
      "calib/step_q_w_n": 1460.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 2937.0,
      "completions/max_terminated_length": 2937.0,
      "completions/mean_length": 698.19140625,
      "completions/mean_terminated_length": 712.099609375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 345.0,
      "epoch": 0.1248,
      "grad_norm": 0.16612644493579865,
      "learning_rate": 2.3333333333333336e-06,
      "loss": -0.0003,
      "num_tokens": 30273235.0,
      "reward": 0.9765625,
      "reward_std": 0.3108536899089813,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/format_reward_step": 0.9765625,
      "step": 117
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 10.29296875,
      "calib/ece": 0.43342231075697224,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0013442728442730756,
      "calib/mean_conf": 0.991191235059761,
      "calib/mu_c": 0.9917857142857143,
      "calib/mu_w": 0.9904414414414412,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.43342231075697224,
      "calib/std_conf": 0.004648960264656658,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.911376488095238,
      "calib/step_q_c_n": 1344.0,
      "calib/step_q_gap": -0.012913209813360194,
      "calib/step_q_w": 0.9242896979085982,
      "calib/step_q_w_n": 1291.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 1806.0,
      "completions/max_terminated_length": 1806.0,
      "completions/mean_length": 679.640625,
      "completions/mean_terminated_length": 693.1793212890625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 334.0,
      "epoch": 0.12586666666666665,
      "grad_norm": 0.14906100928783417,
      "learning_rate": 2.305555555555556e-06,
      "loss": -0.0502,
      "num_tokens": 30551231.0,
      "reward": 1.037109375,
      "reward_std": 0.2077845185995102,
      "rewards/accuracy_reward_step": 0.546875,
      "rewards/format_reward_step": 0.98046875,
      "step": 118
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 9.76953125,
      "calib/ece": 0.4146774193548388,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 9.057609057583349e-05,
      "calib/mean_conf": 0.9912903225806453,
      "calib/mu_c": 0.9913286713286711,
      "calib/mu_w": 0.9912380952380953,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4146774193548388,
      "calib/std_conf": 0.0036956255604482613,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9124657534246575,
      "calib/step_q_c_n": 1168.0,
      "calib/step_q_gap": 0.00017768140665308518,
      "calib/step_q_w": 0.9122880720180044,
      "calib/step_q_w_n": 1333.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 2870.0,
      "completions/max_terminated_length": 2870.0,
      "completions/mean_length": 726.0859375,
      "completions/mean_terminated_length": 746.4979858398438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 319.0,
      "epoch": 0.12693333333333334,
      "grad_norm": 0.15550687909126282,
      "learning_rate": 2.277777777777778e-06,
      "loss": -0.0534,
      "num_tokens": 30842173.0,
      "reward": 1.04296875,
      "reward_std": 0.2845311462879181,
      "rewards/accuracy_reward_step": 0.55859375,
      "rewards/format_reward_step": 0.96875,
      "step": 119
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 9.30078125,
      "calib/ece": 0.3415243741765482,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.9960474308300395,
      "calib/gap": -0.0036297979797981217,
      "calib/mean_conf": 0.9884281949934123,
      "calib/mu_c": 0.9871656565656564,
      "calib/mu_w": 0.9907954545454545,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3388893280632412,
      "calib/std_conf": 0.04153036208504551,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9176109643328929,
      "calib/step_q_c_n": 1514.0,
      "calib/step_q_gap": 0.040852025463227304,
      "calib/step_q_w": 0.8767589388696656,
      "calib/step_q_w_n": 867.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 1216.0,
      "completions/max_terminated_length": 1216.0,
      "completions/mean_length": 659.55859375,
      "completions/mean_terminated_length": 667.3794555664062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 374.0,
      "epoch": 0.128,
      "grad_norm": 0.13183537125587463,
      "learning_rate": 2.25e-06,
      "loss": -0.0255,
      "num_tokens": 31117708.0,
      "reward": 1.138671875,
      "reward_std": 0.18556472659111023,
      "rewards/accuracy_reward_step": 0.64453125,
      "rewards/format_reward_step": 0.98828125,
      "step": 120
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 8.5078125,
      "calib/ece": 0.4200396825396827,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00013888888888913264,
      "calib/mean_conf": 0.9914682539682541,
      "calib/mu_c": 0.9915277777777778,
      "calib/mu_w": 0.9913888888888887,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.4200396825396827,
      "calib/std_conf": 0.003649715161611516,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9121318144833198,
      "calib/step_q_c_n": 1229.0,
      "calib/step_q_gap": -0.0012612308275337103,
      "calib/step_q_w": 0.9133930453108535,
      "calib/step_q_w_n": 949.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2249.0,
      "completions/max_terminated_length": 2249.0,
      "completions/mean_length": 719.19140625,
      "completions/mean_terminated_length": 724.8543090820312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 310.0,
      "epoch": 0.12906666666666666,
      "grad_norm": 0.17723466455936432,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.0058,
      "num_tokens": 31406877.0,
      "reward": 1.0546875,
      "reward_std": 0.3372430205345154,
      "rewards/accuracy_reward_step": 0.5625,
      "rewards/format_reward_step": 0.984375,
      "step": 121
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 8.2734375,
      "calib/ece": 0.34686400000000017,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0005077814222903543,
      "calib/mean_conf": 0.9908640000000002,
      "calib/mu_c": 0.9906832298136646,
      "calib/mu_w": 0.991191011235955,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.34686400000000017,
      "calib/std_conf": 0.0038664588449898217,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9112178517397883,
      "calib/step_q_c_n": 1322.0,
      "calib/step_q_gap": -0.0012720980089555578,
      "calib/step_q_w": 0.9124899497487439,
      "calib/step_q_w_n": 796.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2680.0,
      "completions/max_terminated_length": 2680.0,
      "completions/mean_length": 700.76953125,
      "completions/mean_terminated_length": 706.2874145507812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 350.0,
      "epoch": 0.13013333333333332,
      "grad_norm": 0.16055411100387573,
      "learning_rate": 2.1944444444444445e-06,
      "loss": 0.0187,
      "num_tokens": 31693618.0,
      "reward": 1.1171875,
      "reward_std": 0.21959704160690308,
      "rewards/accuracy_reward_step": 0.62890625,
      "rewards/format_reward_step": 0.9765625,
      "step": 122
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 8.6484375,
      "calib/ece": 0.432988047808765,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00025804375804361346,
      "calib/mean_conf": 0.9907569721115538,
      "calib/mu_c": 0.9906428571428572,
      "calib/mu_w": 0.9909009009009008,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.432988047808765,
      "calib/std_conf": 0.0029309281526294084,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9100867302688638,
      "calib/step_q_c_n": 1153.0,
      "calib/step_q_gap": -0.0025051641703445204,
      "calib/step_q_w": 0.9125918944392083,
      "calib/step_q_w_n": 1061.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2605.0,
      "completions/max_terminated_length": 2605.0,
      "completions/mean_length": 735.09765625,
      "completions/mean_terminated_length": 746.7659301757812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 334.0,
      "epoch": 0.1312,
      "grad_norm": 0.1874084323644638,
      "learning_rate": 2.166666666666667e-06,
      "loss": -0.0302,
      "num_tokens": 31987091.0,
      "reward": 1.03515625,
      "reward_std": 0.3761196434497833,
      "rewards/accuracy_reward_step": 0.546875,
      "rewards/format_reward_step": 0.9765625,
      "step": 123
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 7.87109375,
      "calib/ece": 0.2989723320158103,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00014065934065921049,
      "calib/mean_conf": 0.9906719367588933,
      "calib/mu_c": 0.9906285714285712,
      "calib/mu_w": 0.9907692307692304,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.2989723320158103,
      "calib/std_conf": 0.002801588165481323,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9077858176555716,
      "calib/step_q_c_n": 1382.0,
      "calib/step_q_gap": -0.0005396167836069221,
      "calib/step_q_w": 0.9083254344391786,
      "calib/step_q_w_n": 633.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2169.0,
      "completions/max_terminated_length": 2169.0,
      "completions/mean_length": 693.8984375,
      "completions/mean_terminated_length": 702.1265258789062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 321.0,
      "epoch": 0.13226666666666667,
      "grad_norm": 0.13662093877792358,
      "learning_rate": 2.138888888888889e-06,
      "loss": -0.0311,
      "num_tokens": 32271545.0,
      "reward": 1.177734375,
      "reward_std": 0.21031202375888824,
      "rewards/accuracy_reward_step": 0.68359375,
      "rewards/format_reward_step": 0.98828125,
      "step": 124
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 8.41015625,
      "calib/ece": 0.4788709677419356,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0005986854948915044,
      "calib/mean_conf": 0.990967741935484,
      "calib/mu_c": 0.9912598425196849,
      "calib/mu_w": 0.9906611570247934,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.4788709677419356,
      "calib/std_conf": 0.003217731829558068,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9117524752475249,
      "calib/step_q_c_n": 1010.0,
      "calib/step_q_gap": -0.0020358012179169105,
      "calib/step_q_w": 0.9137882764654418,
      "calib/step_q_w_n": 1143.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2538.0,
      "completions/max_terminated_length": 2538.0,
      "completions/mean_length": 723.49609375,
      "completions/mean_terminated_length": 734.980224609375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 390.0,
      "epoch": 0.13333333333333333,
      "grad_norm": 0.16686882078647614,
      "learning_rate": 2.1111111111111114e-06,
      "loss": -0.0043,
      "num_tokens": 32561568.0,
      "reward": 0.978515625,
      "reward_std": 0.2920733094215393,
      "rewards/accuracy_reward_step": 0.49609375,
      "rewards/format_reward_step": 0.96484375,
      "step": 125
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 9.3515625,
      "calib/ece": 0.497404858299595,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -4.052459016390397e-05,
      "calib/mean_conf": 0.9913319838056679,
      "calib/mu_c": 0.991311475409836,
      "calib/mu_w": 0.9913519999999999,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.497404858299595,
      "calib/std_conf": 0.004045679120530505,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9121834862385321,
      "calib/step_q_c_n": 1090.0,
      "calib/step_q_gap": 0.011021676422581117,
      "calib/step_q_w": 0.901161809815951,
      "calib/step_q_w_n": 1304.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 2704.0,
      "completions/max_terminated_length": 2704.0,
      "completions/mean_length": 744.265625,
      "completions/mean_terminated_length": 759.0916748046875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 319.0,
      "epoch": 0.1344,
      "grad_norm": 0.1326979398727417,
      "learning_rate": 2.0833333333333334e-06,
      "loss": -0.0206,
      "num_tokens": 32857564.0,
      "reward": 0.958984375,
      "reward_std": 0.2162768691778183,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/format_reward_step": 0.96484375,
      "step": 126
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 8.7265625,
      "calib/ece": 0.4245119047619048,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.001260601783537596,
      "calib/mean_conf": 0.9919722222222223,
      "calib/mu_c": 0.9925174825174824,
      "calib/mu_w": 0.9912568807339448,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4245119047619048,
      "calib/std_conf": 0.003968529752321603,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9125038520801233,
      "calib/step_q_c_n": 1298.0,
      "calib/step_q_gap": 0.011049792251063506,
      "calib/step_q_w": 0.9014540598290598,
      "calib/step_q_w_n": 936.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2925.0,
      "completions/max_terminated_length": 2925.0,
      "completions/mean_length": 705.546875,
      "completions/mean_terminated_length": 708.3137817382812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 384.0,
      "epoch": 0.13546666666666668,
      "grad_norm": 0.15588776767253876,
      "learning_rate": 2.0555555555555555e-06,
      "loss": 0.0308,
      "num_tokens": 33141856.0,
      "reward": 1.05078125,
      "reward_std": 0.22860543429851532,
      "rewards/accuracy_reward_step": 0.55859375,
      "rewards/format_reward_step": 0.984375,
      "step": 127
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 8.375,
      "calib/ece": 0.43579600000000007,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.996,
      "calib/gap": 0.009125129399585874,
      "calib/mean_conf": 0.987796,
      "calib/mu_c": 0.9918840579710144,
      "calib/mu_w": 0.9827589285714285,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.43579600000000007,
      "calib/std_conf": 0.06272691913365425,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9094268077601411,
      "calib/step_q_c_n": 1134.0,
      "calib/step_q_gap": 0.05834759983934912,
