{
  "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": [
    {
      "aux_distill/lambda": 0.1,
      "aux_distill/loss": 0.7561912158360848,
      "aux_distill/mean_u": 0.09821361044272292,
      "aux_distill/n_active_tok": 54.0,
      "calib/answer_extract_rate": 0.08203125,
      "calib/auroc": 0.6944444444444445,
      "calib/avg_num_step_conf": 0.3359375,
      "calib/ece": 0.6230769230769231,
      "calib/final_conf_rate": 0.05078125,
      "calib/format_rate": 0.04296875,
      "calib/frac_conf_gt_0.9": 0.7692307692307693,
      "calib/gap": 0.03861111111111115,
      "calib/mean_conf": 0.9307692307692309,
      "calib/mu_c": 0.9575,
      "calib/mu_w": 0.9188888888888889,
      "calib/nonempty_final_conf_rate": 0.05078125,
      "calib/nonempty_reasoning_rate": 0.09765625,
      "calib/nonempty_step_conf_rate": 0.0703125,
      "calib/pce": 0.6230769230769231,
      "calib/std_conf": 0.07965903671384378,
      "calib/step_conf_rate": 0.0703125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2955.0,
      "completions/max_terminated_length": 2955.0,
      "completions/mean_length": 613.67578125,
      "completions/mean_terminated_length": 674.2532348632812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0010666666666666667,
      "grad_norm": 0.012635531835258007,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.0627,
      "num_tokens": 264685.0,
      "reward": 0.037574999034404755,
      "reward_std": 0.07449960708618164,
      "rewards/accuracy_reward_step": 0.015625,
      "rewards/final_brier_reward_step": 0.01655624993145466,
      "rewards/format_reward_step": 0.04296875,
      "step": 1
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.4544563018961957,
      "aux_distill/mean_u": 0.09256462183530526,
      "aux_distill/n_active_tok": 58.36842105263158,
      "calib/answer_extract_rate": 0.13671875,
      "calib/auroc": 0.5338345864661654,
      "calib/avg_num_step_conf": 0.55078125,
      "calib/ece": 0.6261538461538463,
      "calib/final_conf_rate": 0.1015625,
      "calib/format_rate": 0.08984375,
      "calib/frac_conf_gt_0.9": 0.7692307692307693,
      "calib/gap": 0.002406015037593856,
      "calib/mean_conf": 0.8953846153846153,
      "calib/mu_c": 0.897142857142857,
      "calib/mu_w": 0.8947368421052632,
      "calib/nonempty_final_conf_rate": 0.1015625,
      "calib/nonempty_reasoning_rate": 0.14453125,
      "calib/nonempty_step_conf_rate": 0.109375,
      "calib/pce": 0.6261538461538463,
      "calib/std_conf": 0.18653172073466937,
      "calib/step_conf_rate": 0.109375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3001.0,
      "completions/max_terminated_length": 3001.0,
      "completions/mean_length": 646.4609375,
      "completions/mean_terminated_length": 683.8594970703125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0021333333333333334,
      "grad_norm": 0.009192095138132572,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.0927,
      "num_tokens": 533467.0,
      "reward": 0.07537207007408142,
      "reward_std": 0.14035090804100037,
      "rewards/accuracy_reward_step": 0.03125,
      "rewards/final_brier_reward_step": 0.02965039201080799,
      "rewards/format_reward_step": 0.08984375,
      "step": 2
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.5128775238990784,
      "aux_distill/mean_u": 0.06377023176615988,
      "aux_distill/n_active_tok": 41.6,
      "calib/answer_extract_rate": 0.0390625,
      "calib/auroc": 0.9444444444444444,
      "calib/avg_num_step_conf": 0.203125,
      "calib/ece": 0.6272727272727273,
      "calib/final_conf_rate": 0.04296875,
      "calib/format_rate": 0.0234375,
      "calib/frac_conf_gt_0.9": 0.7272727272727273,
      "calib/gap": 0.2088888888888889,
      "calib/mean_conf": 0.8090909090909091,
      "calib/mu_c": 0.98,
      "calib/mu_w": 0.7711111111111111,
      "calib/nonempty_final_conf_rate": 0.04296875,
      "calib/nonempty_reasoning_rate": 0.05859375,
      "calib/nonempty_step_conf_rate": 0.04296875,
      "calib/pce": 0.6272727272727273,
      "calib/std_conf": 0.31430889685885144,
      "calib/step_conf_rate": 0.04296875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3020.0,
      "completions/max_terminated_length": 3020.0,
      "completions/mean_length": 650.43359375,
      "completions/mean_terminated_length": 727.1222534179688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0032,
      "grad_norm": 0.005481162574142218,
      "learning_rate": 7.5e-07,
      "loss": 0.0425,
      "num_tokens": 805234.0,
      "reward": 0.020432421937584877,
      "reward_std": 0.04559621587395668,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/final_brier_reward_step": 0.009614843875169754,
      "rewards/format_reward_step": 0.0234375,
      "step": 3
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.6010356366634368,
      "aux_distill/mean_u": 0.11464373060808687,
      "aux_distill/n_active_tok": 42.2,
      "calib/answer_extract_rate": 0.06640625,
      "calib/auroc": 0.35,
      "calib/avg_num_step_conf": 0.3125,
      "calib/ece": 0.7383333333333335,
      "calib/final_conf_rate": 0.046875,
      "calib/format_rate": 0.04296875,
      "calib/frac_conf_gt_0.9": 0.6666666666666666,
      "calib/gap": -0.050000000000000155,
      "calib/mean_conf": 0.8466666666666667,
      "calib/mu_c": 0.8049999999999999,
      "calib/mu_w": 0.8550000000000001,
      "calib/nonempty_final_conf_rate": 0.046875,
      "calib/nonempty_reasoning_rate": 0.0859375,
      "calib/nonempty_step_conf_rate": 0.0703125,
      "calib/pce": 0.7091666666666668,
      "calib/std_conf": 0.269639677264967,
      "calib/step_conf_rate": 0.0703125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.07421875,
      "completions/max_length": 3029.0,
      "completions/max_terminated_length": 3029.0,
      "completions/mean_length": 682.984375,
      "completions/mean_terminated_length": 737.7383422851562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.004266666666666667,
      "grad_norm": 0.014602464623749256,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0349,
      "num_tokens": 1086246.0,
      "reward": 0.03255273401737213,
      "reward_std": 0.07881505787372589,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/final_brier_reward_step": 0.014324218966066837,
      "rewards/format_reward_step": 0.04296875,
      "step": 4
    },
    {
      "aux_distill/lambda": 0.1,
      "aux_distill/loss": 0.5481972178587546,
      "aux_distill/mean_u": 0.0767053597090393,
      "aux_distill/n_active_tok": 38.76923076923077,
      "calib/answer_extract_rate": 0.0859375,
      "calib/auroc": 0.58,
      "calib/avg_num_step_conf": 0.24609375,
      "calib/ece": 0.6186666666666667,
      "calib/final_conf_rate": 0.05859375,
      "calib/format_rate": 0.03515625,
      "calib/frac_conf_gt_0.9": 0.9333333333333333,
      "calib/gap": 0.029999999999999916,
      "calib/mean_conf": 0.9520000000000001,
      "calib/mu_c": 0.9720000000000001,
      "calib/mu_w": 0.9420000000000002,
      "calib/nonempty_final_conf_rate": 0.05859375,
      "calib/nonempty_reasoning_rate": 0.09765625,
      "calib/nonempty_step_conf_rate": 0.05859375,
      "calib/pce": 0.6186666666666667,
      "calib/std_conf": 0.0643117407632541,
      "calib/step_conf_rate": 0.05859375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3059.0,
      "completions/max_terminated_length": 3059.0,
      "completions/mean_length": 663.9609375,
      "completions/mean_terminated_length": 748.7841186523438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.005333333333333333,
      "grad_norm": 0.01045077946037054,
      "learning_rate": 1.25e-06,
      "loss": 0.0418,
      "num_tokens": 1362908.0,
      "reward": 0.03767089545726776,
      "reward_std": 0.09181629866361618,
      "rewards/accuracy_reward_step": 0.01953125,
      "rewards/final_brier_reward_step": 0.02065429650247097,
      "rewards/format_reward_step": 0.03515625,
      "step": 5
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.5697961250940958,
      "aux_distill/mean_u": 0.09371336220695657,
      "aux_distill/n_active_tok": 58.733333333333334,
      "calib/answer_extract_rate": 0.0703125,
      "calib/auroc": 0.65625,
      "calib/avg_num_step_conf": 0.4375,
      "calib/ece": 0.8947058823529411,
      "calib/final_conf_rate": 0.06640625,
      "calib/format_rate": 0.04296875,
      "calib/frac_conf_gt_0.9": 0.8823529411764706,
      "calib/gap": 0.028124999999999956,
      "calib/mean_conf": 0.953529411764706,
      "calib/mu_c": 0.98,
      "calib/mu_w": 0.951875,
      "calib/nonempty_final_conf_rate": 0.06640625,
      "calib/nonempty_reasoning_rate": 0.10546875,
      "calib/nonempty_step_conf_rate": 0.0859375,
      "calib/pce": 0.8947058823529411,
      "calib/std_conf": 0.05465842343678744,
      "calib/step_conf_rate": 0.0859375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2953.0,
      "completions/max_terminated_length": 2953.0,
      "completions/mean_length": 598.453125,
      "completions/mean_terminated_length": 660.362060546875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0064,
      "grad_norm": 0.00990479625761509,
      "learning_rate": 1.5e-06,
      "loss": 0.0339,
      "num_tokens": 1622064.0,
      "reward": 0.02689570188522339,
      "reward_std": 0.06286215037107468,
      "rewards/accuracy_reward_step": 0.00390625,
      "rewards/final_brier_reward_step": 0.006916406098753214,
      "rewards/format_reward_step": 0.04296875,
      "step": 6
    },
    {
      "aux_distill/lambda": 0.1,
      "aux_distill/loss": 0.8123426660895348,
      "aux_distill/mean_u": 0.0868831620532763,
      "aux_distill/n_active_tok": 44.357142857142854,
      "calib/answer_extract_rate": 0.10546875,
      "calib/auroc": 0.8472222222222222,
      "calib/avg_num_step_conf": 0.296875,
      "calib/ece": 0.6080000000000001,
      "calib/final_conf_rate": 0.05859375,
      "calib/format_rate": 0.03125,
      "calib/frac_conf_gt_0.9": 0.7333333333333333,
      "calib/gap": 0.21916666666666684,
      "calib/mean_conf": 0.8079999999999998,
      "calib/mu_c": 0.9833333333333334,
      "calib/mu_w": 0.7641666666666665,
      "calib/nonempty_final_conf_rate": 0.05859375,
      "calib/nonempty_reasoning_rate": 0.14453125,
      "calib/nonempty_step_conf_rate": 0.078125,
      "calib/pce": 0.6080000000000001,
      "calib/std_conf": 0.32449345139771313,
      "calib/step_conf_rate": 0.078125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3030.0,
      "completions/max_terminated_length": 3030.0,
      "completions/mean_length": 734.79296875,
      "completions/mean_terminated_length": 810.8060302734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.007466666666666667,
      "grad_norm": 0.020580844953656197,
      "learning_rate": 1.75e-06,
      "loss": 0.0671,
      "num_tokens": 1917595.0,
      "reward": 0.026225585490465164,
      "reward_std": 0.05754803121089935,
      "rewards/accuracy_reward_step": 0.01171875,
      "rewards/final_brier_reward_step": 0.009482420980930328,
      "rewards/format_reward_step": 0.03125,
      "step": 7
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.5418999955058098,
      "aux_distill/mean_u": 0.10516872279061203,
      "aux_distill/n_active_tok": 58.45,
      "calib/answer_extract_rate": 0.1171875,
      "calib/auroc": 0.5238095238095238,
      "calib/avg_num_step_conf": 0.5703125,