      "calib/step_q_w": 0.851079207920792,
      "calib/step_q_w_n": 1010.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2888.0,
      "completions/max_terminated_length": 2888.0,
      "completions/mean_length": 710.8984375,
      "completions/mean_terminated_length": 722.1825561523438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 326.0,
      "epoch": 0.13653333333333334,
      "grad_norm": 0.16684812307357788,
      "learning_rate": 2.027777777777778e-06,
      "loss": -0.0061,
      "num_tokens": 33430510.0,
      "reward": 1.02734375,
      "reward_std": 0.29256534576416016,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/format_reward_step": 0.9765625,
      "step": 128
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 8.6875,
      "calib/ece": 0.42376640419947514,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.9960629921259843,
      "calib/gap": -0.00461604977327823,
      "calib/mean_conf": 0.9893832020997375,
      "calib/mu_c": 0.9874022988505748,
      "calib/mu_w": 0.9920183486238531,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.42114173228346463,
      "calib/std_conf": 0.04155039059293514,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9116179952644041,
      "calib/step_q_c_n": 1267.0,
      "calib/step_q_gap": 0.00417807885897048,
      "calib/step_q_w": 0.9074399164054336,
      "calib/step_q_w_n": 957.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2578.0,
      "completions/max_terminated_length": 2578.0,
      "completions/mean_length": 669.89453125,
      "completions/mean_terminated_length": 669.89453125,
      "completions/min_length": 317.0,
      "completions/min_terminated_length": 317.0,
      "epoch": 0.1376,
      "grad_norm": 0.1644672006368637,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0153,
      "num_tokens": 33704387.0,
      "reward": 1.060546875,
      "reward_std": 0.23660747706890106,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/format_reward_step": 0.98828125,
      "step": 129
    },
    {
      "calib/answer_extract_rate": 0.99609375,
      "calib/avg_num_step_conf": 8.1875,
      "calib/ece": 0.38050980392156863,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0002505827505827618,
      "calib/mean_conf": 0.9922745098039216,
      "calib/mu_c": 0.9923717948717948,
      "calib/mu_w": 0.9921212121212121,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.38050980392156863,
      "calib/std_conf": 0.00446370298827677,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9108261211644375,
      "calib/step_q_c_n": 1271.0,
      "calib/step_q_gap": 0.0011836969220132554,
      "calib/step_q_w": 0.9096424242424243,
      "calib/step_q_w_n": 825.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2649.0,
      "completions/max_terminated_length": 2649.0,
      "completions/mean_length": 656.65234375,
      "completions/mean_terminated_length": 656.65234375,
      "completions/min_length": 336.0,
      "completions/min_terminated_length": 336.0,
      "epoch": 0.13866666666666666,
      "grad_norm": 0.13385489583015442,
      "learning_rate": 1.9722222222222224e-06,
      "loss": 0.0115,
      "num_tokens": 33977778.0,
      "reward": 1.107421875,
      "reward_std": 0.15861839056015015,
      "rewards/accuracy_reward_step": 0.609375,
      "rewards/format_reward_step": 0.99609375,
      "step": 130
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 8.703125,
      "calib/ece": 0.5781274900398408,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 5.1020408163293673e-05,
      "calib/mean_conf": 0.9924701195219124,
      "calib/mu_c": 0.9924999999999999,
      "calib/mu_w": 0.9924489795918366,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.5781274900398408,
      "calib/std_conf": 0.004835347067234673,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9095091324200913,
      "calib/step_q_c_n": 876.0,
      "calib/step_q_gap": 0.004057948988138604,
      "calib/step_q_w": 0.9054511834319527,
      "calib/step_q_w_n": 1352.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2442.0,
      "completions/max_terminated_length": 2442.0,
      "completions/mean_length": 670.1796875,
      "completions/mean_terminated_length": 680.8175048828125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 405.0,
      "epoch": 0.13973333333333332,
      "grad_norm": 0.14848309755325317,
      "learning_rate": 1.944444444444445e-06,
      "loss": -0.0111,
      "num_tokens": 34255552.0,
      "reward": 0.89453125,
      "reward_std": 0.17193031311035156,
      "rewards/accuracy_reward_step": 0.40625,
      "rewards/format_reward_step": 0.9765625,
      "step": 131
    },
    {
      "calib/answer_extract_rate": 0.99609375,
      "calib/avg_num_step_conf": 8.59375,
      "calib/ece": 0.35124409448818905,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0014529764713813176,
      "calib/mean_conf": 0.992976377952756,
      "calib/mu_c": 0.9934969325153374,
      "calib/mu_w": 0.9920439560439561,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.35124409448818905,
      "calib/std_conf": 0.0052817686236865434,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9092824773413897,
      "calib/step_q_c_n": 1324.0,
      "calib/step_q_gap": 0.03022197505828461,
      "calib/step_q_w": 0.8790605022831051,
      "calib/step_q_w_n": 876.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2436.0,
      "completions/max_terminated_length": 2436.0,
      "completions/mean_length": 699.015625,
      "completions/mean_terminated_length": 701.7568969726562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 326.0,
      "epoch": 0.1408,
      "grad_norm": 0.1599433869123459,
      "learning_rate": 1.916666666666667e-06,
      "loss": 0.0201,
      "num_tokens": 34540092.0,
      "reward": 1.1328125,
      "reward_std": 0.24816462397575378,
      "rewards/accuracy_reward_step": 0.63671875,
      "rewards/format_reward_step": 0.9921875,
      "step": 132
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 8.265625,
      "calib/ece": 0.5425099601593626,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.9960159362549801,
      "calib/gap": 0.0010279594715916218,
      "calib/mean_conf": 0.9927091633466136,
      "calib/mu_c": 0.9932743362831858,
      "calib/mu_w": 0.9922463768115942,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5425099601593626,
      "calib/std_conf": 0.007510356441719955,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9057649667405765,
      "calib/step_q_c_n": 902.0,
      "calib/step_q_gap": -0.0019483775757331845,
      "calib/step_q_w": 0.9077133443163097,
      "calib/step_q_w_n": 1214.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2653.0,
      "completions/max_terminated_length": 2653.0,
      "completions/mean_length": 755.53125,
      "completions/mean_terminated_length": 767.5238647460938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 318.0,
      "epoch": 0.14186666666666667,
      "grad_norm": 0.17682072520256042,
      "learning_rate": 1.888888888888889e-06,
      "loss": 0.0028,
      "num_tokens": 34839852.0,
      "reward": 0.931640625,
      "reward_std": 0.30837512016296387,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/format_reward_step": 0.98046875,
      "step": 133
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 8.7265625,
      "calib/ece": 0.4609274193548387,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00016980146290501708,
      "calib/mean_conf": 0.9931854838709677,
      "calib/mu_c": 0.9931060606060604,
      "calib/mu_w": 0.9932758620689655,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4609274193548387,
      "calib/std_conf": 0.005229980163232401,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9068285714285715,
      "calib/step_q_c_n": 1050.0,
      "calib/step_q_gap": -0.003467036679536517,
      "calib/step_q_w": 0.9102956081081081,
      "calib/step_q_w_n": 1184.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 2262.0,
      "completions/max_terminated_length": 2262.0,
      "completions/mean_length": 744.5546875,
      "completions/mean_terminated_length": 762.4240112304688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 390.0,
      "epoch": 0.14293333333333333,
      "grad_norm": 0.17197610437870026,
      "learning_rate": 1.8611111111111113e-06,
      "loss": -0.028,
      "num_tokens": 35139410.0,
      "reward": 1.0,
      "reward_std": 0.29847827553749084,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/format_reward_step": 0.96875,
      "step": 134
    },
    {
      "calib/answer_extract_rate": 0.9921875,
      "calib/avg_num_step_conf": 8.828125,
      "calib/ece": 0.38011811023622055,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0005075876504448962,
      "calib/mean_conf": 0.9942913385826773,
      "calib/mu_c": 0.9944871794871796,
      "calib/mu_w": 0.9939795918367347,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.38011811023622055,
      "calib/std_conf": 0.004949525128292414,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9111734317343174,
      "calib/step_q_c_n": 1355.0,
      "calib/step_q_gap": 0.001626470408350622,
      "calib/step_q_w": 0.9095469613259668,
      "calib/step_q_w_n": 905.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2160.0,
      "completions/max_terminated_length": 2160.0,
      "completions/mean_length": 747.48828125,
      "completions/mean_terminated_length": 750.419677734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 409.0,
      "epoch": 0.144,
      "grad_norm": 0.18103228509426117,
      "learning_rate": 1.8333333333333333e-06,
      "loss": 0.0169,
      "num_tokens": 35436647.0,
      "reward": 1.103515625,
      "reward_std": 0.2690998911857605,
      "rewards/accuracy_reward_step": 0.609375,
      "rewards/format_reward_step": 0.98828125,
      "step": 135
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 9.12890625,
      "calib/ece": 0.44757936507936513,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0014836003051106328,
      "calib/mean_conf": 0.9951984126984128,
      "calib/mu_c": 0.9958695652173913,
      "calib/mu_w": 0.9943859649122807,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.44757936507936513,
      "calib/std_conf": 0.004996061689081574,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9105586152635721,
      "calib/step_q_c_n": 1271.0,
      "calib/step_q_gap": 0.016046420141620832,
      "calib/step_q_w": 0.8945121951219512,
      "calib/step_q_w_n": 1066.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2100.0,
      "completions/max_terminated_length": 2100.0,
      "completions/mean_length": 711.12109375,
      "completions/mean_terminated_length": 722.4087524414062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 383.0,
      "epoch": 0.14506666666666668,
      "grad_norm": 0.17702247202396393,
      "learning_rate": 1.8055555555555557e-06,
      "loss": -0.024,
      "num_tokens": 35727182.0,
      "reward": 1.029296875,
      "reward_std": 0.28243717551231384,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/format_reward_step": 0.98046875,
      "step": 136
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 9.44140625,
      "calib/ece": 0.4384196547144754,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.9960159362549801,
      "calib/gap": 0.006667951353203083,
      "calib/mean_conf": 0.9922045152722444,
      "calib/mu_c": 0.9951798561151078,
      "calib/mu_w": 0.9885119047619048,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4384196547144754,
      "calib/std_conf": 0.04202500106425627,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9080967741935484,
      "calib/step_q_c_n": 1240.0,
      "calib/step_q_gap": 0.01070510044673445,
      "calib/step_q_w": 0.897391673746814,
      "calib/step_q_w_n": 1177.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2550.0,
      "completions/max_terminated_length": 2550.0,
      "completions/mean_length": 715.8984375,
      "completions/mean_terminated_length": 724.3873901367188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 388.0,
      "epoch": 0.14613333333333334,
      "grad_norm": 0.16152194142341614,
      "learning_rate": 1.777777777777778e-06,
      "loss": 0.0062,
      "num_tokens": 36017436.0,
      "reward": 1.033203125,
      "reward_std": 0.22543761134147644,
      "rewards/accuracy_reward_step": 0.54296875,
      "rewards/format_reward_step": 0.98046875,
      "step": 137
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 9.359375,
      "calib/ece": 0.3457661290322581,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00015063896623102035,
      "calib/mean_conf": 0.9949596774193549,
      "calib/mu_c": 0.9949068322981367,
      "calib/mu_w": 0.9950574712643677,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3457661290322581,
      "calib/std_conf": 0.0049998374063053444,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9099426111908178,
      "calib/step_q_c_n": 1394.0,
      "calib/step_q_gap": 0.03896456727864206,
      "calib/step_q_w": 0.8709780439121757,
      "calib/step_q_w_n": 1002.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 2179.0,
      "completions/max_terminated_length": 2179.0,
      "completions/mean_length": 723.10546875,
      "completions/mean_terminated_length": 737.510009765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 316.0,
      "epoch": 0.1472,
      "grad_norm": 0.18175871670246124,
      "learning_rate": 1.75e-06,
      "loss": -0.0157,
      "num_tokens": 36306887.0,
      "reward": 1.11328125,
      "reward_std": 0.29471495747566223,
      "rewards/accuracy_reward_step": 0.62890625,
      "rewards/format_reward_step": 0.96875,
      "step": 138
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 8.90234375,
      "calib/ece": 0.3306048387096775,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0007601314348301136,
      "calib/mean_conf": 0.9959274193548387,
      "calib/mu_c": 0.9961818181818182,
      "calib/mu_w": 0.995421686746988,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3306048387096775,
      "calib/std_conf": 0.00499463161335987,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9098336948662328,
      "calib/step_q_c_n": 1383.0,
      "calib/step_q_gap": 0.009576998437661355,
      "calib/step_q_w": 0.9002566964285714,
      "calib/step_q_w_n": 896.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2653.0,
      "completions/max_terminated_length": 2653.0,
      "completions/mean_length": 657.640625,
      "completions/mean_terminated_length": 668.0794067382812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 343.0,
      "epoch": 0.14826666666666666,
      "grad_norm": 0.17501947283744812,
      "learning_rate": 1.7222222222222224e-06,
      "loss": -0.0145,
      "num_tokens": 36578339.0,
      "reward": 1.12890625,
      "reward_std": 0.24773260951042175,
      "rewards/accuracy_reward_step": 0.64453125,
      "rewards/format_reward_step": 0.96875,
      "step": 139
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 9.14453125,
      "calib/ece": 0.29529880478087644,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0004363636363635681,
      "calib/mean_conf": 0.9964940239043825,
      "calib/mu_c": 0.9963636363636365,
      "calib/mu_w": 0.9968,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.29529880478087644,
      "calib/std_conf": 0.004771571289746575,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9106408544726302,
      "calib/step_q_c_n": 1498.0,
      "calib/step_q_gap": -7.326177885258467e-05,
      "calib/step_q_w": 0.9107141162514828,
      "calib/step_q_w_n": 843.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2807.0,
      "completions/max_terminated_length": 2807.0,
      "completions/mean_length": 686.55078125,
      "completions/mean_terminated_length": 697.4484252929688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 355.0,
      "epoch": 0.14933333333333335,
      "grad_norm": 0.16100062429904938,
      "learning_rate": 1.6944444444444446e-06,
      "loss": -0.0091,
      "num_tokens": 36859112.0,
      "reward": 1.177734375,
      "reward_std": 0.2540324330329895,
      "rewards/accuracy_reward_step": 0.6875,
      "rewards/format_reward_step": 0.98046875,
      "step": 140
    },
    {
      "calib/answer_extract_rate": 0.9921875,
      "calib/avg_num_step_conf": 8.359375,
      "calib/ece": 0.39397637795275586,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0008250825082509738,
      "calib/mean_conf": 0.9963385826771654,
      "calib/mu_c": 0.9966666666666667,
      "calib/mu_w": 0.9958415841584157,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.39397637795275586,
      "calib/std_conf": 0.004898530186820943,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9082588235294118,
      "calib/step_q_c_n": 1275.0,
      "calib/step_q_gap": -0.00031921115266919653,
      "calib/step_q_w": 0.908578034682081,
      "calib/step_q_w_n": 865.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2985.0,
      "completions/max_terminated_length": 2985.0,
      "completions/mean_length": 741.88671875,
      "completions/mean_terminated_length": 744.796142578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 344.0,
      "epoch": 0.1504,
      "grad_norm": 0.15197478234767914,
      "learning_rate": 1.6666666666666667e-06,
      "loss": 0.0091,
      "num_tokens": 37156131.0,
      "reward": 1.09375,
      "reward_std": 0.22331982851028442,
      "rewards/accuracy_reward_step": 0.59765625,
      "rewards/format_reward_step": 0.9921875,
      "step": 141
    },
    {
      "calib/answer_extract_rate": 0.9921875,
      "calib/avg_num_step_conf": 8.453125,
      "calib/ece": 0.4255204216073781,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.9960474308300395,
      "calib/gap": -0.004131545338441844,
      "calib/mean_conf": 0.993372859025033,
      "calib/mu_c": 0.9916091954022991,
      "calib/mu_w": 0.9957407407407409,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.42288537549407107,
      "calib/std_conf": 0.041865252417557876,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.911143818334735,
      "calib/step_q_c_n": 1189.0,
      "calib/step_q_gap": 0.00012843371935034043,
      "calib/step_q_w": 0.9110153846153847,
      "calib/step_q_w_n": 975.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 1525.0,
      "completions/max_terminated_length": 1525.0,
      "completions/mean_length": 720.359375,
      "completions/mean_terminated_length": 726.031494140625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 337.0,
      "epoch": 0.15146666666666667,
      "grad_norm": 0.15058152377605438,
      "learning_rate": 1.638888888888889e-06,
      "loss": -0.0109,
      "num_tokens": 37445703.0,
      "reward": 1.064453125,
      "reward_std": 0.20312634110450745,
      "rewards/accuracy_reward_step": 0.5703125,
      "rewards/format_reward_step": 0.98828125,
      "step": 142
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 8.65234375,
      "calib/ece": 0.4366532258064517,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0004573955514487471,
      "calib/mean_conf": 0.9971370967741936,
      "calib/mu_c": 0.9973381294964029,
      "calib/mu_w": 0.9968807339449541,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4366532258064517,
      "calib/std_conf": 0.004520267401131438,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9136747967479675,
      "calib/step_q_c_n": 1230.0,
      "calib/step_q_gap": 0.008420989641368504,
      "calib/step_q_w": 0.905253807106599,
      "calib/step_q_w_n": 985.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2607.0,
      "completions/max_terminated_length": 2607.0,
      "completions/mean_length": 722.49609375,
      "completions/mean_terminated_length": 733.96435546875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 348.0,
      "epoch": 0.15253333333333333,
      "grad_norm": 0.19313447177410126,
      "learning_rate": 1.6111111111111113e-06,
      "loss": 0.0046,
      "num_tokens": 37737998.0,
      "reward": 1.02734375,
      "reward_std": 0.2798112630844116,
      "rewards/accuracy_reward_step": 0.54296875,
      "rewards/format_reward_step": 0.96875,
      "step": 143
    },
    {
      "calib/answer_extract_rate": 0.99609375,
      "calib/avg_num_step_conf": 8.4140625,
      "calib/ece": 0.27803921568627454,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00039845173041908044,
      "calib/mean_conf": 0.995686274509804,
      "calib/mu_c": 0.9955737704918033,
      "calib/mu_w": 0.9959722222222224,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.27803921568627454,
      "calib/std_conf": 0.004952678800123569,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9118355263157897,
      "calib/step_q_c_n": 1520.0,
      "calib/step_q_gap": 8.47376722564741e-05,
      "calib/step_q_w": 0.9117507886435332,
      "calib/step_q_w_n": 634.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2520.0,
      "completions/max_terminated_length": 2520.0,
      "completions/mean_length": 728.23828125,
      "completions/mean_terminated_length": 728.23828125,
      "completions/min_length": 399.0,
      "completions/min_terminated_length": 399.0,
      "epoch": 0.1536,
      "grad_norm": 0.13797351717948914,
      "learning_rate": 1.5833333333333333e-06,
      "loss": 0.0165,
      "num_tokens": 38028555.0,
      "reward": 1.212890625,
      "reward_std": 0.14506946504116058,
      "rewards/accuracy_reward_step": 0.71484375,
      "rewards/format_reward_step": 0.99609375,
      "step": 144
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 9.28125,
      "calib/ece": 0.3081600000000001,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 8.944543828381946e-06,
      "calib/mean_conf": 0.99616,
      "calib/mu_c": 0.9961627906976744,
      "calib/mu_w": 0.9961538461538461,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.3081600000000001,
      "calib/std_conf": 0.004863578929142616,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.913320184089415,
      "calib/step_q_c_n": 1521.0,
      "calib/step_q_gap": 0.001121353680058279,
      "calib/step_q_w": 0.9121988304093567,
      "calib/step_q_w_n": 855.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 1635.0,
      "completions/max_terminated_length": 1635.0,
      "completions/mean_length": 718.89453125,
      "completions/mean_terminated_length": 736.1480102539062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 354.0,
      "epoch": 0.15466666666666667,
      "grad_norm": 0.18962031602859497,
      "learning_rate": 1.5555555555555558e-06,
      "loss": -0.0321,
      "num_tokens": 38315296.0,
      "reward": 1.158203125,
      "reward_std": 0.3293054699897766,
      "rewards/accuracy_reward_step": 0.671875,
      "rewards/format_reward_step": 0.97265625,
      "step": 145
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 8.82421875,
      "calib/ece": 0.5,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0008960573476700651,
      "calib/mean_conf": 0.996,
      "calib/mu_c": 0.9964516129032257,
      "calib/mu_w": 0.9955555555555556,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5,
      "calib/std_conf": 0.00489897948556636,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9110076045627378,
      "calib/step_q_c_n": 1052.0,
      "calib/step_q_gap": 0.0008750444964575887,
      "calib/step_q_w": 0.9101325600662802,
      "calib/step_q_w_n": 1207.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2443.0,
      "completions/max_terminated_length": 2443.0,
      "completions/mean_length": 741.7421875,
      "completions/mean_terminated_length": 753.5159301757812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 366.0,
      "epoch": 0.15573333333333333,
      "grad_norm": 0.20323771238327026,
      "learning_rate": 1.527777777777778e-06,
      "loss": -0.0212,
      "num_tokens": 38612398.0,
      "reward": 0.97265625,
      "reward_std": 0.27729031443595886,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/format_reward_step": 0.9765625,
      "step": 146
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 8.69140625,
      "calib/ece": 0.5288047808764942,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.000288302079347047,
      "calib/mean_conf": 0.9949402390438248,
      "calib/mu_c": 0.9947863247863247,
      "calib/mu_w": 0.9950746268656717,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5288047808764942,
      "calib/std_conf": 0.004999642850056097,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9134854368932038,
      "calib/step_q_c_n": 1030.0,
      "calib/step_q_gap": -0.00045598570093852153,
      "calib/step_q_w": 0.9139414225941423,
      "calib/step_q_w_n": 1195.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2852.0,
      "completions/max_terminated_length": 2852.0,
      "completions/mean_length": 753.86328125,
      "completions/mean_terminated_length": 759.7991943359375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 426.0,
      "epoch": 0.1568,
      "grad_norm": 0.1743156462907791,
      "learning_rate": 1.5e-06,
      "loss": 0.0223,
      "num_tokens": 38909067.0,
      "reward": 0.947265625,
      "reward_std": 0.20046457648277283,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/format_reward_step": 0.98046875,
      "step": 147
    },
    {
      "calib/answer_extract_rate": 0.98046875,
      "calib/avg_num_step_conf": 8.46484375,
      "calib/ece": 0.33278884462151404,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.000672572643515057,
      "calib/mean_conf": 0.9941434262948208,
      "calib/mu_c": 0.9939156626506025,
      "calib/mu_w": 0.9945882352941176,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.33278884462151404,
      "calib/std_conf": 0.004926081758111154,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9119335219236209,
      "calib/step_q_c_n": 1414.0,
      "calib/step_q_gap": 0.010459418337963378,
      "calib/step_q_w": 0.9014741035856575,
      "calib/step_q_w_n": 753.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2852.0,
      "completions/max_terminated_length": 2852.0,
      "completions/mean_length": 730.67578125,
      "completions/mean_terminated_length": 736.4291381835938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 406.0,
      "epoch": 0.15786666666666666,
      "grad_norm": 0.1720433533191681,
      "learning_rate": 1.4722222222222225e-06,
      "loss": 0.0099,
      "num_tokens": 39201232.0,
      "reward": 1.138671875,
      "reward_std": 0.21238605678081512,
      "rewards/accuracy_reward_step": 0.6484375,
      "rewards/format_reward_step": 0.98046875,
      "step": 148
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 8.140625,
      "calib/ece": 0.4248387096774193,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0011904288460262702,
      "calib/mean_conf": 0.9933870967741936,
      "calib/mu_c": 0.9939007092198581,
      "calib/mu_w": 0.9927102803738318,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4248387096774193,
      "calib/std_conf": 0.004732709919716523,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9122546419098143,
      "calib/step_q_c_n": 1131.0,
      "calib/step_q_gap": 0.002338587345281251,
      "calib/step_q_w": 0.9099160545645331,
      "calib/step_q_w_n": 953.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2728.0,
      "completions/max_terminated_length": 2728.0,
      "completions/mean_length": 774.7109375,
      "completions/mean_terminated_length": 783.8972778320312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 320.0,
      "epoch": 0.15893333333333334,
      "grad_norm": 0.18674218654632568,
      "learning_rate": 1.4444444444444445e-06,
      "loss": 0.0311,
      "num_tokens": 39504014.0,
      "reward": 1.03515625,
      "reward_std": 0.30080240964889526,
      "rewards/accuracy_reward_step": 0.55078125,
      "rewards/format_reward_step": 0.96875,
      "step": 149
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 9.09375,
      "calib/ece": 0.35622132253711214,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.9959514170040485,
      "calib/gap": -0.0036106765277578168,
      "calib/mean_conf": 0.9904993252361675,