      "calib/ece": 0.7068,
      "calib/final_conf_rate": 0.09765625,
      "calib/format_rate": 0.0703125,
      "calib/frac_conf_gt_0.9": 0.8,
      "calib/gap": 0.11690476190476173,
      "calib/mean_conf": 0.8667999999999999,
      "calib/mu_c": 0.965,
      "calib/mu_w": 0.8480952380952382,
      "calib/nonempty_final_conf_rate": 0.09765625,
      "calib/nonempty_reasoning_rate": 0.1484375,
      "calib/nonempty_step_conf_rate": 0.12109375,
      "calib/pce": 0.7068,
      "calib/std_conf": 0.2573980574907278,
      "calib/step_conf_rate": 0.12109375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2845.0,
      "completions/max_terminated_length": 2845.0,
      "completions/mean_length": 674.76171875,
      "completions/mean_terminated_length": 735.0595703125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.008533333333333334,
      "grad_norm": 0.012400745414197445,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0829,
      "num_tokens": 2196846.0,
      "reward": 0.0567142590880394,
      "reward_std": 0.1300020068883896,
      "rewards/accuracy_reward_step": 0.01953125,
      "rewards/final_brier_reward_step": 0.023584768176078796,
      "rewards/format_reward_step": 0.0703125,
      "step": 8
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.625441624359651,
      "aux_distill/mean_u": 0.05235877670716145,
      "aux_distill/n_active_tok": 34.18181818181818,
      "calib/answer_extract_rate": 0.08203125,
      "calib/auroc": 0.33333333333333337,
      "calib/avg_num_step_conf": 0.18359375,
      "calib/ece": 0.8238461538461539,
      "calib/final_conf_rate": 0.05078125,
      "calib/format_rate": 0.04296875,
      "calib/frac_conf_gt_0.9": 0.6923076923076923,
      "calib/gap": 0.02083333333333315,
      "calib/mean_conf": 0.900769230769231,
      "calib/mu_c": 0.92,
      "calib/mu_w": 0.8991666666666669,
      "calib/nonempty_final_conf_rate": 0.05078125,
      "calib/nonempty_reasoning_rate": 0.0859375,
      "calib/nonempty_step_conf_rate": 0.05078125,
      "calib/pce": 0.8238461538461539,
      "calib/std_conf": 0.12218910554870632,
      "calib/step_conf_rate": 0.05078125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3016.0,
      "completions/max_terminated_length": 3016.0,
      "completions/mean_length": 711.82421875,
      "completions/mean_terminated_length": 762.4560546875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 4.0,
      "epoch": 0.0096,
      "grad_norm": 0.011278004385530949,
      "learning_rate": 2.25e-06,
      "loss": 0.053,
      "num_tokens": 2486609.0,
      "reward": 0.02900156006217003,
      "reward_std": 0.062437769025564194,
      "rewards/accuracy_reward_step": 0.00390625,
      "rewards/final_brier_reward_step": 0.01112812478095293,
      "rewards/format_reward_step": 0.04296875,
      "step": 9
    },
    {
      "aux_distill/lambda": 0.1,
      "aux_distill/loss": 0.5412973931857518,
      "aux_distill/mean_u": 0.049759440712768645,
      "aux_distill/n_active_tok": 41.357142857142854,
      "calib/answer_extract_rate": 0.078125,
      "calib/avg_num_step_conf": 0.25,
      "calib/ece": 0.8383333333333334,
      "calib/final_conf_rate": 0.046875,
      "calib/format_rate": 0.0390625,
      "calib/frac_conf_gt_0.9": 0.75,
      "calib/mean_conf": 0.8383333333333334,
      "calib/mu_c": NaN,
      "calib/mu_w": 0.8383333333333334,
      "calib/nonempty_final_conf_rate": 0.046875,
      "calib/nonempty_reasoning_rate": 0.09765625,
      "calib/nonempty_step_conf_rate": 0.05859375,
      "calib/pce": 0.8383333333333334,
      "calib/std_conf": 0.25951343360647483,
      "calib/step_conf_rate": 0.05859375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2979.0,
      "completions/max_terminated_length": 2979.0,
      "completions/mean_length": 702.6953125,
      "completions/mean_terminated_length": 755.8403930664062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.010666666666666666,
      "grad_norm": 0.00804956629872322,
      "learning_rate": 2.5e-06,
      "loss": 0.0301,
      "num_tokens": 2773299.0,
      "reward": 0.024574805051088333,
      "reward_std": 0.05259711667895317,
      "rewards/accuracy_reward_step": 0.0,
      "rewards/final_brier_reward_step": 0.010087109170854092,
      "rewards/format_reward_step": 0.0390625,
      "step": 10
    },
    {
      "aux_distill/lambda": 0.1,
      "aux_distill/loss": 0.43450634820120676,
      "aux_distill/mean_u": 0.04201065652126385,
      "aux_distill/n_active_tok": 59.42857142857143,
      "calib/answer_extract_rate": 0.1171875,
      "calib/auroc": 0.49264705882352944,
      "calib/avg_num_step_conf": 0.421875,
      "calib/ece": 0.7297142857142858,
      "calib/final_conf_rate": 0.08203125,
      "calib/format_rate": 0.06640625,
      "calib/frac_conf_gt_0.9": 0.7142857142857143,
      "calib/gap": -0.02788235294117658,
      "calib/mean_conf": 0.8725714285714284,
      "calib/mu_c": 0.8499999999999999,
      "calib/mu_w": 0.8778823529411764,
      "calib/nonempty_final_conf_rate": 0.08203125,
      "calib/nonempty_reasoning_rate": 0.140625,
      "calib/nonempty_step_conf_rate": 0.09375,
      "calib/pce": 0.7059047619047619,
      "calib/std_conf": 0.2245436793712489,
      "calib/step_conf_rate": 0.09375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2973.0,
      "completions/max_terminated_length": 2973.0,
      "completions/mean_length": 684.203125,
      "completions/mean_terminated_length": 729.8167114257812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.011733333333333333,
      "grad_norm": 0.008146646432578564,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.067,
      "num_tokens": 3052935.0,
      "reward": 0.05362946167588234,
      "reward_std": 0.10116725414991379,
      "rewards/accuracy_reward_step": 0.015625,
      "rewards/final_brier_reward_step": 0.02522767148911953,
      "rewards/format_reward_step": 0.06640625,
      "step": 11
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.4705760436398642,
      "aux_distill/mean_u": 0.07098841757193944,
      "aux_distill/n_active_tok": 57.095238095238095,
      "calib/answer_extract_rate": 0.1640625,
      "calib/auroc": 0.6043956043956045,
      "calib/avg_num_step_conf": 0.56640625,
      "calib/ece": 0.506764705882353,
      "calib/final_conf_rate": 0.1328125,
      "calib/format_rate": 0.09765625,
      "calib/frac_conf_gt_0.9": 0.6470588235294118,
      "calib/gap": 0.07805860805860809,
      "calib/mean_conf": 0.8479411764705883,
      "calib/mu_c": 0.8961538461538462,
      "calib/mu_w": 0.8180952380952381,
      "calib/nonempty_final_conf_rate": 0.1328125,
      "calib/nonempty_reasoning_rate": 0.1796875,
      "calib/nonempty_step_conf_rate": 0.125,
      "calib/pce": 0.4861764705882353,
      "calib/std_conf": 0.2521761687178314,
      "calib/step_conf_rate": 0.125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.07421875,
      "completions/max_length": 3005.0,
      "completions/max_terminated_length": 3005.0,
      "completions/mean_length": 620.20703125,
      "completions/mean_terminated_length": 669.92822265625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 4.0,
      "epoch": 0.0128,
      "grad_norm": 0.010749871842563152,
      "learning_rate": 3e-06,
      "loss": 0.1102,
      "num_tokens": 3315884.0,
      "reward": 0.10153749585151672,
      "reward_std": 0.17852166295051575,
      "rewards/accuracy_reward_step": 0.05078125,
      "rewards/final_brier_reward_step": 0.054637499153614044,
      "rewards/format_reward_step": 0.09765625,
      "step": 12
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.4806722551584244,
      "aux_distill/mean_u": 0.06838403580369118,
      "aux_distill/n_active_tok": 65.14285714285714,
      "calib/answer_extract_rate": 0.15234375,
      "calib/auroc": 0.3920454545454546,
      "calib/avg_num_step_conf": 0.66796875,
      "calib/ece": 0.601,
      "calib/final_conf_rate": 0.1171875,
      "calib/format_rate": 0.1015625,
      "calib/frac_conf_gt_0.9": 0.6,
      "calib/gap": -0.028749999999999942,
      "calib/mean_conf": 0.8223333333333332,
      "calib/mu_c": 0.80125,
      "calib/mu_w": 0.83,
      "calib/nonempty_final_conf_rate": 0.1171875,
      "calib/nonempty_reasoning_rate": 0.203125,
      "calib/nonempty_step_conf_rate": 0.16015625,
      "calib/pce": 0.5783333333333333,
      "calib/std_conf": 0.26203922522316303,
      "calib/step_conf_rate": 0.16015625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.05859375,
      "completions/max_length": 3065.0,
      "completions/max_terminated_length": 3065.0,
      "completions/mean_length": 678.55859375,
      "completions/mean_terminated_length": 720.7925415039062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 4.0,
      "epoch": 0.013866666666666666,
      "grad_norm": 0.011776494793593884,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.1223,
      "num_tokens": 3594187.0,
      "reward": 0.08667441457509995,
      "reward_std": 0.17229244112968445,
      "rewards/accuracy_reward_step": 0.03125,
      "rewards/final_brier_reward_step": 0.04053632542490959,
      "rewards/format_reward_step": 0.1015625,
      "step": 13
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.45552198818096745,
      "aux_distill/mean_u": 0.052501629064482844,
      "aux_distill/n_active_tok": 59.42307692307692,
      "calib/answer_extract_rate": 0.19140625,
      "calib/auroc": 0.5651041666666667,
      "calib/avg_num_step_conf": 0.75390625,
      "calib/ece": 0.6221052631578947,
      "calib/final_conf_rate": 0.1484375,
      "calib/format_rate": 0.1171875,
      "calib/frac_conf_gt_0.9": 0.5789473684210527,
      "calib/gap": 0.1048958333333333,
      "calib/mean_conf": 0.78,
      "calib/mu_c": 0.8683333333333333,
      "calib/mu_w": 0.7634375,
      "calib/nonempty_final_conf_rate": 0.1484375,
      "calib/nonempty_reasoning_rate": 0.22265625,
      "calib/nonempty_step_conf_rate": 0.171875,
      "calib/pce": 0.6221052631578947,
      "calib/std_conf": 0.3280885631398295,
      "calib/step_conf_rate": 0.171875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3009.0,
      "completions/max_terminated_length": 3009.0,
      "completions/mean_length": 637.84765625,
      "completions/mean_terminated_length": 671.97119140625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.014933333333333333,
      "grad_norm": 0.010614952072501183,
      "learning_rate": 3.5e-06,
      "loss": 0.0726,
      "num_tokens": 3862876.0,
      "reward": 0.09558359533548355,
      "reward_std": 0.15353244543075562,
      "rewards/accuracy_reward_step": 0.02734375,
      "rewards/final_brier_reward_step": 0.0466359406709671,
      "rewards/format_reward_step": 0.1171875,
      "step": 14
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4393505273120744,
      "aux_distill/mean_u": 0.0932373114978294,
      "aux_distill/n_active_tok": 72.17857142857143,
      "calib/answer_extract_rate": 0.2265625,
      "calib/auroc": 0.45959595959595956,
      "calib/avg_num_step_conf": 0.98828125,
      "calib/ece": 0.6157446808510638,
      "calib/final_conf_rate": 0.18359375,
      "calib/format_rate": 0.12890625,
      "calib/frac_conf_gt_0.9": 0.5531914893617021,
      "calib/gap": -0.07212121212121203,
      "calib/mean_conf": 0.8097872340425533,
      "calib/mu_c": 0.7545454545454546,
      "calib/mu_w": 0.8266666666666667,
      "calib/nonempty_final_conf_rate": 0.18359375,