      "calib/mu_c": 0.9891983122362871,
      "calib/mu_w": 0.9928089887640449,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.35352226720647784,
      "calib/std_conf": 0.042155943753956086,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9152635486265775,
      "calib/step_q_c_n": 1347.0,
      "calib/step_q_gap": 0.003714109278973088,
      "calib/step_q_w": 0.9115494393476045,
      "calib/step_q_w_n": 981.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 2365.0,
      "completions/max_terminated_length": 2365.0,
      "completions/mean_length": 683.05078125,
      "completions/mean_terminated_length": 705.0846557617188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 318.0,
      "epoch": 0.16,
      "grad_norm": 0.1734960377216339,
      "learning_rate": 1.4166666666666667e-06,
      "loss": -0.0502,
      "num_tokens": 39783835.0,
      "reward": 1.099609375,
      "reward_std": 0.20597529411315918,
      "rewards/accuracy_reward_step": 0.6171875,
      "rewards/format_reward_step": 0.96484375,
      "step": 150
    },
    {
      "calib/answer_extract_rate": 0.95703125,
      "calib/avg_num_step_conf": 8.23046875,
      "calib/ece": 0.5041463414634146,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0007460317460317167,
      "calib/mean_conf": 0.9919512195121951,
      "calib/mu_c": 0.9923333333333334,
      "calib/mu_w": 0.9915873015873017,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5041463414634146,
      "calib/std_conf": 0.003962945563237058,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9136016096579478,
      "calib/step_q_c_n": 994.0,
      "calib/step_q_gap": 0.0042485099274894855,
      "calib/step_q_w": 0.9093530997304583,
      "calib/step_q_w_n": 1113.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 2807.0,
      "completions/max_terminated_length": 2807.0,
      "completions/mean_length": 829.67578125,
      "completions/mean_terminated_length": 839.5138549804688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 424.0,
      "epoch": 0.16106666666666666,
      "grad_norm": 0.18515925109386444,
      "learning_rate": 1.3888888888888892e-06,
      "loss": 0.0253,
      "num_tokens": 40103256.0,
      "reward": 0.947265625,
      "reward_std": 0.31336262822151184,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/format_reward_step": 0.95703125,
      "step": 151
    },
    {
      "calib/answer_extract_rate": 0.94140625,
      "calib/avg_num_step_conf": 9.5625,
      "calib/ece": 0.46032365145228227,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -8.0475663716717e-05,
      "calib/mean_conf": 0.9914439834024897,
      "calib/mu_c": 0.99140625,
      "calib/mu_w": 0.9914867256637168,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.46032365145228227,
      "calib/std_conf": 0.0035054751342345866,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9150719424460431,
      "calib/step_q_c_n": 1112.0,
      "calib/step_q_gap": 0.00028825981131241907,
      "calib/step_q_w": 0.9147836826347306,
      "calib/step_q_w_n": 1336.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 2681.0,
      "completions/max_terminated_length": 2681.0,
      "completions/mean_length": 802.953125,
      "completions/mean_terminated_length": 825.5260620117188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 436.0,
      "epoch": 0.16213333333333332,
      "grad_norm": 0.18708793818950653,
      "learning_rate": 1.3611111111111112e-06,
      "loss": -0.0153,
      "num_tokens": 40414204.0,
      "reward": 0.970703125,
      "reward_std": 0.2786449193954468,
      "rewards/accuracy_reward_step": 0.5,
      "rewards/format_reward_step": 0.94140625,
      "step": 152
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 8.43359375,
      "calib/ece": 0.40223577235772345,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00036394673950157497,
      "calib/mean_conf": 0.9916666666666666,
      "calib/mu_c": 0.9915172413793102,
      "calib/mu_w": 0.9918811881188118,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.40223577235772345,
      "calib/std_conf": 0.003726779962499653,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9149527085124677,
      "calib/step_q_c_n": 1163.0,
      "calib/step_q_gap": -0.00010753245138783729,
      "calib/step_q_w": 0.9150602409638555,
      "calib/step_q_w_n": 996.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 3014.0,
      "completions/max_terminated_length": 3014.0,
      "completions/mean_length": 781.5625,
      "completions/mean_terminated_length": 803.5341186523438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 372.0,
      "epoch": 0.1632,
      "grad_norm": 0.16443714499473572,
      "learning_rate": 1.3333333333333334e-06,
      "loss": -0.0299,
      "num_tokens": 40721604.0,
      "reward": 1.05078125,
      "reward_std": 0.2820996046066284,
      "rewards/accuracy_reward_step": 0.5703125,
      "rewards/format_reward_step": 0.9609375,
      "step": 153
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 8.9375,
      "calib/ece": 0.48970281124498005,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0005599354838710102,
      "calib/mean_conf": 0.9917108433734941,
      "calib/mu_c": 0.9914319999999999,
      "calib/mu_w": 0.9919919354838709,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.48970281124498005,
      "calib/std_conf": 0.0037476529323794885,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.920658424908425,
      "calib/step_q_c_n": 1092.0,
      "calib/step_q_gap": -0.000581541646759054,
      "calib/step_q_w": 0.921239966555184,
      "calib/step_q_w_n": 1196.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 2099.0,
      "completions/max_terminated_length": 2099.0,
      "completions/mean_length": 731.42578125,
      "completions/mean_terminated_length": 751.9879150390625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 382.0,
      "epoch": 0.16426666666666667,
      "grad_norm": 0.16835734248161316,
      "learning_rate": 1.3055555555555556e-06,
      "loss": -0.0415,
      "num_tokens": 41013289.0,
      "reward": 0.974609375,
      "reward_std": 0.212571382522583,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/format_reward_step": 0.97265625,
      "step": 154
    },
    {
      "calib/answer_extract_rate": 0.9765625,
      "calib/avg_num_step_conf": 9.4140625,
      "calib/ece": 0.4033960000000001,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.000409286044515178,
      "calib/mean_conf": 0.991396,
      "calib/mu_c": 0.9915646258503401,
      "calib/mu_w": 0.9911553398058249,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4033960000000001,
      "calib/std_conf": 0.003460517880317919,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9184092606422704,
      "calib/step_q_c_n": 1339.0,
      "calib/step_q_gap": -0.000744800982379501,
      "calib/step_q_w": 0.9191540616246499,
      "calib/step_q_w_n": 1071.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2281.0,
      "completions/max_terminated_length": 2281.0,
      "completions/mean_length": 785.921875,
      "completions/mean_terminated_length": 798.3968505859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 436.0,
      "epoch": 0.16533333333333333,
      "grad_norm": 0.18231578171253204,
      "learning_rate": 1.2777777777777779e-06,
      "loss": -0.0423,
      "num_tokens": 41321701.0,
      "reward": 1.0625,
      "reward_std": 0.26381197571754456,
      "rewards/accuracy_reward_step": 0.57421875,
      "rewards/format_reward_step": 0.9765625,
      "step": 155
    },
    {
      "calib/answer_extract_rate": 0.9375,
      "calib/avg_num_step_conf": 11.0078125,
      "calib/ece": 0.4370333333333335,
      "calib/final_conf_rate": 0.9375,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00015852715901898407,
      "calib/mean_conf": 0.9912000000000002,
      "calib/mu_c": 0.9912706766917293,
      "calib/mu_w": 0.9911121495327103,
      "calib/nonempty_final_conf_rate": 0.9375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4370333333333335,
      "calib/std_conf": 0.0032380549717384386,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9186949286846275,
      "calib/step_q_c_n": 1262.0,
      "calib/step_q_gap": -0.00517396591691488,
      "calib/step_q_w": 0.9238688946015424,
      "calib/step_q_w_n": 1556.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.04296875,
      "completions/max_length": 3028.0,
      "completions/max_terminated_length": 3028.0,
      "completions/mean_length": 781.55078125,
      "completions/mean_terminated_length": 816.6408081054688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 383.0,
      "epoch": 0.1664,
      "grad_norm": 0.1721334308385849,
      "learning_rate": 1.25e-06,
      "loss": -0.0415,
      "num_tokens": 41626538.0,
      "reward": 0.98828125,
      "reward_std": 0.24574384093284607,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/format_reward_step": 0.9375,
      "step": 156
    },
    {
      "calib/answer_extract_rate": 0.95703125,
      "calib/avg_num_step_conf": 11.01953125,
      "calib/ece": 0.34683673469387755,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00010548523206788119,
      "calib/mean_conf": 0.991734693877551,
      "calib/mu_c": 0.9917721518987341,
      "calib/mu_w": 0.9916666666666663,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.34683673469387755,
      "calib/std_conf": 0.004374207847604431,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9220989974937344,
      "calib/step_q_c_n": 1596.0,
      "calib/step_q_gap": -0.006978553526673759,
      "calib/step_q_w": 0.9290775510204081,
      "calib/step_q_w_n": 1225.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 2527.0,
      "completions/max_terminated_length": 2527.0,
      "completions/mean_length": 792.1328125,
      "completions/mean_terminated_length": 824.3333129882812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 462.0,
      "epoch": 0.16746666666666668,
      "grad_norm": 0.1614348441362381,
      "learning_rate": 1.2222222222222223e-06,
      "loss": -0.0904,
      "num_tokens": 41933052.0,
      "reward": 1.095703125,
      "reward_std": 0.23183158040046692,
      "rewards/accuracy_reward_step": 0.6171875,
      "rewards/format_reward_step": 0.95703125,
      "step": 157
    },
    {
      "calib/answer_extract_rate": 0.94921875,
      "calib/avg_num_step_conf": 10.63671875,
      "calib/ece": 0.34127572016460905,
      "calib/final_conf_rate": 0.94921875,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00046909903201775016,
      "calib/mean_conf": 0.9914814814814815,
      "calib/mu_c": 0.9916455696202529,
      "calib/mu_w": 0.9911764705882351,
      "calib/nonempty_final_conf_rate": 0.94921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.34127572016460905,
      "calib/std_conf": 0.0035524677950464624,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9242605863192183,
      "calib/step_q_c_n": 1535.0,
      "calib/step_q_gap": -0.010730996172364171,
      "calib/step_q_w": 0.9349915824915824,
      "calib/step_q_w_n": 1188.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 2925.0,
      "completions/max_terminated_length": 2925.0,
      "completions/mean_length": 771.078125,
      "completions/mean_terminated_length": 802.4227294921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 465.0,
      "epoch": 0.16853333333333334,
      "grad_norm": 0.17561709880828857,
      "learning_rate": 1.1944444444444446e-06,
      "loss": -0.0555,
      "num_tokens": 42235688.0,
      "reward": 1.091796875,
      "reward_std": 0.28867921233177185,
      "rewards/accuracy_reward_step": 0.6171875,
      "rewards/format_reward_step": 0.94921875,
      "step": 158
    },
    {
      "calib/answer_extract_rate": 0.90234375,
      "calib/avg_num_step_conf": 13.3984375,
      "calib/ece": 0.3978744588744588,
      "calib/final_conf_rate": 0.90234375,
      "calib/format_rate": 0.90234375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0007016617487191068,
      "calib/mean_conf": 0.9909480519480519,
      "calib/mu_c": 0.9912335766423358,
      "calib/mu_w": 0.9905319148936167,
      "calib/nonempty_final_conf_rate": 0.90234375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3978744588744588,
      "calib/std_conf": 0.0029227993337472484,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9257992481203008,
      "calib/step_q_c_n": 1330.0,
      "calib/step_q_gap": -0.025595989974937394,
      "calib/step_q_w": 0.9513952380952382,
      "calib/step_q_w_n": 2100.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.08984375,
      "completions/max_length": 2909.0,
      "completions/max_terminated_length": 2909.0,
      "completions/mean_length": 742.56640625,
      "completions/mean_terminated_length": 815.866943359375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 503.0,
      "epoch": 0.1696,
      "grad_norm": 0.19220517575740814,
      "learning_rate": 1.1666666666666668e-06,
      "loss": -0.1328,
      "num_tokens": 42530569.0,
      "reward": 0.986328125,
      "reward_std": 0.4323299527168274,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/format_reward_step": 0.90234375,
      "step": 159
    },
    {
      "calib/answer_extract_rate": 0.90234375,
      "calib/avg_num_step_conf": 12.61328125,
      "calib/ece": 0.3415541125541126,
      "calib/final_conf_rate": 0.90234375,
      "calib/format_rate": 0.90234375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0006518518518516814,
      "calib/mean_conf": 0.990904761904762,
      "calib/mu_c": 0.9911333333333332,
      "calib/mu_w": 0.9904814814814815,
      "calib/nonempty_final_conf_rate": 0.90234375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3415541125541126,
      "calib/std_conf": 0.002861828769222511,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9232820855614974,
      "calib/step_q_c_n": 1496.0,
      "calib/step_q_gap": -0.017290909245196162,
      "calib/step_q_w": 0.9405729948066935,
      "calib/step_q_w_n": 1733.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0703125,
      "completions/max_length": 2659.0,
      "completions/max_terminated_length": 2659.0,
      "completions/mean_length": 787.7890625,
      "completions/mean_terminated_length": 847.3698120117188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 477.0,
      "epoch": 0.17066666666666666,
      "grad_norm": 0.1917860209941864,
      "learning_rate": 1.138888888888889e-06,
      "loss": -0.062,
      "num_tokens": 42837083.0,
      "reward": 1.037109375,
      "reward_std": 0.3205576539039612,
      "rewards/accuracy_reward_step": 0.5859375,
      "rewards/format_reward_step": 0.90234375,
      "step": 160
    },
    {
      "calib/answer_extract_rate": 0.8671875,
      "calib/avg_num_step_conf": 14.1953125,
      "calib/ece": 0.2570045045045045,
      "calib/final_conf_rate": 0.8671875,
      "calib/format_rate": 0.86328125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00030206925236542226,
      "calib/mean_conf": 0.9912387387387388,
      "calib/mu_c": 0.9913190184049079,
      "calib/mu_w": 0.9910169491525425,
      "calib/nonempty_final_conf_rate": 0.8671875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.2570045045045045,
      "calib/std_conf": 0.0032772398618231705,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9238732394366197,
      "calib/step_q_c_n": 1633.0,
      "calib/step_q_gap": -0.0283731373749746,
      "calib/step_q_w": 0.9522463768115943,
      "calib/step_q_w_n": 2001.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.11328125,
      "completions/max_length": 3000.0,
      "completions/max_terminated_length": 3000.0,
      "completions/mean_length": 737.0234375,
      "completions/mean_terminated_length": 831.1806030273438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 461.0,
      "epoch": 0.17173333333333332,
      "grad_norm": 0.18176619708538055,
      "learning_rate": 1.111111111111111e-06,
      "loss": -0.1118,
      "num_tokens": 43129681.0,
      "reward": 1.068359375,
      "reward_std": 0.368930459022522,
      "rewards/accuracy_reward_step": 0.63671875,
      "rewards/format_reward_step": 0.86328125,
      "step": 161
    },
    {
      "calib/answer_extract_rate": 0.89453125,
      "calib/avg_num_step_conf": 13.8125,
      "calib/ece": 0.3053580786026201,
      "calib/final_conf_rate": 0.89453125,
      "calib/format_rate": 0.89453125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0009972576079262385,
      "calib/mean_conf": 0.9909475982532752,
      "calib/mu_c": 0.9912611464968152,
      "calib/mu_w": 0.9902638888888889,
      "calib/nonempty_final_conf_rate": 0.89453125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3053580786026201,
      "calib/std_conf": 0.002908631319540183,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9286242955541641,
      "calib/step_q_c_n": 1597.0,
      "calib/step_q_gap": -0.021603141268940584,
      "calib/step_q_w": 0.9502274368231047,
      "calib/step_q_w_n": 1939.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.09375,
      "completions/max_length": 2772.0,
      "completions/max_terminated_length": 2772.0,
      "completions/mean_length": 747.8828125,
      "completions/mean_terminated_length": 825.25,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 465.0,
      "epoch": 0.1728,
      "grad_norm": 0.1692075878381729,
      "learning_rate": 1.0833333333333335e-06,
      "loss": -0.1072,
      "num_tokens": 43425283.0,
      "reward": 1.060546875,
      "reward_std": 0.302878737449646,
      "rewards/accuracy_reward_step": 0.61328125,
      "rewards/format_reward_step": 0.89453125,
      "step": 162
    },
    {
      "calib/answer_extract_rate": 0.89453125,
      "calib/avg_num_step_conf": 13.2734375,
      "calib/ece": 0.412695652173913,
      "calib/final_conf_rate": 0.8984375,
      "calib/format_rate": 0.89453125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0006635144562439077,
      "calib/mean_conf": 0.9909565217391304,
      "calib/mu_c": 0.9906766917293233,
      "calib/mu_w": 0.9913402061855672,
      "calib/nonempty_final_conf_rate": 0.8984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.412695652173913,
      "calib/std_conf": 0.0043686458042019065,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9277293233082707,
      "calib/step_q_c_n": 1330.0,
      "calib/step_q_gap": -0.019925415569872373,
      "calib/step_q_w": 0.9476547388781431,
      "calib/step_q_w_n": 2068.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.08203125,
      "completions/max_length": 3063.0,
      "completions/max_terminated_length": 3063.0,
      "completions/mean_length": 806.09765625,
      "completions/mean_terminated_length": 878.1318969726562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 427.0,
      "epoch": 0.17386666666666667,
      "grad_norm": 0.16379491984844208,
      "learning_rate": 1.0555555555555557e-06,
      "loss": -0.1066,
      "num_tokens": 43736476.0,
      "reward": 0.966796875,
      "reward_std": 0.30104950070381165,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/format_reward_step": 0.89453125,
      "step": 163
    },
    {
      "calib/answer_extract_rate": 0.91015625,
      "calib/avg_num_step_conf": 11.84765625,
      "calib/ece": 0.3810987124463521,
      "calib/final_conf_rate": 0.91015625,
      "calib/format_rate": 0.91015625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 5.7885776195587546e-05,
      "calib/mean_conf": 0.990540772532189,
      "calib/mu_c": 0.9905633802816899,
      "calib/mu_w": 0.9905054945054943,
      "calib/nonempty_final_conf_rate": 0.91015625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3810987124463521,
      "calib/std_conf": 0.0024122920742030173,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9231641791044776,
      "calib/step_q_c_n": 1340.0,
      "calib/step_q_gap": -0.014627905951635878,
      "calib/step_q_w": 0.9377920850561134,
      "calib/step_q_w_n": 1693.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 2666.0,
      "completions/max_terminated_length": 2666.0,
      "completions/mean_length": 821.84765625,
      "completions/mean_terminated_length": 891.4957885742188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 505.0,
      "epoch": 0.17493333333333333,
      "grad_norm": 0.16550104320049286,
      "learning_rate": 1.0277777777777777e-06,
      "loss": -0.091,
      "num_tokens": 44053005.0,
      "reward": 1.009765625,
      "reward_std": 0.2756240665912628,
      "rewards/accuracy_reward_step": 0.5546875,
      "rewards/format_reward_step": 0.91015625,
      "step": 164
    },
    {
      "calib/answer_extract_rate": 0.94140625,
      "calib/avg_num_step_conf": 10.8671875,
      "calib/ece": 0.4410247933884298,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00012788852866119083,
      "calib/mean_conf": 0.9906115702479339,
      "calib/mu_c": 0.9906691729323307,
      "calib/mu_w": 0.9905412844036695,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4410247933884298,
      "calib/std_conf": 0.0023806100365015024,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9178924103419517,
      "calib/step_q_c_n": 1199.0,
      "calib/step_q_gap": -0.019873856240486876,
      "calib/step_q_w": 0.9377662665824386,
      "calib/step_q_w_n": 1583.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 2114.0,
      "completions/max_terminated_length": 2114.0,
      "completions/mean_length": 832.63671875,
      "completions/mean_terminated_length": 873.5859985351562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 484.0,
      "epoch": 0.176,
      "grad_norm": 0.1745368391275406,
      "learning_rate": 1.0000000000000002e-06,
      "loss": -0.0757,
      "num_tokens": 44371736.0,
      "reward": 0.990234375,
      "reward_std": 0.3081328272819519,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/format_reward_step": 0.94140625,
      "step": 165
    },
    {
      "calib/answer_extract_rate": 0.875,
      "calib/avg_num_step_conf": 10.984375,
      "calib/ece": 0.28958035714285724,
      "calib/final_conf_rate": 0.875,
      "calib/format_rate": 0.875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00010019963874885462,
      "calib/mean_conf": 0.9904732142857143,
      "calib/mu_c": 0.9905031847133755,
      "calib/mu_w": 0.9904029850746267,
      "calib/nonempty_final_conf_rate": 0.875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.28958035714285724,
      "calib/std_conf": 0.0020850651815289846,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9221847307430131,
      "calib/step_q_c_n": 1467.0,
      "calib/step_q_gap": -0.00874612427557453,
      "calib/step_q_w": 0.9309308550185876,
      "calib/step_q_w_n": 1345.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.09765625,
      "completions/max_length": 2750.0,
      "completions/max_terminated_length": 2750.0,
      "completions/mean_length": 808.18359375,
      "completions/mean_terminated_length": 895.6493530273438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 452.0,
      "epoch": 0.17706666666666668,
      "grad_norm": 0.1618948131799698,
      "learning_rate": 9.722222222222224e-07,
      "loss": -0.1082,
      "num_tokens": 44684815.0,
      "reward": 1.05078125,
      "reward_std": 0.3207697868347168,
      "rewards/accuracy_reward_step": 0.61328125,
      "rewards/format_reward_step": 0.875,
      "step": 166
    },
    {
      "calib/answer_extract_rate": 0.93359375,
      "calib/avg_num_step_conf": 10.26953125,
      "calib/ece": 0.32927615062761517,
      "calib/final_conf_rate": 0.93359375,
      "calib/format_rate": 0.93359375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00017768401312734383,
      "calib/mean_conf": 0.9903640167364017,
      "calib/mu_c": 0.9903037974683542,
      "calib/mu_w": 0.9904814814814815,
      "calib/nonempty_final_conf_rate": 0.93359375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.32927615062761517,
      "calib/std_conf": 0.0018424680372639151,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9210435365036839,
      "calib/step_q_c_n": 1493.0,
      "calib/step_q_gap": -0.014997836735752834,
      "calib/step_q_w": 0.9360413732394367,
      "calib/step_q_w_n": 1136.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.05078125,
      "completions/max_length": 2963.0,
      "completions/max_terminated_length": 2963.0,
      "completions/mean_length": 782.7734375,
      "completions/mean_terminated_length": 824.6502075195312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 500.0,
      "epoch": 0.17813333333333334,
      "grad_norm": 0.1658693253993988,
      "learning_rate": 9.444444444444445e-07,
      "loss": -0.0328,
      "num_tokens": 44990813.0,
      "reward": 1.083984375,
      "reward_std": 0.25721055269241333,
      "rewards/accuracy_reward_step": 0.6171875,
      "rewards/format_reward_step": 0.93359375,
      "step": 167
    },
    {
      "calib/answer_extract_rate": 0.91796875,
      "calib/avg_num_step_conf": 10.578125,
      "calib/ece": 0.30531489361702135,
      "calib/final_conf_rate": 0.91796875,
      "calib/format_rate": 0.91796875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0001740809132110277,
      "calib/mean_conf": 0.9904212765957447,
      "calib/mu_c": 0.9903664596273293,
      "calib/mu_w": 0.9905405405405403,
      "calib/nonempty_final_conf_rate": 0.91796875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.30531489361702135,
      "calib/std_conf": 0.001999248387515709,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9184727272727273,
      "calib/step_q_c_n": 1485.0,
      "calib/step_q_gap": -0.005324492678213,
      "calib/step_q_w": 0.9237972199509403,
      "calib/step_q_w_n": 1223.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 2418.0,
      "completions/max_terminated_length": 2418.0,
      "completions/mean_length": 827.09375,
      "completions/mean_terminated_length": 882.2333984375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 467.0,
      "epoch": 0.1792,
      "grad_norm": 0.18554162979125977,
      "learning_rate": 9.166666666666666e-07,
      "loss": -0.1005,
      "num_tokens": 45307221.0,
      "reward": 1.087890625,
      "reward_std": 0.41108518838882446,
      "rewards/accuracy_reward_step": 0.62890625,
      "rewards/format_reward_step": 0.91796875,
      "step": 168
    },
    {
      "calib/answer_extract_rate": 0.9453125,
      "calib/avg_num_step_conf": 10.1875,
      "calib/ece": 0.3830495867768595,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.9453125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00030017901897638755,
      "calib/mean_conf": 0.9904876033057851,
      "calib/mu_c": 0.9906054421768707,
      "calib/mu_w": 0.9903052631578944,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3830495867768595,
      "calib/std_conf": 0.002136327668862951,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9197799999999999,
      "calib/step_q_c_n": 1350.0,
      "calib/step_q_gap": -0.0173026709062003,
      "calib/step_q_w": 0.9370826709062002,
      "calib/step_q_w_n": 1258.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 2902.0,
      "completions/max_terminated_length": 2902.0,
      "completions/mean_length": 835.328125,
      "completions/mean_terminated_length": 876.4097900390625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 432.0,
      "epoch": 0.18026666666666666,
      "grad_norm": 0.1450500339269638,
      "learning_rate": 8.88888888888889e-07,
      "loss": -0.0578,
      "num_tokens": 45625249.0,
      "reward": 1.046875,
      "reward_std": 0.22004562616348267,
      "rewards/accuracy_reward_step": 0.57421875,
      "rewards/format_reward_step": 0.9453125,
      "step": 169
    },
    {
      "calib/answer_extract_rate": 0.9453125,
      "calib/avg_num_step_conf": 9.140625,
      "calib/ece": 0.3747272727272729,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.9453125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0001228260085154309,
      "calib/mean_conf": 0.9904297520661158,
      "calib/mu_c": 0.9903825503355703,
      "calib/mu_w": 0.9905053763440858,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3747272727272729,