      "calib/nonempty_reasoning_rate": 0.2734375,
      "calib/nonempty_step_conf_rate": 0.203125,
      "calib/pce": 0.5957446808510639,
      "calib/std_conf": 0.21594413542456847,
      "calib/step_conf_rate": 0.203125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2660.0,
      "completions/max_terminated_length": 2660.0,
      "completions/mean_length": 580.484375,
      "completions/mean_terminated_length": 624.3865966796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.016,
      "grad_norm": 0.010281026363372803,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.0852,
      "num_tokens": 4119360.0,
      "reward": 0.11895253509283066,
      "reward_std": 0.19674092531204224,
      "rewards/accuracy_reward_step": 0.05078125,
      "rewards/final_brier_reward_step": 0.05821758136153221,
      "rewards/format_reward_step": 0.12890625,
      "step": 15
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.404011321755556,
      "aux_distill/mean_u": 0.1309152058316472,
      "aux_distill/n_active_tok": 102.11538461538461,
      "calib/answer_extract_rate": 0.234375,
      "calib/auroc": 0.4892344497607655,
      "calib/avg_num_step_conf": 1.2890625,
      "calib/ece": 0.6683673469387755,
      "calib/final_conf_rate": 0.19140625,
      "calib/format_rate": 0.15625,
      "calib/frac_conf_gt_0.9": 0.6326530612244898,
      "calib/gap": -0.07406698564593306,
      "calib/mean_conf": 0.8165306122448979,
      "calib/mu_c": 0.759090909090909,
      "calib/mu_w": 0.8331578947368421,
      "calib/nonempty_final_conf_rate": 0.19140625,
      "calib/nonempty_reasoning_rate": 0.2890625,
      "calib/nonempty_step_conf_rate": 0.23046875,
      "calib/pce": 0.6302040816326531,
      "calib/std_conf": 0.2564399203251316,
      "calib/step_conf_rate": 0.23046875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3064.0,
      "completions/max_terminated_length": 3064.0,
      "completions/mean_length": 669.92578125,
      "completions/mean_terminated_length": 726.6991577148438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.017066666666666667,
      "grad_norm": 0.009635679423809052,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.1339,
      "num_tokens": 4399709.0,
      "reward": 0.130092591047287,
      "reward_std": 0.24008221924304962,
      "rewards/accuracy_reward_step": 0.046875,
      "rewards/final_brier_reward_step": 0.057060156017541885,
      "rewards/format_reward_step": 0.15625,
      "step": 16
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.46860240880520115,
      "aux_distill/mean_u": 0.11381851062136948,
      "aux_distill/n_active_tok": 94.06451612903226,
      "calib/answer_extract_rate": 0.2890625,
      "calib/auroc": 0.36937590711175616,
      "calib/avg_num_step_conf": 1.453125,
      "calib/ece": 0.643478787878788,
      "calib/final_conf_rate": 0.2578125,
      "calib/format_rate": 0.21875,
      "calib/frac_conf_gt_0.9": 0.4696969696969697,
      "calib/gap": -0.1606316400580553,
      "calib/mean_conf": 0.7259151515151515,
      "calib/mu_c": 0.5969230769230769,
      "calib/mu_w": 0.7575547169811322,
      "calib/nonempty_final_conf_rate": 0.2578125,
      "calib/nonempty_reasoning_rate": 0.36328125,
      "calib/nonempty_step_conf_rate": 0.296875,
      "calib/pce": 0.5862121212121213,
      "calib/std_conf": 0.34071722223569406,
      "calib/step_conf_rate": 0.296875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3031.0,
      "completions/max_terminated_length": 3031.0,
      "completions/mean_length": 517.45703125,
      "completions/mean_terminated_length": 556.5924682617188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.018133333333333335,
      "grad_norm": 0.013831501826643944,
      "learning_rate": 4.25e-06,
      "loss": 0.1361,
      "num_tokens": 4635706.0,
      "reward": 0.1753574013710022,
      "reward_std": 0.2824881672859192,
      "rewards/accuracy_reward_step": 0.05078125,
      "rewards/final_brier_reward_step": 0.08118356764316559,
      "rewards/format_reward_step": 0.21875,
      "step": 17
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36261486814868066,
      "aux_distill/mean_u": 0.09953009255725216,
      "aux_distill/n_active_tok": 94.64516129032258,
      "calib/answer_extract_rate": 0.31640625,
      "calib/auroc": 0.6474137931034483,
      "calib/avg_num_step_conf": 1.44921875,
      "calib/ece": 0.42455882352941177,
      "calib/final_conf_rate": 0.265625,
      "calib/format_rate": 0.21875,
      "calib/frac_conf_gt_0.9": 0.3235294117647059,
      "calib/gap": 0.2293448275862069,
      "calib/mean_conf": 0.5383823529411763,
      "calib/mu_c": 0.734,
      "calib/mu_w": 0.5046551724137931,
      "calib/nonempty_final_conf_rate": 0.265625,
      "calib/nonempty_reasoning_rate": 0.37890625,
      "calib/nonempty_step_conf_rate": 0.2890625,
      "calib/pce": 0.40794117647058825,
      "calib/std_conf": 0.39705019053795754,
      "calib/step_conf_rate": 0.2890625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.05859375,
      "completions/max_length": 2844.0,
      "completions/max_terminated_length": 2844.0,
      "completions/mean_length": 541.7109375,
      "completions/mean_terminated_length": 575.4274291992188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.0192,
      "grad_norm": 0.014123844914138317,
      "learning_rate": 4.5e-06,
      "loss": 0.1028,
      "num_tokens": 4885104.0,
      "reward": 0.19429530203342438,
      "reward_std": 0.3070595860481262,
      "rewards/accuracy_reward_step": 0.0390625,
      "rewards/final_brier_reward_step": 0.13077811896800995,
      "rewards/format_reward_step": 0.21875,
      "step": 18
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3752623365726322,
      "aux_distill/mean_u": 0.10054922271830097,
      "aux_distill/n_active_tok": 219.3125,
      "calib/answer_extract_rate": 0.51953125,
      "calib/auroc": 0.5744301994301995,
      "calib/avg_num_step_conf": 3.453125,
      "calib/ece": 0.346603305785124,
      "calib/final_conf_rate": 0.47265625,
      "calib/format_rate": 0.40625,
      "calib/frac_conf_gt_0.9": 0.2396694214876033,
      "calib/gap": 0.10920726495726507,
      "calib/mean_conf": 0.39637190082644624,
      "calib/mu_c": 0.49384615384615393,
      "calib/mu_w": 0.38463888888888886,
      "calib/nonempty_final_conf_rate": 0.47265625,
      "calib/nonempty_reasoning_rate": 0.6875,
      "calib/nonempty_step_conf_rate": 0.58984375,
      "calib/pce": 0.3177685950413224,
      "calib/std_conf": 0.40467524235171576,
      "calib/step_conf_rate": 0.58984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2929.0,
      "completions/max_terminated_length": 2929.0,
      "completions/mean_length": 457.26953125,
      "completions/mean_terminated_length": 470.1244812011719,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.020266666666666665,
      "grad_norm": 0.013274524360895157,
      "learning_rate": 4.75e-06,
      "loss": 0.1721,
      "num_tokens": 5106925.0,
      "reward": 0.3702906370162964,
      "reward_std": 0.429590106010437,
      "rewards/accuracy_reward_step": 0.05078125,
      "rewards/final_brier_reward_step": 0.2835499942302704,
      "rewards/format_reward_step": 0.40625,
      "step": 19
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36305526876822114,
      "aux_distill/mean_u": 0.08904974974893316,
      "aux_distill/n_active_tok": 238.84375,
      "calib/answer_extract_rate": 0.66796875,
      "calib/auroc": 0.4521428571428571,
      "calib/avg_num_step_conf": 3.734375,
      "calib/ece": 0.32581699346405235,
      "calib/final_conf_rate": 0.59765625,
      "calib/format_rate": 0.5,
      "calib/frac_conf_gt_0.9": 0.16339869281045752,
      "calib/gap": -0.03130857142857146,
      "calib/mean_conf": 0.3191503267973856,
      "calib/mu_c": 0.29357142857142854,
      "calib/mu_w": 0.32488,
      "calib/nonempty_final_conf_rate": 0.59765625,
      "calib/nonempty_reasoning_rate": 0.859375,
      "calib/nonempty_step_conf_rate": 0.734375,
      "calib/pce": 0.23098039215686278,
      "calib/std_conf": 0.3729548987301565,
      "calib/step_conf_rate": 0.734375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2900.0,
      "completions/max_terminated_length": 2900.0,
      "completions/mean_length": 361.8671875,
      "completions/mean_terminated_length": 373.5403137207031,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.021333333333333333,
      "grad_norm": 0.017645180225372314,
      "learning_rate": 5e-06,
      "loss": 0.1372,
      "num_tokens": 5304435.0,
      "reward": 0.4850044846534729,
      "reward_std": 0.4788362383842468,
      "rewards/accuracy_reward_step": 0.12890625,
      "rewards/final_brier_reward_step": 0.3411027193069458,
      "rewards/format_reward_step": 0.5,
      "step": 20
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3560049566440284,
      "aux_distill/mean_u": 0.07944683044637237,
      "aux_distill/n_active_tok": 284.84375,
      "calib/answer_extract_rate": 0.7890625,
      "calib/auroc": 0.43239113827349124,
      "calib/avg_num_step_conf": 4.3828125,
      "calib/ece": 0.3241648936170212,
      "calib/final_conf_rate": 0.734375,
      "calib/format_rate": 0.671875,
      "calib/frac_conf_gt_0.9": 0.11702127659574468,
      "calib/gap": -0.09717761650114587,
      "calib/mean_conf": 0.26725,
      "calib/mu_c": 0.18764705882352944,
      "calib/mu_w": 0.2848246753246753,
      "calib/nonempty_final_conf_rate": 0.734375,
      "calib/nonempty_reasoning_rate": 0.9375,
      "calib/nonempty_step_conf_rate": 0.87109375,
      "calib/pce": 0.20528191489361697,
      "calib/std_conf": 0.3527064602108503,
      "calib/step_conf_rate": 0.87109375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2802.0,
      "completions/max_terminated_length": 2802.0,
      "completions/mean_length": 342.6796875,
      "completions/mean_terminated_length": 346.74310302734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 24.0,
      "epoch": 0.0224,
      "grad_norm": 0.015373860485851765,
      "learning_rate": 4.9722222222222224e-06,
      "loss": 0.1704,
      "num_tokens": 5495121.0,
      "reward": 0.64105623960495,
      "reward_std": 0.4382461607456207,
      "rewards/accuracy_reward_step": 0.14453125,
      "rewards/final_brier_reward_step": 0.4657062888145447,
      "rewards/format_reward_step": 0.671875,
      "step": 21
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3883116855286062,
      "aux_distill/mean_u": 0.095899419626436,
      "aux_distill/n_active_tok": 275.0625,
      "calib/answer_extract_rate": 0.78125,
      "calib/auroc": 0.5436607142857143,
      "calib/avg_num_step_conf": 4.16796875,
      "calib/ece": 0.2713825641025641,
      "calib/final_conf_rate": 0.76171875,
      "calib/format_rate": 0.6796875,
      "calib/frac_conf_gt_0.9": 0.14358974358974358,
      "calib/gap": 0.08447071428571423,
      "calib/mean_conf": 0.2749764102564103,
      "calib/mu_c": 0.34428571428571425,
      "calib/mu_w": 0.259815,
      "calib/nonempty_final_conf_rate": 0.76171875,
      "calib/nonempty_reasoning_rate": 0.9296875,
      "calib/nonempty_step_conf_rate": 0.87109375,
      "calib/pce": 0.18343589743589744,
      "calib/std_conf": 0.3698652772847972,
      "calib/step_conf_rate": 0.87109375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2997.0,
      "completions/max_terminated_length": 2997.0,
      "completions/mean_length": 328.2890625,