      "calib/std_conf": 0.0019722305460971212,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9167304542069992,
      "calib/step_q_c_n": 1343.0,
      "calib/step_q_gap": -0.007478171670633915,
      "calib/step_q_w": 0.9242086258776331,
      "calib/step_q_w_n": 997.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 2763.0,
      "completions/max_terminated_length": 2763.0,
      "completions/mean_length": 811.234375,
      "completions/mean_terminated_length": 844.2113647460938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 377.0,
      "epoch": 0.18133333333333335,
      "grad_norm": 0.180429607629776,
      "learning_rate": 8.611111111111112e-07,
      "loss": -0.0272,
      "num_tokens": 45937077.0,
      "reward": 1.0546875,
      "reward_std": 0.3228003978729248,
      "rewards/accuracy_reward_step": 0.58203125,
      "rewards/format_reward_step": 0.9453125,
      "step": 170
    },
    {
      "calib/answer_extract_rate": 0.953125,
      "calib/avg_num_step_conf": 9.11328125,
      "calib/ece": 0.5066803278688524,
      "calib/final_conf_rate": 0.953125,
      "calib/format_rate": 0.953125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0002273338714018358,
      "calib/mean_conf": 0.9902868852459016,
      "calib/mu_c": 0.9901694915254236,
      "calib/mu_w": 0.9903968253968254,
      "calib/nonempty_final_conf_rate": 0.953125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5066803278688524,
      "calib/std_conf": 0.0016692960536406824,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9185578747628085,
      "calib/step_q_c_n": 1054.0,
      "calib/step_q_gap": 0.0031591257401344386,
      "calib/step_q_w": 0.915398749022674,
      "calib/step_q_w_n": 1279.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 2559.0,
      "completions/max_terminated_length": 2559.0,
      "completions/mean_length": 805.4296875,
      "completions/mean_terminated_length": 838.1707153320312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 463.0,
      "epoch": 0.1824,
      "grad_norm": 0.17943838238716125,
      "learning_rate": 8.333333333333333e-07,
      "loss": -0.0565,
      "num_tokens": 46250163.0,
      "reward": 0.9375,
      "reward_std": 0.32925909757614136,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/format_reward_step": 0.953125,
      "step": 171
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 8.5859375,
      "calib/ece": 0.2671767068273093,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0002864734299514726,
      "calib/mean_conf": 0.9900682730923696,
      "calib/mu_c": 0.9899888888888888,
      "calib/mu_w": 0.9902753623188403,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.2671767068273093,
      "calib/std_conf": 0.003195025851300865,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9176888454011741,
      "calib/step_q_c_n": 1533.0,
      "calib/step_q_gap": 0.0047114017921515705,
      "calib/step_q_w": 0.9129774436090226,
      "calib/step_q_w_n": 665.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 2987.0,
      "completions/max_terminated_length": 2987.0,
      "completions/mean_length": 776.5703125,
      "completions/mean_terminated_length": 795.2080078125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 484.0,
      "epoch": 0.18346666666666667,
      "grad_norm": 0.1873815506696701,
      "learning_rate": 8.055555555555557e-07,
      "loss": -0.0292,
      "num_tokens": 46552317.0,
      "reward": 1.189453125,
      "reward_std": 0.30480220913887024,
      "rewards/accuracy_reward_step": 0.703125,
      "rewards/format_reward_step": 0.97265625,
      "step": 172
    },
    {
      "calib/answer_extract_rate": 0.953125,
      "calib/avg_num_step_conf": 9.6171875,
      "calib/ece": 0.33864344262295076,
      "calib/final_conf_rate": 0.953125,
      "calib/format_rate": 0.953125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00027014428412885927,
      "calib/mean_conf": 0.9902827868852458,
      "calib/mu_c": 0.9901886792452826,
      "calib/mu_w": 0.9904588235294115,
      "calib/nonempty_final_conf_rate": 0.953125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.33864344262295076,
      "calib/std_conf": 0.0016465160746528724,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9175739644970413,
      "calib/step_q_c_n": 1521.0,
      "calib/step_q_gap": -0.0043410195624697945,
      "calib/step_q_w": 0.9219149840595111,
      "calib/step_q_w_n": 941.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 2843.0,
      "completions/max_terminated_length": 2843.0,
      "completions/mean_length": 846.43359375,
      "completions/mean_terminated_length": 870.2288818359375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 371.0,
      "epoch": 0.18453333333333333,
      "grad_norm": 0.16905918717384338,
      "learning_rate": 7.777777777777779e-07,
      "loss": 0.001,
      "num_tokens": 46872164.0,
      "reward": 1.09765625,
      "reward_std": 0.251212477684021,
      "rewards/accuracy_reward_step": 0.62109375,
      "rewards/format_reward_step": 0.953125,
      "step": 173
    },
    {
      "calib/answer_extract_rate": 0.921875,
      "calib/avg_num_step_conf": 9.7734375,
      "calib/ece": 0.4460168776371308,
      "calib/final_conf_rate": 0.92578125,
      "calib/format_rate": 0.921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -4.026701119708154e-05,
      "calib/mean_conf": 0.9903206751054853,
      "calib/mu_c": 0.9903023255813954,
      "calib/mu_w": 0.9903425925925925,
      "calib/nonempty_final_conf_rate": 0.92578125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4460168776371308,
      "calib/std_conf": 0.0019483073817899558,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9168514588859417,
      "calib/step_q_c_n": 1131.0,
      "calib/step_q_gap": -0.004921699392687051,
      "calib/step_q_w": 0.9217731582786287,
      "calib/step_q_w_n": 1371.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0546875,
      "completions/max_length": 3056.0,
      "completions/max_terminated_length": 3056.0,
      "completions/mean_length": 894.421875,
      "completions/mean_terminated_length": 946.1652221679688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 491.0,
      "epoch": 0.1856,
      "grad_norm": 0.15766571462154388,
      "learning_rate": 7.5e-07,
      "loss": -0.0503,
      "num_tokens": 47205368.0,
      "reward": 0.96484375,
      "reward_std": 0.3347104787826538,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/format_reward_step": 0.921875,
      "step": 174
    },
    {
      "calib/answer_extract_rate": 0.94140625,
      "calib/avg_num_step_conf": 8.74609375,
      "calib/ece": 0.5440375,
      "calib/final_conf_rate": 0.9375,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 0.9958333333333333,
      "calib/gap": -0.0007787927763334146,
      "calib/mean_conf": 0.9898708333333334,
      "calib/mu_c": 0.9894392523364485,
      "calib/mu_w": 0.9902180451127819,
      "calib/nonempty_final_conf_rate": 0.9375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.5440375,
      "calib/std_conf": 0.00601283205366286,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9203290870488323,
      "calib/step_q_c_n": 942.0,
      "calib/step_q_gap": 0.0061270824227721565,
      "calib/step_q_w": 0.9142020046260602,
      "calib/step_q_w_n": 1297.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03515625,
      "completions/max_length": 2921.0,
      "completions/max_terminated_length": 2921.0,
      "completions/mean_length": 832.51953125,
      "completions/mean_terminated_length": 862.854248046875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 507.0,
      "epoch": 0.18666666666666668,
      "grad_norm": 0.1711643487215042,
      "learning_rate": 7.222222222222222e-07,
      "loss": -0.0372,
      "num_tokens": 47524317.0,
      "reward": 0.890625,
      "reward_std": 0.2728787660598755,
      "rewards/accuracy_reward_step": 0.421875,
      "rewards/format_reward_step": 0.9375,
      "step": 175
    },
    {
      "calib/answer_extract_rate": 0.94921875,
      "calib/avg_num_step_conf": 9.43359375,
      "calib/ece": 0.41413168724279836,
      "calib/final_conf_rate": 0.94921875,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00020436893203912287,
      "calib/mean_conf": 0.9902633744855968,
      "calib/mu_c": 0.99035,
      "calib/mu_w": 0.9901456310679608,
      "calib/nonempty_final_conf_rate": 0.94921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.41413168724279836,
      "calib/std_conf": 0.002713435579808148,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9192688870836719,
      "calib/step_q_c_n": 1231.0,
      "calib/step_q_gap": 0.018897265462050394,
      "calib/step_q_w": 0.9003716216216215,
      "calib/step_q_w_n": 1184.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03515625,
      "completions/max_length": 2535.0,
      "completions/max_terminated_length": 2535.0,
      "completions/mean_length": 817.59765625,
      "completions/mean_terminated_length": 847.388671875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 389.0,
      "epoch": 0.18773333333333334,
      "grad_norm": 0.17403370141983032,
      "learning_rate": 6.944444444444446e-07,
      "loss": -0.0415,
      "num_tokens": 47837686.0,
      "reward": 1.021484375,
      "reward_std": 0.3301824629306793,
      "rewards/accuracy_reward_step": 0.546875,
      "rewards/format_reward_step": 0.94921875,
      "step": 176
    },
    {
      "calib/answer_extract_rate": 0.94921875,
      "calib/avg_num_step_conf": 8.359375,
      "calib/ece": 0.36323770491803276,
      "calib/final_conf_rate": 0.953125,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00010701716584071708,
      "calib/mean_conf": 0.9902868852459016,
      "calib/mu_c": 0.9903267973856208,
      "calib/mu_w": 0.9902197802197801,
      "calib/nonempty_final_conf_rate": 0.953125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.36323770491803276,
      "calib/std_conf": 0.001669296053640682,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9181977186311787,
      "calib/step_q_c_n": 1315.0,
      "calib/step_q_gap": 0.0026583246917848413,
      "calib/step_q_w": 0.9155393939393939,
      "calib/step_q_w_n": 825.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 2395.0,
      "completions/max_terminated_length": 2395.0,
      "completions/mean_length": 813.44140625,
      "completions/mean_terminated_length": 836.3092041015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 500.0,
      "epoch": 0.1888,
      "grad_norm": 0.1733151078224182,
      "learning_rate": 6.666666666666667e-07,
      "loss": -0.0216,
      "num_tokens": 48149759.0,
      "reward": 1.072265625,
      "reward_std": 0.222038134932518,
      "rewards/accuracy_reward_step": 0.59765625,
      "rewards/format_reward_step": 0.94921875,
      "step": 177
    },
    {
      "calib/answer_extract_rate": 0.94140625,
      "calib/avg_num_step_conf": 8.87890625,
      "calib/ece": 0.3460167364016735,
      "calib/final_conf_rate": 0.93359375,
      "calib/format_rate": 0.93359375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.0008709702062640323,
      "calib/mean_conf": 0.99036820083682,
      "calib/mu_c": 0.9900584415584415,
      "calib/mu_w": 0.9909294117647055,
      "calib/nonempty_final_conf_rate": 0.93359375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3460167364016735,
      "calib/std_conf": 0.0018630949263079577,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9145264750378214,
      "calib/step_q_c_n": 1322.0,
      "calib/step_q_gap": 0.0018482415993356094,
      "calib/step_q_w": 0.9126782334384858,
      "calib/step_q_w_n": 951.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.04296875,
      "completions/max_length": 3021.0,
      "completions/max_terminated_length": 3021.0,
      "completions/mean_length": 782.7109375,
      "completions/mean_terminated_length": 817.85302734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 470.0,
      "epoch": 0.18986666666666666,
      "grad_norm": 0.18498072028160095,
      "learning_rate": 6.388888888888889e-07,
      "loss": -0.0341,
      "num_tokens": 48456205.0,
      "reward": 1.068359375,
      "reward_std": 0.31419384479522705,
      "rewards/accuracy_reward_step": 0.6015625,
      "rewards/format_reward_step": 0.93359375,
      "step": 178
    },
    {
      "calib/answer_extract_rate": 0.9296875,
      "calib/avg_num_step_conf": 8.5,
      "calib/ece": 0.3417907949790795,
      "calib/final_conf_rate": 0.93359375,
      "calib/format_rate": 0.92578125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00021267281105996627,
      "calib/mean_conf": 0.990326359832636,
      "calib/mu_c": 0.9902516129032256,
      "calib/mu_w": 0.9904642857142856,
      "calib/nonempty_final_conf_rate": 0.93359375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.3417907949790795,
      "calib/std_conf": 0.0017554981567835014,
      "calib/step_conf_rate": 0.9921875,
      "calib/step_q_c": 0.9156312111801242,
      "calib/step_q_c_n": 1288.0,
      "calib/step_q_gap": 0.013353058026971176,
      "calib/step_q_w": 0.9022781531531531,
      "calib/step_q_w_n": 888.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 2873.0,
      "completions/max_terminated_length": 2873.0,
      "completions/mean_length": 790.3671875,
      "completions/mean_terminated_length": 829.2376708984375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 432.0,
      "epoch": 0.19093333333333334,
      "grad_norm": 0.17732445895671844,
      "learning_rate": 6.111111111111112e-07,
      "loss": -0.0293,
      "num_tokens": 48764803.0,
      "reward": 1.068359375,
      "reward_std": 0.3088341951370239,
      "rewards/accuracy_reward_step": 0.60546875,
      "rewards/format_reward_step": 0.92578125,
      "step": 179
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 8.3828125,
      "calib/ece": 0.3602357723577235,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0001718539524990792,
      "calib/mean_conf": 0.9903170731707317,