      "completions/mean_terminated_length": 330.8740234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 34.0,
      "epoch": 0.023466666666666667,
      "grad_norm": 0.016197707504034042,
      "learning_rate": 4.944444444444445e-06,
      "loss": 0.1975,
      "num_tokens": 5680979.0,
      "reward": 0.6641405820846558,
      "reward_std": 0.45880556106567383,
      "rewards/accuracy_reward_step": 0.13671875,
      "rewards/final_brier_reward_step": 0.5118749141693115,
      "rewards/format_reward_step": 0.6796875,
      "step": 22
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37039018934592605,
      "aux_distill/mean_u": 0.07887884510197941,
      "aux_distill/n_active_tok": 290.09375,
      "calib/answer_extract_rate": 0.76953125,
      "calib/auroc": 0.5085059171597632,
      "calib/avg_num_step_conf": 4.359375,
      "calib/ece": 0.19217616580310878,
      "calib/final_conf_rate": 0.75390625,
      "calib/format_rate": 0.69140625,
      "calib/frac_conf_gt_0.9": 0.06217616580310881,
      "calib/gap": 0.019437869822485215,
      "calib/mean_conf": 0.13797927461139897,
      "calib/mu_c": 0.155,
      "calib/mu_w": 0.13556213017751478,
      "calib/nonempty_final_conf_rate": 0.75390625,
      "calib/nonempty_reasoning_rate": 0.953125,
      "calib/nonempty_step_conf_rate": 0.921875,
      "calib/pce": 0.10290155440414508,
      "calib/std_conf": 0.2753284606220166,
      "calib/step_conf_rate": 0.921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2716.0,
      "completions/max_terminated_length": 2716.0,
      "completions/mean_length": 339.54296875,
      "completions/mean_terminated_length": 340.8745422363281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 34.0,
      "epoch": 0.024533333333333334,
      "grad_norm": 0.013020611368119717,
      "learning_rate": 4.9166666666666665e-06,
      "loss": 0.1838,
      "num_tokens": 5871838.0,
      "reward": 0.6801663637161255,
      "reward_std": 0.4263819456100464,
      "rewards/accuracy_reward_step": 0.10546875,
      "rewards/final_brier_reward_step": 0.5634578466415405,
      "rewards/format_reward_step": 0.69140625,
      "step": 23
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3795403325930238,
      "aux_distill/mean_u": 0.0993213219486208,
      "aux_distill/n_active_tok": 288.4375,
      "calib/answer_extract_rate": 0.86328125,
      "calib/auroc": 0.42058823529411765,
      "calib/avg_num_step_conf": 4.515625,
      "calib/ece": 0.22866036866359452,
      "calib/final_conf_rate": 0.84765625,
      "calib/format_rate": 0.78125,
      "calib/frac_conf_gt_0.9": 0.06912442396313365,
      "calib/gap": -0.0714048128342246,
      "calib/mean_conf": 0.1655331797235023,
      "calib/mu_c": 0.104,
      "calib/mu_w": 0.1754048128342246,
      "calib/nonempty_final_conf_rate": 0.84765625,
      "calib/nonempty_reasoning_rate": 0.97265625,
      "calib/nonempty_step_conf_rate": 0.9375,
      "calib/pce": 0.12797235023041476,
      "calib/std_conf": 0.2959538720526848,
      "calib/step_conf_rate": 0.9375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2398.0,
      "completions/max_terminated_length": 2398.0,
      "completions/mean_length": 293.7265625,
      "completions/mean_terminated_length": 297.2095031738281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 35.0,
      "epoch": 0.0256,
      "grad_norm": 0.012711533345282078,
      "learning_rate": 4.888888888888889e-06,
      "loss": 0.1428,
      "num_tokens": 6051544.0,
      "reward": 0.7570474147796631,
      "reward_std": 0.35965046286582947,
      "rewards/accuracy_reward_step": 0.125,
      "rewards/final_brier_reward_step": 0.607844889163971,
      "rewards/format_reward_step": 0.78125,
      "step": 24
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34714192571118474,
      "aux_distill/mean_u": 0.09413421996178056,
      "aux_distill/n_active_tok": 284.34375,
      "calib/answer_extract_rate": 0.90625,
      "calib/auroc": 0.5238383838383839,
      "calib/avg_num_step_conf": 4.3828125,
      "calib/ece": 0.1623762331838565,
      "calib/final_conf_rate": 0.87109375,
      "calib/format_rate": 0.81640625,
      "calib/frac_conf_gt_0.9": 0.026905829596412557,
      "calib/gap": 0.012165151515151504,
      "calib/mean_conf": 0.12479865470852018,
      "calib/mu_c": 0.1356,
      "calib/mu_w": 0.1234348484848485,
      "calib/nonempty_final_conf_rate": 0.87109375,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.94140625,
      "calib/pce": 0.08753363228699551,
      "calib/std_conf": 0.23934835832936371,
      "calib/step_conf_rate": 0.94140625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2226.0,
      "completions/max_terminated_length": 2226.0,
      "completions/mean_length": 271.77734375,
      "completions/mean_terminated_length": 273.9173278808594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 24.0,
      "epoch": 0.02666666666666667,
      "grad_norm": 0.01145833171904087,
      "learning_rate": 4.861111111111111e-06,
      "loss": 0.1131,
      "num_tokens": 6224343.0,
      "reward": 0.8035451173782349,
      "reward_std": 0.3558991849422455,
      "rewards/accuracy_reward_step": 0.10546875,
      "rewards/final_brier_reward_step": 0.6852152347564697,
      "rewards/format_reward_step": 0.81640625,
      "step": 25
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3926220517605543,
      "aux_distill/mean_u": 0.11919352905940954,
      "aux_distill/n_active_tok": 289.59375,
      "calib/answer_extract_rate": 0.93359375,
      "calib/auroc": 0.43568840579710144,
      "calib/avg_num_step_conf": 4.484375,
      "calib/ece": 0.15626017316017315,
      "calib/final_conf_rate": 0.90234375,
      "calib/format_rate": 0.875,
      "calib/frac_conf_gt_0.9": 0.030303030303030304,
      "calib/gap": -0.061667391304347834,
      "calib/mean_conf": 0.09317705627705629,
      "calib/mu_c": 0.03791666666666666,
      "calib/mu_w": 0.0995840579710145,
      "calib/nonempty_final_conf_rate": 0.90234375,
      "calib/nonempty_reasoning_rate": 0.98046875,
      "calib/nonempty_step_conf_rate": 0.953125,
      "calib/pce": 0.07277056277056278,
      "calib/std_conf": 0.21375208328797457,
      "calib/step_conf_rate": 0.953125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2267.0,
      "completions/max_terminated_length": 2267.0,
      "completions/mean_length": 266.21484375,
      "completions/mean_terminated_length": 266.21484375,
      "completions/min_length": 20.0,
      "completions/min_terminated_length": 20.0,
      "epoch": 0.027733333333333332,
      "grad_norm": 0.011086120270192623,
      "learning_rate": 4.833333333333333e-06,
      "loss": 0.1295,
      "num_tokens": 6397734.0,
      "reward": 0.8561491966247559,
      "reward_std": 0.2689933180809021,
      "rewards/accuracy_reward_step": 0.09765625,
      "rewards/final_brier_reward_step": 0.7396421432495117,
      "rewards/format_reward_step": 0.875,
      "step": 26
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3392999805510044,
      "aux_distill/mean_u": 0.08765503889892604,
      "aux_distill/n_active_tok": 276.25,
      "calib/answer_extract_rate": 0.953125,
      "calib/auroc": 0.5318181818181819,
      "calib/avg_num_step_conf": 4.28125,
      "calib/ece": 0.1106625,
      "calib/final_conf_rate": 0.9375,
      "calib/format_rate": 0.90234375,
      "calib/frac_conf_gt_0.9": 0.016666666666666666,
      "calib/gap": -0.03391363636363635,
      "calib/mean_conf": 0.07008750000000001,
      "calib/mu_c": 0.039,
      "calib/mu_w": 0.07291363636363635,
      "calib/nonempty_final_conf_rate": 0.9375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9765625,
      "calib/pce": 0.04870833333333333,
      "calib/std_conf": 0.1805147635063404,
      "calib/step_conf_rate": 0.9765625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1882.0,
      "completions/max_terminated_length": 1882.0,
      "completions/mean_length": 233.49609375,
      "completions/mean_terminated_length": 234.41177368164062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 25.0,
      "epoch": 0.0288,
      "grad_norm": 0.01071285642683506,
      "learning_rate": 4.805555555555556e-06,
      "loss": 0.0644,
      "num_tokens": 6562725.0,
      "reward": 0.8917468786239624,
      "reward_std": 0.2230653464794159,
      "rewards/accuracy_reward_step": 0.08203125,
      "rewards/final_brier_reward_step": 0.7991188168525696,
      "rewards/format_reward_step": 0.90234375,
      "step": 27
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3293367293663323,
      "aux_distill/mean_u": 0.08486807697068835,
      "aux_distill/n_active_tok": 266.28125,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.4142295597484277,
      "calib/avg_num_step_conf": 4.203125,
      "calib/ece": 0.14791818181818184,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.9140625,
      "calib/frac_conf_gt_0.9": 0.024793388429752067,
      "calib/gap": 0.01924622641509434,
      "calib/mean_conf": 0.0851396694214876,
      "calib/mu_c": 0.10200000000000001,
      "calib/mu_w": 0.08275377358490567,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.054545454545454536,
      "calib/std_conf": 0.20000299294427465,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1156.0,
      "completions/max_terminated_length": 1156.0,
      "completions/mean_length": 224.82421875,
      "completions/mean_terminated_length": 225.70590209960938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.029866666666666666,
      "grad_norm": 0.010027695447206497,
      "learning_rate": 4.777777777777778e-06,
      "loss": 0.0176,
      "num_tokens": 6727224.0,
      "reward": 0.9075891971588135,
      "reward_std": 0.22711993753910065,
      "rewards/accuracy_reward_step": 0.12109375,
      "rewards/final_brier_reward_step": 0.7800223231315613,
      "rewards/format_reward_step": 0.9140625,
      "step": 28
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33028204878792167,
      "aux_distill/mean_u": 0.09123819166510017,
      "aux_distill/n_active_tok": 288.34375,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.477989352989353,
      "calib/avg_num_step_conf": 4.4609375,
      "calib/ece": 0.08967540983606559,
      "calib/final_conf_rate": 0.953125,
      "calib/format_rate": 0.9296875,
      "calib/frac_conf_gt_0.9": 0.00819672131147541,
      "calib/gap": 0.04145323505323507,
      "calib/mean_conf": 0.04819344262295082,
      "calib/mu_c": 0.08590909090909092,
      "calib/mu_w": 0.044455855855855854,
      "calib/nonempty_final_conf_rate": 0.953125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.023852459016393445,
      "calib/std_conf": 0.13082393858854915,
      "calib/step_conf_rate": 0.98828125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2240.0,
      "completions/max_terminated_length": 2240.0,
      "completions/mean_length": 240.50390625,
      "completions/mean_terminated_length": 241.4470672607422,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 56.0,
      "epoch": 0.030933333333333334,
      "grad_norm": 0.0092391986399889,
      "learning_rate": 4.75e-06,
      "loss": 0.0501,
      "num_tokens": 6895921.0,
      "reward": 0.9342766404151917,
      "reward_std": 0.15640118718147278,
      "rewards/accuracy_reward_step": 0.09765625,
      "rewards/final_brier_reward_step": 0.8412095904350281,
      "rewards/format_reward_step": 0.9296875,
      "step": 29
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3043099972419441,
      "aux_distill/mean_u": 0.07361558707575262,
      "aux_distill/n_active_tok": 300.25,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.5543650793650794,
      "calib/avg_num_step_conf": 4.703125,
      "calib/ece": 0.11186788617886177,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.93359375,
      "calib/frac_conf_gt_0.9": 0.0040650406504065045,
      "calib/gap": 0.04093412698412698,
      "calib/mean_conf": 0.055611788617886175,
      "calib/mu_c": 0.09055555555555556,
      "calib/mu_w": 0.04962142857142857,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.010569105691056912,
      "calib/std_conf": 0.12115992749981405,
      "calib/step_conf_rate": 0.98828125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 768.0,
      "completions/max_terminated_length": 768.0,
      "completions/mean_length": 233.4140625,
      "completions/mean_terminated_length": 234.3294219970703,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 43.0,
      "epoch": 0.032,
      "grad_norm": 0.00907308142632246,
      "learning_rate": 4.722222222222222e-06,
      "loss": 0.0428,
      "num_tokens": 7062659.0,
      "reward": 0.9381251335144043,
      "reward_std": 0.18297097086906433,
      "rewards/accuracy_reward_step": 0.140625,
      "rewards/final_brier_reward_step": 0.8020316362380981,
      "rewards/format_reward_step": 0.93359375,
      "step": 30
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.30257489811629057,
      "aux_distill/mean_u": 0.06740870386265194,
      "aux_distill/n_active_tok": 296.875,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.45075974770642196,
      "calib/avg_num_step_conf": 4.6875,
      "calib/ece": 0.12351999999999999,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.012,
      "calib/gap": -0.014423738532110097,
      "calib/mean_conf": 0.06664,
      "calib/mu_c": 0.0540625,
      "calib/mu_w": 0.0684862385321101,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.031079999999999997,
      "calib/std_conf": 0.14081729439241475,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 739.0,
      "completions/max_terminated_length": 739.0,
      "completions/mean_length": 222.49609375,
      "completions/mean_terminated_length": 223.36863708496094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 42.0,
      "epoch": 0.03306666666666667,
      "grad_norm": 0.009486602619290352,
      "learning_rate": 4.694444444444445e-06,
      "loss": 0.0173,
      "num_tokens": 7225530.0,
      "reward": 0.9653845429420471,
      "reward_std": 0.11241470277309418,
      "rewards/accuracy_reward_step": 0.125,
      "rewards/final_brier_reward_step": 0.837019145488739,
      "rewards/format_reward_step": 0.96875,
      "step": 31
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32631619879975915,
      "aux_distill/mean_u": 0.08648384814445001,
      "aux_distill/n_active_tok": 294.1875,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.5512232415902141,
      "calib/avg_num_step_conf": 4.63671875,
      "calib/ece": 0.13063266932270917,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.01195219123505976,
      "calib/gap": -0.00442585487906591,
      "calib/mean_conf": 0.07566215139442233,
      "calib/mu_c": 0.07181818181818181,
      "calib/mu_w": 0.07624403669724772,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03741035856573705,
      "calib/std_conf": 0.16396980567488864,
      "calib/step_conf_rate": 1.0,
      "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": 540.0,
      "completions/max_terminated_length": 540.0,
      "completions/mean_length": 213.46875,
      "completions/mean_terminated_length": 214.30589294433594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 36.0,
      "epoch": 0.034133333333333335,
      "grad_norm": 0.009173893369734287,
      "learning_rate": 4.666666666666667e-06,
      "loss": 0.0242,
      "num_tokens": 7386882.0,
      "reward": 0.9543381929397583,
      "reward_std": 0.13994824886322021,
      "rewards/accuracy_reward_step": 0.12890625,
      "rewards/final_brier_reward_step": 0.8188326954841614,
      "rewards/format_reward_step": 0.9609375,
      "step": 32
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3093249648809433,
      "aux_distill/mean_u": 0.07432332380150891,
      "aux_distill/n_active_tok": 329.1875,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5285885989010989,
      "calib/avg_num_step_conf": 5.13671875,
      "calib/ece": 0.09990800000000001,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.003763392857142854,
      "calib/mean_conf": 0.063372,
      "calib/mu_c": 0.060000000000000005,
      "calib/mu_w": 0.06376339285714286,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.02964,
      "calib/std_conf": 0.13201373267959662,
      "calib/step_conf_rate": 1.0,
      "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": 2364.0,
      "completions/max_terminated_length": 2364.0,
      "completions/mean_length": 252.53125,
      "completions/mean_terminated_length": 253.52159118652344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 78.0,
      "epoch": 0.0352,
      "grad_norm": 0.008318406529724598,
      "learning_rate": 4.638888888888889e-06,
      "loss": 0.0882,
      "num_tokens": 7558402.0,
      "reward": 0.9643739461898804,
      "reward_std": 0.1018943265080452,
      "rewards/accuracy_reward_step": 0.1015625,
      "rewards/final_brier_reward_step": 0.8584355115890503,
      "rewards/format_reward_step": 0.96875,
      "step": 33
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2938751680776477,
      "aux_distill/mean_u": 0.08131013044518337,
      "aux_distill/n_active_tok": 315.78125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.47363569321533927,
      "calib/avg_num_step_conf": 5.0078125,
      "calib/ece": 0.07644000000000001,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.004,
      "calib/gap": 0.03626474926253687,
      "calib/mean_conf": 0.0518,
      "calib/mu_c": 0.08458333333333334,
      "calib/mu_w": 0.04831858407079647,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.016120000000000002,
      "calib/std_conf": 0.09212795449807837,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 842.0,
      "completions/max_terminated_length": 842.0,
      "completions/mean_length": 219.78125,
      "completions/mean_terminated_length": 220.6431427001953,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 61.0,
      "epoch": 0.03626666666666667,
      "grad_norm": 0.009772083722054958,
      "learning_rate": 4.611111111111112e-06,
      "loss": 0.0595,
      "num_tokens": 7719778.0,
      "reward": 0.9713041186332703,
      "reward_std": 0.1016630083322525,
      "rewards/accuracy_reward_step": 0.09375,
      "rewards/final_brier_reward_step": 0.8801082372665405,
      "rewards/format_reward_step": 0.96875,
      "step": 34
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31735930871218443,
      "aux_distill/mean_u": 0.0893676437073024,
      "aux_distill/n_active_tok": 315.375,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.37512800819252434,
      "calib/avg_num_step_conf": 4.91796875,
      "calib/ece": 0.13023715415019763,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0259536610343062,
      "calib/mean_conf": 0.05031620553359684,
      "calib/mu_c": 0.028055555555555556,
      "calib/mu_w": 0.054009216589861755,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.019130434782608695,
      "calib/std_conf": 0.0948636363430505,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1565.0,
      "completions/max_terminated_length": 1565.0,
      "completions/mean_length": 234.33984375,
      "completions/mean_terminated_length": 234.33984375,
      "completions/min_length": 78.0,
      "completions/min_terminated_length": 78.0,
      "epoch": 0.037333333333333336,
      "grad_norm": 0.008290213532745838,
      "learning_rate": 4.583333333333333e-06,
      "loss": 0.0982,
      "num_tokens": 7889025.0,
      "reward": 0.9709469079971313,
      "reward_std": 0.07739226520061493,
      "rewards/accuracy_reward_step": 0.140625,
      "rewards/final_brier_reward_step": 0.8286125063896179,
      "rewards/format_reward_step": 0.97265625,
      "step": 35
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32667525624856353,
      "aux_distill/mean_u": 0.0935250968970002,
      "aux_distill/n_active_tok": 312.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4863888106636628,
      "calib/avg_num_step_conf": 4.90625,
      "calib/ece": 0.2004330708661417,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.015177884164085234,
      "calib/mean_conf": 0.043897637795275585,
      "calib/mu_c": 0.031886792452830194,
      "calib/mu_w": 0.04706467661691543,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.017834645669291337,
      "calib/std_conf": 0.09072848953508278,
      "calib/step_conf_rate": 1.0,
      "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": 1743.0,
      "completions/max_terminated_length": 1743.0,
      "completions/mean_length": 227.453125,
      "completions/mean_terminated_length": 227.453125,
      "completions/min_length": 51.0,
      "completions/min_terminated_length": 51.0,
      "epoch": 0.0384,
      "grad_norm": 0.008621211163699627,
      "learning_rate": 4.555555555555556e-06,
      "loss": 0.0632,
      "num_tokens": 8049965.0,
      "reward": 0.9937493801116943,
      "reward_std": 0.04545110464096069,
      "rewards/accuracy_reward_step": 0.20703125,
      "rewards/final_brier_reward_step": 0.7882800698280334,
      "rewards/format_reward_step": 0.9921875,
      "step": 36
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.27830373821780086,
      "aux_distill/mean_u": 0.06014688570484389,
      "aux_distill/n_active_tok": 271.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5403703703703703,
      "calib/avg_num_step_conf": 4.24609375,
      "calib/ece": 0.0991764705882353,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00017777777777778017,
      "calib/mean_conf": 0.029843137254901966,
      "calib/mu_c": 0.030000000000000006,
      "calib/mu_w": 0.029822222222222226,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.005686274509803923,
      "calib/std_conf": 0.04472985253502062,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 748.0,
      "completions/max_terminated_length": 748.0,
      "completions/mean_length": 196.78515625,
      "completions/mean_terminated_length": 197.55686950683594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 54.0,
      "epoch": 0.039466666666666664,
      "grad_norm": 0.009745295159518719,
      "learning_rate": 4.527777777777778e-06,
      "loss": 0.0497,
      "num_tokens": 8207438.0,
      "reward": 0.9884085655212402,
      "reward_std": 0.047429949045181274,
      "rewards/accuracy_reward_step": 0.1171875,
      "rewards/final_brier_reward_step": 0.8752546906471252,
      "rewards/format_reward_step": 0.984375,
      "step": 37
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34966559894382954,
      "aux_distill/mean_u": 0.09391435826298852,
      "aux_distill/n_active_tok": 261.375,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5124125479575716,
      "calib/avg_num_step_conf": 4.08203125,
      "calib/ece": 0.15146245059288538,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.004239449334236058,
      "calib/mean_conf": 0.029011857707509883,