      "calib/mu_c": 0.9903806451612902,
      "calib/mu_w": 0.9902087912087911,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3602357723577235,
      "calib/std_conf": 0.0017311919535429633,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9111107668474052,
      "calib/step_q_c_n": 1291.0,
      "calib/step_q_gap": -0.004353560637974874,
      "calib/step_q_w": 0.9154643274853801,
      "calib/step_q_w_n": 855.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 2997.0,
      "completions/max_terminated_length": 2997.0,
      "completions/mean_length": 836.5234375,
      "completions/mean_terminated_length": 860.0401000976562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 477.0,
      "epoch": 0.192,
      "grad_norm": 0.16332538425922394,
      "learning_rate": 5.833333333333334e-07,
      "loss": -0.0279,
      "num_tokens": 49082809.0,
      "reward": 1.08984375,
      "reward_std": 0.2516862750053406,
      "rewards/accuracy_reward_step": 0.609375,
      "rewards/format_reward_step": 0.9609375,
      "step": 180
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 8.2890625,
      "calib/ece": 0.43997590361445804,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00015888946819575533,
      "calib/mean_conf": 0.9901767068273094,
      "calib/mu_c": 0.9902481751824816,
      "calib/mu_w": 0.9900892857142859,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.43997590361445804,
      "calib/std_conf": 0.0012646330635942267,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9141155595996361,
      "calib/step_q_c_n": 1099.0,
      "calib/step_q_gap": -0.0005667473407353629,
      "calib/step_q_w": 0.9146823069403714,
      "calib/step_q_w_n": 1023.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 2656.0,
      "completions/max_terminated_length": 2656.0,
      "completions/mean_length": 782.58203125,
      "completions/mean_terminated_length": 798.1713256835938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 414.0,
      "epoch": 0.19306666666666666,
      "grad_norm": 0.19966793060302734,
      "learning_rate": 5.555555555555555e-07,
      "loss": -0.0271,
      "num_tokens": 49389414.0,
      "reward": 1.01953125,
      "reward_std": 0.29831817746162415,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/format_reward_step": 0.96875,
      "step": 181
    },
    {
      "calib/answer_extract_rate": 0.984375,
      "calib/avg_num_step_conf": 7.88671875,
      "calib/ece": 0.39507936507936514,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 3.9215686274518546e-05,
      "calib/mean_conf": 0.9903174603174604,
      "calib/mu_c": 0.9903333333333331,
      "calib/mu_w": 0.9902941176470585,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.39507936507936514,
      "calib/std_conf": 0.0017532319074900432,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9131224322103533,
      "calib/step_q_c_n": 1217.0,
      "calib/step_q_gap": 0.002523928469704839,
      "calib/step_q_w": 0.9105985037406484,
      "calib/step_q_w_n": 802.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 3039.0,
      "completions/max_terminated_length": 3039.0,
      "completions/mean_length": 802.16015625,
      "completions/mean_terminated_length": 808.4763793945312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 473.0,
      "epoch": 0.19413333333333332,
      "grad_norm": 0.17675307393074036,
      "learning_rate": 5.277777777777779e-07,
      "loss": 0.0234,
      "num_tokens": 49700927.0,
      "reward": 1.076171875,
      "reward_std": 0.2512344717979431,
      "rewards/accuracy_reward_step": 0.5859375,
      "rewards/format_reward_step": 0.98046875,
      "step": 182
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 7.47265625,
      "calib/ece": 0.42914634146341457,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00021739130434783593,
      "calib/mean_conf": 0.9901219512195122,
      "calib/mu_c": 0.9902173913043477,
      "calib/mu_w": 0.9899999999999999,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.42914634146341457,
      "calib/std_conf": 0.0010975609756097569,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9101256038647343,
      "calib/step_q_c_n": 1035.0,
      "calib/step_q_gap": 0.002125603864734371,
      "calib/step_q_w": 0.9079999999999999,
      "calib/step_q_w_n": 878.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 2599.0,
      "completions/max_terminated_length": 2599.0,
      "completions/mean_length": 835.4375,
      "completions/mean_terminated_length": 858.9236450195312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 414.0,
      "epoch": 0.1952,
      "grad_norm": 0.20332731306552887,
      "learning_rate": 5.000000000000001e-07,
      "loss": -0.0296,
      "num_tokens": 50021479.0,
      "reward": 1.01953125,
      "reward_std": 0.33503273129463196,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/format_reward_step": 0.9609375,
      "step": 183
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 8.26953125,
      "calib/ece": 0.3126612903225807,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00011904761904768524,
      "calib/mean_conf": 0.9900806451612904,
      "calib/mu_c": 0.9901190476190477,
      "calib/mu_w": 0.99,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3126612903225807,
      "calib/std_conf": 0.0008943981053555988,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9111671924290222,
      "calib/step_q_c_n": 1268.0,
      "calib/step_q_gap": 0.009423965102756116,
      "calib/step_q_w": 0.9017432273262661,
      "calib/step_q_w_n": 849.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 2445.0,
      "completions/max_terminated_length": 2445.0,
      "completions/mean_length": 798.4609375,
      "completions/mean_terminated_length": 817.6240234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 476.0,
      "epoch": 0.19626666666666667,
      "grad_norm": 0.18218982219696045,
      "learning_rate": 4.7222222222222226e-07,
      "loss": -0.0186,
      "num_tokens": 50331165.0,
      "reward": 1.140625,
      "reward_std": 0.23907458782196045,
      "rewards/accuracy_reward_step": 0.65625,
      "rewards/format_reward_step": 0.96875,
      "step": 184
    },
    {
      "calib/answer_extract_rate": 0.9375,
      "calib/avg_num_step_conf": 7.78125,
      "calib/ece": 0.42362500000000003,
      "calib/final_conf_rate": 0.9375,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0003450226244343657,
      "calib/mean_conf": 0.9902916666666667,
      "calib/mu_c": 0.9904411764705882,
      "calib/mu_w": 0.9900961538461538,
      "calib/nonempty_final_conf_rate": 0.9375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.42362500000000003,
      "calib/std_conf": 0.0016827350421923905,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9101146131805159,
      "calib/step_q_c_n": 1047.0,
      "calib/step_q_gap": 0.0370458301117329,
      "calib/step_q_w": 0.873068783068783,
      "calib/step_q_w_n": 945.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 2840.0,
      "completions/max_terminated_length": 2840.0,
      "completions/mean_length": 813.578125,
      "completions/mean_terminated_length": 839.8225708007812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 399.0,
      "epoch": 0.19733333333333333,
      "grad_norm": 0.16892407834529877,
      "learning_rate": 4.444444444444445e-07,
      "loss": -0.0162,
      "num_tokens": 50646361.0,
      "reward": 1.0,
      "reward_std": 0.2559589147567749,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/format_reward_step": 0.9375,
      "step": 185
    },
    {
      "calib/answer_extract_rate": 0.94921875,
      "calib/avg_num_step_conf": 8.07421875,
      "calib/ece": 0.4303662551440329,
      "calib/final_conf_rate": 0.94921875,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -8.411214953252255e-05,
      "calib/mean_conf": 0.990037037037037,
      "calib/mu_c": 0.9900000000000001,
      "calib/mu_w": 0.9900841121495326,
      "calib/nonempty_final_conf_rate": 0.94921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4303662551440329,
      "calib/std_conf": 0.0005761610809668172,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9128520179372198,
      "calib/step_q_c_n": 1115.0,
      "calib/step_q_gap": 0.001009580962429979,
      "calib/step_q_w": 0.9118424369747898,
      "calib/step_q_w_n": 952.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 2818.0,
      "completions/max_terminated_length": 2818.0,
      "completions/mean_length": 812.67578125,
      "completions/mean_terminated_length": 845.7113647460938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 400.0,
      "epoch": 0.1984,
      "grad_norm": 0.1680852174758911,
      "learning_rate": 4.1666666666666667e-07,
      "loss": -0.0338,
      "num_tokens": 50959446.0,
      "reward": 1.005859375,
      "reward_std": 0.3241080045700073,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/format_reward_step": 0.94921875,
      "step": 186
    },
    {
      "calib/answer_extract_rate": 0.9609375,
      "calib/avg_num_step_conf": 7.8125,
      "calib/ece": 0.40235510204081626,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00011571782178176893,
      "calib/mean_conf": 0.9901102040816326,
      "calib/mu_c": 0.9900625000000001,
      "calib/mu_w": 0.9901782178217818,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.40235510204081626,
      "calib/std_conf": 0.0009939089799330242,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9107363717605005,
      "calib/step_q_c_n": 1119.0,
      "calib/step_q_gap": 0.008383363814984057,
      "calib/step_q_w": 0.9023530079455164,
      "calib/step_q_w_n": 881.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 2901.0,
      "completions/max_terminated_length": 2901.0,
      "completions/mean_length": 843.18359375,
      "completions/mean_terminated_length": 859.9801025390625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 492.0,
      "epoch": 0.19946666666666665,
      "grad_norm": 0.18715514242649078,
      "learning_rate": 3.8888888888888895e-07,
      "loss": 0.0284,
      "num_tokens": 51276845.0,
      "reward": 1.041015625,
      "reward_std": 0.3368094861507416,
      "rewards/accuracy_reward_step": 0.5625,
      "rewards/format_reward_step": 0.95703125,
      "step": 187
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 7.51171875,
      "calib/ece": 0.37746640316205526,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 3.719552337055898e-05,
      "calib/mean_conf": 0.9901146245059288,
      "calib/mu_c": 0.9901290322580644,
      "calib/mu_w": 0.9900918367346938,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.37746640316205526,
      "calib/std_conf": 0.001047632167986198,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.910450997398092,
      "calib/step_q_c_n": 1153.0,
      "calib/step_q_gap": -0.00034640519931072866,
      "calib/step_q_w": 0.9107974025974027,
      "calib/step_q_w_n": 770.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 3044.0,
      "completions/max_terminated_length": 3044.0,
      "completions/mean_length": 873.3203125,
      "completions/mean_terminated_length": 876.7451782226562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 385.0,
      "epoch": 0.20053333333333334,
      "grad_norm": 0.15324848890304565,
      "learning_rate": 3.611111111111111e-07,
      "loss": -0.0018,
      "num_tokens": 51604487.0,
      "reward": 1.099609375,
      "reward_std": 0.23567822575569153,
      "rewards/accuracy_reward_step": 0.60546875,
      "rewards/format_reward_step": 0.98828125,
      "step": 188
    },
    {
      "calib/answer_extract_rate": 0.94921875,
      "calib/avg_num_step_conf": 7.66015625,
      "calib/ece": 0.35780991735537204,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.9453125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00011235955056165814,
      "calib/mean_conf": 0.9900413223140497,
      "calib/mu_c": 0.99,
      "calib/mu_w": 0.9901123595505616,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.35780991735537204,
      "calib/std_conf": 0.0006414948221595057,
      "calib/step_conf_rate": 0.99609375,
      "calib/step_q_c": 0.9097991631799164,
      "calib/step_q_c_n": 1195.0,
      "calib/step_q_gap": 0.0024623485585064664,
      "calib/step_q_w": 0.9073368146214099,
      "calib/step_q_w_n": 766.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0390625,
      "completions/max_length": 2874.0,
      "completions/max_terminated_length": 2874.0,
      "completions/mean_length": 791.296875,
      "completions/mean_terminated_length": 823.46337890625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 438.0,
      "epoch": 0.2016,
      "grad_norm": 0.19075053930282593,
      "learning_rate": 3.3333333333333335e-07,
      "loss": -0.0486,
      "num_tokens": 51914827.0,
      "reward": 1.0703125,
      "reward_std": 0.27192389965057373,
      "rewards/accuracy_reward_step": 0.59765625,
      "rewards/format_reward_step": 0.9453125,
      "step": 189
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 8.15234375,
      "calib/ece": 0.33951807228915676,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.0001851851851855102,
      "calib/mean_conf": 0.990120481927711,
      "calib/mu_c": 0.9901851851851852,
      "calib/mu_w": 0.9899999999999997,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.33951807228915676,
      "calib/std_conf": 0.0010910102576069185,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9096304849884528,
      "calib/step_q_c_n": 1299.0,
      "calib/step_q_gap": -0.0019177383618010913,
      "calib/step_q_w": 0.9115482233502539,
      "calib/step_q_w_n": 788.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 1823.0,
      "completions/max_terminated_length": 1823.0,
      "completions/mean_length": 818.33984375,
      "completions/mean_terminated_length": 841.3453369140625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 417.0,
      "epoch": 0.20266666666666666,
      "grad_norm": 0.1583220213651657,
      "learning_rate": 3.055555555555556e-07,
      "loss": -0.0538,
      "num_tokens": 52229930.0,