      "calib/mu_c": 0.02547619047619048,
      "calib/mu_w": 0.029715639810426537,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0072332015810276686,
      "calib/std_conf": 0.0670983280726559,
      "calib/step_conf_rate": 1.0,
      "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": 1823.0,
      "completions/max_terminated_length": 1823.0,
      "completions/mean_length": 193.234375,
      "completions/mean_terminated_length": 193.234375,
      "completions/min_length": 48.0,
      "completions/min_terminated_length": 48.0,
      "epoch": 0.04053333333333333,
      "grad_norm": 0.010434729978442192,
      "learning_rate": 4.5e-06,
      "loss": 0.0401,
      "num_tokens": 8363794.0,
      "reward": 0.9839221239089966,
      "reward_std": 0.06251996010541916,
      "rewards/accuracy_reward_step": 0.1640625,
      "rewards/final_brier_reward_step": 0.8233128786087036,
      "rewards/format_reward_step": 0.98046875,
      "step": 38
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32260613050311804,
      "aux_distill/mean_u": 0.07570191593785686,
      "aux_distill/n_active_tok": 233.0625,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.43281970046082946,
      "calib/avg_num_step_conf": 3.609375,
      "calib/ece": 0.10807725490196078,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.00392156862745098,
      "calib/gap": -0.008174063940092163,
      "calib/mean_conf": 0.02556745098039215,
      "calib/mu_c": 0.01838709677419355,
      "calib/mu_w": 0.026561160714285714,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.006038039215686275,
      "calib/std_conf": 0.06936021292239343,
      "calib/step_conf_rate": 1.0,
      "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": 2179.0,
      "completions/max_terminated_length": 2179.0,
      "completions/mean_length": 181.72265625,
      "completions/mean_terminated_length": 181.72265625,
      "completions/min_length": 49.0,
      "completions/min_terminated_length": 49.0,
      "epoch": 0.0416,
      "grad_norm": 0.008618353866040707,
      "learning_rate": 4.472222222222223e-06,
      "loss": 0.0462,
      "num_tokens": 8516403.0,
      "reward": 0.9838817119598389,
      "reward_std": 0.053902022540569305,
      "rewards/accuracy_reward_step": 0.12109375,
      "rewards/final_brier_reward_step": 0.8622946739196777,
      "rewards/format_reward_step": 0.984375,
      "step": 39
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31119333766400814,
      "aux_distill/mean_u": 0.0803044466646787,
      "aux_distill/n_active_tok": 235.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.43339945071711933,
      "calib/avg_num_step_conf": 3.70703125,
      "calib/ece": 0.10207843137254902,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.012967653341470858,
      "calib/mean_conf": 0.025058823529411765,
      "calib/mu_c": 0.036551724137931035,
      "calib/mu_w": 0.023584070796460178,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.006705882352941176,
      "calib/std_conf": 0.0748272289133522,
      "calib/step_conf_rate": 1.0,
      "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": 628.0,
      "completions/max_terminated_length": 628.0,
      "completions/mean_length": 176.61328125,
      "completions/mean_terminated_length": 177.30589294433594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 50.0,
      "epoch": 0.042666666666666665,
      "grad_norm": 0.008439777418971062,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.0166,
      "num_tokens": 8668376.0,
      "reward": 0.9945722222328186,
      "reward_std": 0.02659289352595806,
      "rewards/accuracy_reward_step": 0.11328125,
      "rewards/final_brier_reward_step": 0.883675754070282,
      "rewards/format_reward_step": 0.9921875,
      "step": 40
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3391628819517791,
      "aux_distill/mean_u": 0.09605003539857801,
      "aux_distill/n_active_tok": 230.78125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5066517895459617,
      "calib/avg_num_step_conf": 3.609375,
      "calib/ece": 0.23242063492063492,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0034065745429576865,
      "calib/mean_conf": 0.0175,
      "calib/mu_c": 0.014918032786885247,
      "calib/mu_w": 0.018324607329842934,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.003928571428571429,
      "calib/std_conf": 0.03396572735745012,
      "calib/step_conf_rate": 1.0,
      "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": 516.0,
      "completions/max_terminated_length": 516.0,
      "completions/mean_length": 171.4296875,
      "completions/mean_terminated_length": 172.1019744873047,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 42.0,
      "epoch": 0.04373333333333333,
      "grad_norm": 0.009633166715502739,
      "learning_rate": 4.416666666666667e-06,
      "loss": 0.0363,
      "num_tokens": 8819510.0,
      "reward": 0.9852579832077026,
      "reward_std": 0.056308452039957047,
      "rewards/accuracy_reward_step": 0.23828125,
      "rewards/final_brier_reward_step": 0.75176602602005,
      "rewards/format_reward_step": 0.98046875,
      "step": 41
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32519798586145043,
      "aux_distill/mean_u": 0.08523037325117232,
      "aux_distill/n_active_tok": 212.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4938311688311688,
      "calib/avg_num_step_conf": 3.34765625,
      "calib/ece": 0.1280392156862745,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0016428571428571438,
      "calib/mean_conf": 0.012274509803921568,
      "calib/mu_c": 0.010857142857142857,
      "calib/mu_w": 0.0125,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0015294117647058822,
      "calib/std_conf": 0.022200581542791604,
      "calib/step_conf_rate": 1.0,
      "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": 401.0,
      "completions/max_terminated_length": 401.0,
      "completions/mean_length": 152.13671875,
      "completions/mean_terminated_length": 152.73333740234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 51.0,
      "epoch": 0.0448,
      "grad_norm": 0.00855032168328762,
      "learning_rate": 4.388888888888889e-06,
      "loss": 0.023,
      "num_tokens": 8962825.0,
      "reward": 0.9972575902938843,
      "reward_std": 0.015008872374892235,
      "rewards/accuracy_reward_step": 0.13671875,
      "rewards/final_brier_reward_step": 0.8617027401924133,
      "rewards/format_reward_step": 0.99609375,
      "step": 42
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3105672858655453,
      "aux_distill/mean_u": 0.08021462575046566,
      "aux_distill/n_active_tok": 229.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4931987713909609,
      "calib/avg_num_step_conf": 3.58984375,
      "calib/ece": 0.15682352941176472,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.007620666959192628,
      "calib/mean_conf": 0.011803921568627451,
      "calib/mu_c": 0.018139534883720932,
      "calib/mu_w": 0.010518867924528304,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.023777580153468786,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 385.0,
      "completions/max_terminated_length": 385.0,
      "completions/mean_length": 167.234375,
      "completions/mean_terminated_length": 167.8902130126953,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.04586666666666667,
      "grad_norm": 0.008470780216157436,
      "learning_rate": 4.361111111111112e-06,
      "loss": 0.0409,
      "num_tokens": 9110861.0,
      "reward": 0.9948850870132446,
      "reward_std": 0.02814546786248684,
      "rewards/accuracy_reward_step": 0.16796875,
      "rewards/final_brier_reward_step": 0.8296140432357788,
      "rewards/format_reward_step": 0.9921875,
      "step": 43
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3330973004922271,
      "aux_distill/mean_u": 0.09241892886200444,
      "aux_distill/n_active_tok": 237.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5287971457696229,
      "calib/avg_num_step_conf": 3.71484375,
      "calib/ece": 0.13338582677165353,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.002250254841997963,
      "calib/mean_conf": 0.008346456692913387,
      "calib/mu_c": 0.01027777777777778,
      "calib/mu_w": 0.008027522935779817,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.012344387581723506,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2142.0,
      "completions/max_terminated_length": 2142.0,
      "completions/mean_length": 181.2265625,
      "completions/mean_terminated_length": 181.2265625,
      "completions/min_length": 51.0,
      "completions/min_terminated_length": 51.0,
      "epoch": 0.046933333333333334,
      "grad_norm": 0.007521971128880978,
      "learning_rate": 4.333333333333334e-06,
      "loss": 0.0436,
      "num_tokens": 9263575.0,
      "reward": 0.9896165728569031,
      "reward_std": 0.02841258980333805,
      "rewards/accuracy_reward_step": 0.140625,
      "rewards/final_brier_reward_step": 0.8503269553184509,
      "rewards/format_reward_step": 0.98828125,
      "step": 44
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3044560784474015,
      "aux_distill/mean_u": 0.0746863951893971,
      "aux_distill/n_active_tok": 214.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.507775403608737,
      "calib/avg_num_step_conf": 3.359375,
      "calib/ece": 0.14403921568627454,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00038817663817663876,
      "calib/mean_conf": 0.008901960784313726,
      "calib/mu_c": 0.009230769230769232,
      "calib/mu_w": 0.008842592592592593,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.014347582842075764,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2297.0,
      "completions/max_terminated_length": 2297.0,
      "completions/mean_length": 171.30078125,
      "completions/mean_terminated_length": 171.30078125,
      "completions/min_length": 45.0,
      "completions/min_terminated_length": 45.0,
      "epoch": 0.048,
      "grad_norm": 0.007751181721687317,
      "learning_rate": 4.305555555555556e-06,
      "loss": 0.0843,
      "num_tokens": 9412476.0,
      "reward": 0.9895457029342651,
      "reward_std": 0.03634835034608841,
      "rewards/accuracy_reward_step": 0.15234375,
      "rewards/final_brier_reward_step": 0.8384664058685303,
      "rewards/format_reward_step": 0.98828125,
      "step": 45
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3505524811334908,
      "aux_distill/mean_u": 0.11305923610502411,
      "aux_distill/n_active_tok": 223.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5502726952850104,
      "calib/avg_num_step_conf": 3.515625,
      "calib/ece": 0.2225984251968504,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0014883884588318087,
      "calib/mean_conf": 0.005748031496062992,
      "calib/mu_c": 0.006896551724137932,
      "calib/mu_w": 0.005408163265306123,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.010311654215518736,
      "calib/step_conf_rate": 1.0,
      "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": 603.0,
      "completions/max_terminated_length": 603.0,
      "completions/mean_length": 162.703125,
      "completions/mean_terminated_length": 163.3411865234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 47.0,
      "epoch": 0.04906666666666667,
      "grad_norm": 0.0071896458975970745,
      "learning_rate": 4.277777777777778e-06,
      "loss": 0.0083,
      "num_tokens": 9558896.0,
      "reward": 0.9936808347702026,
      "reward_std": 0.017596811056137085,
      "rewards/accuracy_reward_step": 0.2265625,
      "rewards/final_brier_reward_step": 0.7686116695404053,
      "rewards/format_reward_step": 0.9921875,
      "step": 46
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3146101264283061,
      "aux_distill/mean_u": 0.08592837101606565,
      "aux_distill/n_active_tok": 228.375,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5255847953216374,
      "calib/avg_num_step_conf": 3.578125,
      "calib/ece": 0.14346456692913387,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0036086744639376245,
      "calib/mean_conf": 0.006141732283464567,
      "calib/mu_c": 0.009210526315789476,
      "calib/mu_w": 0.005601851851851852,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.009683758957676255,
      "calib/step_conf_rate": 1.0,
      "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": 479.0,
      "completions/max_terminated_length": 479.0,
      "completions/mean_length": 163.14453125,
      "completions/mean_terminated_length": 163.78431701660156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 51.0,
      "epoch": 0.050133333333333335,
      "grad_norm": 0.006908603478223085,
      "learning_rate": 4.25e-06,
      "loss": 0.0494,
      "num_tokens": 9706637.0,
      "reward": 0.9954426288604736,
      "reward_std": 0.01957503706216812,
      "rewards/accuracy_reward_step": 0.15234375,
      "rewards/final_brier_reward_step": 0.8463538885116577,
      "rewards/format_reward_step": 0.9921875,
      "step": 47
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3150333953090012,
      "aux_distill/mean_u": 0.08773427079714621,
      "aux_distill/n_active_tok": 217.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5326086956521738,
      "calib/avg_num_step_conf": 3.4140625,
      "calib/ece": 0.18262745098039218,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0014824879227053143,
      "calib/mean_conf": 0.008588235294117647,
      "calib/mu_c": 0.009791666666666667,
      "calib/mu_w": 0.008309178743961353,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0014901960784313726,
      "calib/std_conf": 0.021263221004214024,
      "calib/step_conf_rate": 1.0,
      "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": 444.0,
      "completions/max_terminated_length": 444.0,
      "completions/mean_length": 158.1796875,
      "completions/mean_terminated_length": 158.8000030517578,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 51.0,
      "epoch": 0.0512,
      "grad_norm": 0.0069527393206954,
      "learning_rate": 4.222222222222223e-06,
      "loss": 0.022,
      "num_tokens": 9850819.0,
      "reward": 0.9976677298545837,
      "reward_std": 0.01528632640838623,
      "rewards/accuracy_reward_step": 0.1875,
      "rewards/final_brier_reward_step": 0.811741828918457,
      "rewards/format_reward_step": 0.99609375,
      "step": 48
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32575283851474524,
      "aux_distill/mean_u": 0.08622161203094775,
      "aux_distill/n_active_tok": 232.0,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5482870370370371,
      "calib/avg_num_step_conf": 3.62890625,
      "calib/ece": 0.20685039370078742,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.003048148148148149,
      "calib/mean_conf": 0.005748031496062992,
      "calib/mu_c": 0.00814814814814815,
      "calib/mu_w": 0.0051,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.011434144545147353,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2156.0,
      "completions/max_terminated_length": 2156.0,
      "completions/mean_length": 182.9765625,
      "completions/mean_terminated_length": 182.9765625,
      "completions/min_length": 46.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.05226666666666667,
      "grad_norm": 0.005945167969912291,
      "learning_rate": 4.194444444444445e-06,
      "loss": 0.0707,
      "num_tokens": 10002197.0,
      "reward": 0.9899195432662964,
      "reward_std": 0.0371810719370842,
      "rewards/accuracy_reward_step": 0.2109375,
      "rewards/final_brier_reward_step": 0.7806203365325928,
      "rewards/format_reward_step": 0.98828125,
      "step": 49
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3211708148010075,
      "aux_distill/mean_u": 0.08810421113629638,
      "aux_distill/n_active_tok": 236.5,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.49433962264150944,
      "calib/avg_num_step_conf": 3.6953125,
      "calib/ece": 0.20478260869565218,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0007330188679245285,
      "calib/mean_conf": 0.004703557312252965,
      "calib/mu_c": 0.005283018867924529,
      "calib/mu_w": 0.00455,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.009176091721549455,
      "calib/step_conf_rate": 1.0,
      "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": 2891.0,
      "completions/max_terminated_length": 2891.0,
      "completions/mean_length": 203.78125,
      "completions/mean_terminated_length": 203.78125,
      "completions/min_length": 52.0,
      "completions/min_terminated_length": 52.0,
      "epoch": 0.05333333333333334,
      "grad_norm": 0.004969415720552206,
      "learning_rate": 4.166666666666667e-06,
      "loss": 0.0815,
      "num_tokens": 10159725.0,
      "reward": 0.9893224239349365,
      "reward_std": 0.028084326535463333,
      "rewards/accuracy_reward_step": 0.20703125,
      "rewards/final_brier_reward_step": 0.7833324074745178,
      "rewards/format_reward_step": 0.98828125,
      "step": 50
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.29707011906430125,
      "aux_distill/mean_u": 0.07452163606529925,
      "aux_distill/n_active_tok": 241.5,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5441661813240148,
      "calib/avg_num_step_conf": 3.7734375,
      "calib/ece": 0.19521739130434784,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0013618714812657744,
      "calib/mean_conf": 0.006363636363636363,
      "calib/mu_c": 0.007450980392156864,
      "calib/mu_w": 0.006089108910891089,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.010862879726242943,
      "calib/step_conf_rate": 1.0,
      "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": 2042.0,
      "completions/max_terminated_length": 2042.0,
      "completions/mean_length": 193.67578125,
      "completions/mean_terminated_length": 193.67578125,
      "completions/min_length": 54.0,
      "completions/min_terminated_length": 54.0,
      "epoch": 0.0544,
      "grad_norm": 0.006089591421186924,
      "learning_rate": 4.138888888888889e-06,
      "loss": 0.0822,
      "num_tokens": 10318602.0,
      "reward": 0.9896872639656067,
      "reward_std": 0.03538576513528824,
      "rewards/accuracy_reward_step": 0.19921875,
      "rewards/final_brier_reward_step": 0.7918745279312134,
      "rewards/format_reward_step": 0.98828125,
      "step": 51
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3404705720022321,
      "aux_distill/mean_u": 0.1005887897967428,
      "aux_distill/n_active_tok": 212.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5432149037972106,
      "calib/avg_num_step_conf": 3.3203125,
      "calib/ece": 0.18183593750000002,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0019372900335946248,
      "calib/mean_conf": 0.0075390625,
      "calib/mu_c": 0.005957446808510639,
      "calib/mu_w": 0.007894736842105263,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.002890625,
      "calib/std_conf": 0.04034759176978341,
      "calib/step_conf_rate": 1.0,
      "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": 520.0,
      "completions/max_terminated_length": 520.0,
      "completions/mean_length": 170.72265625,
      "completions/mean_terminated_length": 171.3921661376953,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 50.0,
      "epoch": 0.055466666666666664,
      "grad_norm": 0.006997254677116871,
      "learning_rate": 4.111111111111111e-06,
      "loss": 0.0022,
      "num_tokens": 10470259.0,
      "reward": 1.000251293182373,
      "reward_std": 0.004490233957767487,
      "rewards/accuracy_reward_step": 0.18359375,
      "rewards/final_brier_reward_step": 0.8169089555740356,
      "rewards/format_reward_step": 1.0,
      "step": 52
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3243614980019629,
      "aux_distill/mean_u": 0.09754663784882035,
      "aux_distill/n_active_tok": 236.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.49855324074074076,
      "calib/avg_num_step_conf": 3.703125,
      "calib/ece": 0.24121568627450982,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0013045634920634932,
      "calib/mean_conf": 0.0058431372549019615,
      "calib/mu_c": 0.0068253968253968265,
      "calib/mu_w": 0.005520833333333333,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.011130901253117435,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 466.0,
      "completions/max_terminated_length": 466.0,
      "completions/mean_length": 184.984375,
      "completions/mean_terminated_length": 185.70980834960938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 54.0,
      "epoch": 0.05653333333333333,
      "grad_norm": 0.00639935489743948,
      "learning_rate": 4.083333333333334e-06,
      "loss": 0.0273,
      "num_tokens": 10623439.0,
      "reward": 0.989882230758667,
      "reward_std": 0.029333556070923805,
      "rewards/accuracy_reward_step": 0.24609375,
      "rewards/final_brier_reward_step": 0.745389461517334,
      "rewards/format_reward_step": 0.98828125,
      "step": 53
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.30648385314270854,
      "aux_distill/mean_u": 0.08786370281314095,
      "aux_distill/n_active_tok": 262.0,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.4715535227090355,
      "calib/avg_num_step_conf": 4.1015625,
      "calib/ece": 0.2612648221343874,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0015904349221633764,
      "calib/mean_conf": 0.0035573122529644276,
      "calib/mu_c": 0.0023880597014925378,
      "calib/mu_w": 0.003978494623655914,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.008340201932212852,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3049.0,
      "completions/max_terminated_length": 3049.0,
      "completions/mean_length": 218.15625,
      "completions/mean_terminated_length": 218.15625,
      "completions/min_length": 55.0,
      "completions/min_terminated_length": 55.0,
      "epoch": 0.0576,
      "grad_norm": 0.005304010584950447,
      "learning_rate": 4.055555555555556e-06,
      "loss": 0.0789,
      "num_tokens": 10785519.0,
      "reward": 0.9830062389373779,
      "reward_std": 0.046094685792922974,
      "rewards/accuracy_reward_step": 0.26171875,
      "rewards/final_brier_reward_step": 0.7238249778747559,
      "rewards/format_reward_step": 0.98046875,
      "step": 54
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3295765253715217,
      "aux_distill/mean_u": 0.09773883690950656,
      "aux_distill/n_active_tok": 266.1875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4909261067708333,
      "calib/avg_num_step_conf": 4.15234375,
      "calib/ece": 0.2450390625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.000572916666666666,