      "reward": 1.119140625,
      "reward_std": 0.244254007935524,
      "rewards/accuracy_reward_step": 0.6328125,
      "rewards/format_reward_step": 0.97265625,
      "step": 190
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 8.2421875,
      "calib/ece": 0.39330645161290323,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -3.2432432432449865e-05,
      "calib/mean_conf": 0.9900806451612904,
      "calib/mu_c": 0.9900675675675674,
      "calib/mu_w": 0.9900999999999999,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.39330645161290323,
      "calib/std_conf": 0.0008943981053555988,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9113084112149532,
      "calib/step_q_c_n": 1177.0,
      "calib/step_q_gap": 0.008811090743356065,
      "calib/step_q_w": 0.9024973204715971,
      "calib/step_q_w_n": 933.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 2827.0,
      "completions/max_terminated_length": 2827.0,
      "completions/mean_length": 807.09765625,
      "completions/mean_terminated_length": 823.17529296875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 439.0,
      "epoch": 0.20373333333333332,
      "grad_norm": 0.15239450335502625,
      "learning_rate": 2.7777777777777776e-07,
      "loss": 0.0041,
      "num_tokens": 52540715.0,
      "reward": 1.0625,
      "reward_std": 0.2156136929988861,
      "rewards/accuracy_reward_step": 0.578125,
      "rewards/format_reward_step": 0.96875,
      "step": 191
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 7.78515625,
      "calib/ece": 0.3033694779116467,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00018511021142553208,
      "calib/mean_conf": 0.9901164658634539,
      "calib/mu_c": 0.9900584795321639,
      "calib/mu_w": 0.9902435897435894,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3033694779116467,
      "calib/std_conf": 0.0010559118139668953,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9111232676878191,
      "calib/step_q_c_n": 1371.0,
      "calib/step_q_gap": 0.012917479906468743,
      "calib/step_q_w": 0.8982057877813504,
      "calib/step_q_w_n": 622.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2299.0,
      "completions/max_terminated_length": 2299.0,
      "completions/mean_length": 818.88671875,
      "completions/mean_terminated_length": 831.8849487304688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 386.0,
      "epoch": 0.2048,
      "grad_norm": 0.1535855233669281,
      "learning_rate": 2.5000000000000004e-07,
      "loss": -0.0124,
      "num_tokens": 52855326.0,
      "reward": 1.154296875,
      "reward_std": 0.2417459487915039,
      "rewards/accuracy_reward_step": 0.66796875,
      "rewards/format_reward_step": 0.97265625,
      "step": 192
    },
    {
      "calib/answer_extract_rate": 0.96875,
      "calib/avg_num_step_conf": 8.0,
      "calib/ece": 0.4038669354838709,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.9919354838709677,
      "calib/gap": 0.014229166666666515,
      "calib/mean_conf": 0.9845120967741936,
      "calib/mu_c": 0.9904791666666666,
      "calib/mu_w": 0.9762500000000001,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4038669354838709,
      "calib/std_conf": 0.06710944250269774,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9107306985294119,
      "calib/step_q_c_n": 1088.0,
      "calib/step_q_gap": 0.015761948529411818,
      "calib/step_q_w": 0.8949687500000001,
      "calib/step_q_w_n": 960.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 1859.0,
      "completions/max_terminated_length": 1859.0,
      "completions/mean_length": 813.01953125,
      "completions/mean_terminated_length": 829.2151489257812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 488.0,
      "epoch": 0.20586666666666667,
      "grad_norm": 0.18594703078269958,
      "learning_rate": 2.2222222222222224e-07,
      "loss": -0.0141,
      "num_tokens": 53169171.0,
      "reward": 1.046875,
      "reward_std": 0.31658273935317993,
      "rewards/accuracy_reward_step": 0.5625,
      "rewards/format_reward_step": 0.96875,
      "step": 193
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 7.84375,
      "calib/ece": 0.39098380566801616,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00035005460005466116,
      "calib/mean_conf": 0.9901740890688259,
      "calib/mu_c": 0.9900337837837836,
      "calib/mu_w": 0.9903838383838383,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.39098380566801616,
      "calib/std_conf": 0.0012395542003270886,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9095937770095073,
      "calib/step_q_c_n": 1157.0,
      "calib/step_q_gap": 0.007634905094113575,
      "calib/step_q_w": 0.9019588719153937,
      "calib/step_q_w_n": 851.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 3048.0,
      "completions/max_terminated_length": 3048.0,
      "completions/mean_length": 811.765625,
      "completions/mean_terminated_length": 821.391357421875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 473.0,
      "epoch": 0.20693333333333333,
      "grad_norm": 0.1659308820962906,
      "learning_rate": 1.9444444444444447e-07,
      "loss": 0.0241,
      "num_tokens": 53482927.0,
      "reward": 1.060546875,
      "reward_std": 0.2599868178367615,
      "rewards/accuracy_reward_step": 0.578125,
      "rewards/format_reward_step": 0.96484375,
      "step": 194
    },
    {
      "calib/answer_extract_rate": 0.94921875,
      "calib/avg_num_step_conf": 8.26171875,
      "calib/ece": 0.38534567901234573,
      "calib/final_conf_rate": 0.94921875,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 0.00015943877550983476,
      "calib/mean_conf": 0.990283950617284,
      "calib/mu_c": 0.9903469387755099,
      "calib/mu_w": 0.9901875000000001,
      "calib/nonempty_final_conf_rate": 0.94921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.38534567901234573,
      "calib/std_conf": 0.001567949838850028,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9125693617021279,
      "calib/step_q_c_n": 1175.0,
      "calib/step_q_gap": 0.007242765957447106,
      "calib/step_q_w": 0.9053265957446808,
      "calib/step_q_w_n": 940.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 2934.0,
      "completions/max_terminated_length": 2934.0,
      "completions/mean_length": 810.203125,
      "completions/mean_terminated_length": 836.3386840820312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 441.0,
      "epoch": 0.208,
      "grad_norm": 0.16117671132087708,
      "learning_rate": 1.6666666666666668e-07,
      "loss": 0.0094,
      "num_tokens": 53796323.0,
      "reward": 1.048828125,
      "reward_std": 0.2848566770553589,
      "rewards/accuracy_reward_step": 0.57421875,
      "rewards/format_reward_step": 0.94921875,
      "step": 195
    },
    {
      "calib/answer_extract_rate": 0.98828125,
      "calib/avg_num_step_conf": 7.98046875,
      "calib/ece": 0.33410671936758896,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00015261044176695382,
      "calib/mean_conf": 0.9902332015810277,
      "calib/mu_c": 0.9901807228915661,
      "calib/mu_w": 0.9903333333333331,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.33410671936758896,
      "calib/std_conf": 0.0014973508975364576,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9123646071700993,
      "calib/step_q_c_n": 1311.0,
      "calib/step_q_gap": 0.00786460717009918,
      "calib/step_q_w": 0.9045000000000001,
      "calib/step_q_w_n": 732.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 1867.0,
      "completions/max_terminated_length": 1867.0,
      "completions/mean_length": 735.75,
      "completions/mean_terminated_length": 744.474365234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 444.0,
      "epoch": 0.20906666666666668,
      "grad_norm": 0.1402597725391388,
      "learning_rate": 1.3888888888888888e-07,
      "loss": -0.0177,
      "num_tokens": 54087219.0,
      "reward": 1.142578125,
      "reward_std": 0.12430930137634277,
      "rewards/accuracy_reward_step": 0.6484375,
      "rewards/format_reward_step": 0.98828125,
      "step": 196
    },
    {
      "calib/answer_extract_rate": 0.9609375,
      "calib/avg_num_step_conf": 8.14453125,
      "calib/ece": 0.4454430894308944,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00018430170575711724,
      "calib/mean_conf": 0.9901585365853659,
      "calib/mu_c": 0.9900746268656716,
      "calib/mu_w": 0.9902589285714287,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.4454430894308944,
      "calib/std_conf": 0.0012343608220068457,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9119418386491558,
      "calib/step_q_c_n": 1066.0,
      "calib/step_q_gap": 0.0061381095029339194,
      "calib/step_q_w": 0.9058037291462219,
      "calib/step_q_w_n": 1019.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 2900.0,
      "completions/max_terminated_length": 2900.0,
      "completions/mean_length": 809.0,
      "completions/mean_terminated_length": 828.416015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 504.0,
      "epoch": 0.21013333333333334,
      "grad_norm": 0.19100052118301392,
      "learning_rate": 1.1111111111111112e-07,
      "loss": -0.0037,
      "num_tokens": 54399379.0,
      "reward": 1.00390625,
      "reward_std": 0.3031455874443054,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/format_reward_step": 0.9609375,
      "step": 197
    },
    {
      "calib/answer_extract_rate": 0.96484375,
      "calib/avg_num_step_conf": 7.859375,
      "calib/ece": 0.34650607287449386,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": 5.4316752428063e-06,
      "calib/mean_conf": 0.9902307692307691,
      "calib/mu_c": 0.9902327044025155,
      "calib/mu_w": 0.9902272727272727,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.34650607287449386,
      "calib/std_conf": 0.0014646249106620441,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9112080924855491,
      "calib/step_q_c_n": 1211.0,
      "calib/step_q_gap": -0.004846838850281254,
      "calib/step_q_w": 0.9160549313358304,
      "calib/step_q_w_n": 801.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03515625,
      "completions/max_length": 2812.0,
      "completions/max_terminated_length": 2812.0,
      "completions/mean_length": 771.47265625,
      "completions/mean_terminated_length": 799.5830078125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 414.0,
      "epoch": 0.2112,
      "grad_norm": 0.17476636171340942,
      "learning_rate": 8.333333333333334e-08,
      "loss": -0.0466,
      "num_tokens": 54702260.0,
      "reward": 1.103515625,
      "reward_std": 0.23565450310707092,
      "rewards/accuracy_reward_step": 0.62109375,
      "rewards/format_reward_step": 0.96484375,
      "step": 198
    },
    {
      "calib/answer_extract_rate": 0.9375,
      "calib/avg_num_step_conf": 7.6171875,
      "calib/ece": 0.31928750000000017,
      "calib/final_conf_rate": 0.9375,
      "calib/format_rate": 0.93359375,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.00019726393584351687,
      "calib/mean_conf": 0.9901208333333334,
      "calib/mu_c": 0.990055900621118,
      "calib/mu_w": 0.9902531645569616,
      "calib/nonempty_final_conf_rate": 0.9375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.31928750000000017,
      "calib/std_conf": 0.0010752825855973355,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9103620414673047,
      "calib/step_q_c_n": 1254.0,
      "calib/step_q_gap": 0.0015258345707528376,
      "calib/step_q_w": 0.9088362068965519,
      "calib/step_q_w_n": 696.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02734375,
      "completions/max_length": 3068.0,
      "completions/max_terminated_length": 3068.0,
      "completions/mean_length": 871.69921875,
      "completions/mean_terminated_length": 896.2047729492188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 427.0,
      "epoch": 0.21226666666666666,
      "grad_norm": 0.19598470628261566,
      "learning_rate": 5.555555555555556e-08,
      "loss": -0.028,
      "num_tokens": 55029615.0,
      "reward": 1.099609375,
      "reward_std": 0.325309157371521,
      "rewards/accuracy_reward_step": 0.6328125,
      "rewards/format_reward_step": 0.93359375,
      "step": 199
    },
    {
      "calib/answer_extract_rate": 0.97265625,
      "calib/avg_num_step_conf": 7.5078125,
      "calib/ece": 0.3274096385542169,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -8.874458874452706e-05,
      "calib/mean_conf": 0.9900602409638555,
      "calib/mu_c": 0.9900303030303029,
      "calib/mu_w": 0.9901190476190475,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.3274096385542169,
      "calib/std_conf": 0.0007059596719376287,
      "calib/step_conf_rate": 1.0,
      "calib/step_q_c": 0.9145186953062848,
      "calib/step_q_c_n": 1257.0,
      "calib/step_q_gap": 0.0061878682386157235,
      "calib/step_q_w": 0.9083308270676691,
      "calib/step_q_w_n": 665.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0234375,
      "completions/max_length": 2309.0,
      "completions/max_terminated_length": 2309.0,
      "completions/mean_length": 811.74609375,
      "completions/mean_terminated_length": 831.22802734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 459.0,
      "epoch": 0.21333333333333335,
      "grad_norm": 0.14802969992160797,
      "learning_rate": 2.777777777777778e-08,
      "loss": -0.0269,
      "num_tokens": 55345470.0,
      "reward": 1.12890625,
      "reward_std": 0.2350997030735016,
      "rewards/accuracy_reward_step": 0.64453125,
      "rewards/format_reward_step": 0.96875,
      "step": 200
    },
    {
      "epoch": 0.21333333333333335,
      "step": 200,
      "total_flos": 0.0,
      "train_loss": 0.006713523315265774,
      "train_runtime": 6552.8806,
      "train_samples_per_second": 7.813,
      "train_steps_per_second": 0.031
    }
  ],
  "logging_steps": 1,
  "max_steps": 200,
  "num_input_tokens_seen": 55345470,
  "num_train_epochs": 1,
  "save_steps": 25,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}