      "calib/mean_conf": 0.0049609375,
      "calib/mu_c": 0.0045312500000000006,
      "calib/mu_w": 0.005104166666666667,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.009682379569150022,
      "calib/step_conf_rate": 1.0,
      "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": 596.0,
      "completions/max_terminated_length": 596.0,
      "completions/mean_length": 201.58203125,
      "completions/mean_terminated_length": 202.37255859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 52.0,
      "epoch": 0.058666666666666666,
      "grad_norm": 0.0058689904399216175,
      "learning_rate": 4.027777777777779e-06,
      "loss": 0.0203,
      "num_tokens": 10944948.0,
      "reward": 0.9913079738616943,
      "reward_std": 0.029984353110194206,
      "rewards/accuracy_reward_step": 0.25,
      "rewards/final_brier_reward_step": 0.7443346977233887,
      "rewards/format_reward_step": 0.98828125,
      "step": 55
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3124923906289041,
      "aux_distill/mean_u": 0.09151797595988466,
      "aux_distill/n_active_tok": 287.15625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.47722727272727267,
      "calib/avg_num_step_conf": 4.515625,
      "calib/ece": 0.21137254901960786,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0015590909090909093,
      "calib/mean_conf": 0.004313725490196079,
      "calib/mu_c": 0.003090909090909091,
      "calib/mu_w": 0.0046500000000000005,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.008779410396207511,
      "calib/step_conf_rate": 1.0,
      "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": 696.0,
      "completions/max_terminated_length": 696.0,
      "completions/mean_length": 211.5234375,
      "completions/mean_terminated_length": 212.3529510498047,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 59.0,
      "epoch": 0.05973333333333333,
      "grad_norm": 0.0049371616914868355,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0203,
      "num_tokens": 11105938.0,
      "reward": 0.9928038716316223,
      "reward_std": 0.023647766560316086,
      "rewards/accuracy_reward_step": 0.21484375,
      "rewards/final_brier_reward_step": 0.7785765528678894,
      "rewards/format_reward_step": 0.9921875,
      "step": 56
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2937448946759105,
      "aux_distill/mean_u": 0.0755081459411437,
      "aux_distill/n_active_tok": 311.5625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.46857605862044677,
      "calib/avg_num_step_conf": 4.86328125,
      "calib/ece": 0.3214173228346457,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0011526808990347352,
      "calib/mean_conf": 0.005354330708661418,
      "calib/mu_c": 0.004578313253012049,
      "calib/mu_w": 0.0057309941520467846,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.011891588742458496,
      "calib/step_conf_rate": 1.0,
      "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": 2507.0,
      "completions/max_terminated_length": 2507.0,
      "completions/mean_length": 255.3125,
      "completions/mean_terminated_length": 255.3125,
      "completions/min_length": 63.0,
      "completions/min_terminated_length": 63.0,
      "epoch": 0.0608,
      "grad_norm": 0.0056275613605976105,
      "learning_rate": 3.972222222222223e-06,
      "loss": 0.0815,
      "num_tokens": 11278090.0,
      "reward": 0.9896812438964844,
      "reward_std": 0.036557819694280624,
      "rewards/accuracy_reward_step": 0.32421875,
      "rewards/final_brier_reward_step": 0.6668624877929688,
      "rewards/format_reward_step": 0.98828125,
      "step": 57
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2732988477218896,
      "aux_distill/mean_u": 0.0651957247729289,
      "aux_distill/n_active_tok": 315.46875,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5159128978224456,
      "calib/avg_num_step_conf": 4.93359375,
      "calib/ece": 0.218671875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.000488406947015782,
      "calib/mean_conf": 0.0052343749999999994,
      "calib/mu_c": 0.0056140350877192996,
      "calib/mu_w": 0.0051256281407035175,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.000625,
      "calib/std_conf": 0.013108015424135531,
      "calib/step_conf_rate": 0.98828125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 910.0,
      "completions/max_terminated_length": 910.0,
      "completions/mean_length": 246.265625,
      "completions/mean_terminated_length": 247.23138427734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.06186666666666667,
      "grad_norm": 0.004538458306342363,
      "learning_rate": 3.944444444444445e-06,
      "loss": 0.0216,
      "num_tokens": 11447454.0,
      "reward": 0.9894318580627441,
      "reward_std": 0.03593946248292923,
      "rewards/accuracy_reward_step": 0.22265625,
      "rewards/final_brier_reward_step": 0.7679261565208435,
      "rewards/format_reward_step": 0.98828125,
      "step": 58
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.26986421295441687,
      "aux_distill/mean_u": 0.07765561929139546,
      "aux_distill/n_active_tok": 341.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5320283473235172,
      "calib/avg_num_step_conf": 5.38671875,
      "calib/ece": 0.33254901960784317,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0018473923214531451,
      "calib/mean_conf": 0.004705882352941177,
      "calib/mu_c": 0.005930232558139535,
      "calib/mu_w": 0.00408284023668639,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.010655559384193809,
      "calib/step_conf_rate": 1.0,
      "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": 788.0,
      "completions/max_terminated_length": 788.0,
      "completions/mean_length": 256.7109375,
      "completions/mean_terminated_length": 257.7176513671875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 78.0,
      "epoch": 0.06293333333333333,
      "grad_norm": 0.005329262465238571,
      "learning_rate": 3.916666666666667e-06,
      "loss": 0.0169,
      "num_tokens": 11619420.0,
      "reward": 0.9980183839797974,
      "reward_std": 0.015194371342658997,
      "rewards/accuracy_reward_step": 0.3359375,
      "rewards/final_brier_reward_step": 0.66400545835495,
      "rewards/format_reward_step": 0.99609375,
      "step": 59
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3008674378506839,
      "aux_distill/mean_u": 0.07826646190527813,
      "aux_distill/n_active_tok": 353.53125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.540191652807476,
      "calib/avg_num_step_conf": 5.5234375,
      "calib/ece": 0.26999999999999996,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0014255167498218105,
      "calib/mean_conf": 0.004761904761904762,
      "calib/mu_c": 0.005797101449275362,
      "calib/mu_w": 0.004371584699453552,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.00047619047619047614,
      "calib/std_conf": 0.011106575037800574,
      "calib/step_conf_rate": 1.0,
      "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": 1197.0,
      "completions/max_terminated_length": 1197.0,
      "completions/mean_length": 269.203125,
      "completions/mean_terminated_length": 270.25885009765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 107.0,
      "epoch": 0.064,
      "grad_norm": 0.005264188162982464,
      "learning_rate": 3.88888888888889e-06,
      "loss": 0.017,
      "num_tokens": 11797192.0,
      "reward": 0.9858656525611877,
      "reward_std": 0.04708104953169823,
      "rewards/accuracy_reward_step": 0.26953125,
      "rewards/final_brier_reward_step": 0.7178249359130859,
      "rewards/format_reward_step": 0.984375,
      "step": 60
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.28734378842636943,
      "aux_distill/mean_u": 0.08205361872395535,
      "aux_distill/n_active_tok": 350.25,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4551039697542534,
      "calib/avg_num_step_conf": 5.47265625,
      "calib/ece": 0.35873517786561265,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0012111801242236028,
      "calib/mean_conf": 0.004901185770750989,
      "calib/mu_c": 0.004130434782608695,
      "calib/mu_w": 0.005341614906832298,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.008081240083161419,
      "calib/step_conf_rate": 1.0,
      "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": 2514.0,
      "completions/max_terminated_length": 2514.0,
      "completions/mean_length": 263.265625,
      "completions/mean_terminated_length": 263.265625,
      "completions/min_length": 90.0,
      "completions/min_terminated_length": 90.0,
      "epoch": 0.06506666666666666,
      "grad_norm": 0.00512561434879899,
      "learning_rate": 3.861111111111112e-06,
      "loss": 0.085,
      "num_tokens": 11968652.0,
      "reward": 0.989721417427063,
      "reward_std": 0.03596086427569389,
      "rewards/accuracy_reward_step": 0.359375,
      "rewards/final_brier_reward_step": 0.6317867040634155,
      "rewards/format_reward_step": 0.98828125,
      "step": 61
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.26067466707900167,
      "aux_distill/mean_u": 0.06894861401193524,
      "aux_distill/n_active_tok": 385.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.45507317073170733,
      "calib/avg_num_step_conf": 6.11328125,
      "calib/ece": 0.1912156862745098,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0018195121951219517,
      "calib/mean_conf": 0.004862745098039216,
      "calib/mu_c": 0.0034,
      "calib/mu_w": 0.0052195121951219515,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.008157805375397733,
      "calib/step_conf_rate": 1.0,
      "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": 1024.0,
      "completions/max_terminated_length": 1024.0,
      "completions/mean_length": 294.1484375,
      "completions/mean_terminated_length": 295.3019714355469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.06613333333333334,
      "grad_norm": 0.004156938754022121,
      "learning_rate": 3.833333333333334e-06,
      "loss": -0.0044,
      "num_tokens": 12151034.0,
      "reward": 0.996712863445282,
      "reward_std": 0.01257465872913599,
      "rewards/accuracy_reward_step": 0.1953125,
      "rewards/final_brier_reward_step": 0.8020195364952087,
      "rewards/format_reward_step": 0.99609375,
      "step": 62
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2901492426171899,
      "aux_distill/mean_u": 0.08588818541921382,
      "aux_distill/n_active_tok": 386.03125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4864945213624394,
      "calib/avg_num_step_conf": 6.109375,
      "calib/ece": 0.2366535433070866,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0013751804977490865,
      "calib/mean_conf": 0.0035039370078740156,
      "calib/mu_c": 0.002459016393442623,
      "calib/mu_w": 0.0038341968911917096,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.008780233720409064,
      "calib/step_conf_rate": 1.0,
      "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": 883.0,
      "completions/max_terminated_length": 883.0,
      "completions/mean_length": 292.73046875,
      "completions/mean_terminated_length": 293.8784484863281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 96.0,
      "epoch": 0.0672,
      "grad_norm": 0.004469459876418114,
      "learning_rate": 3.8055555555555556e-06,
      "loss": 0.0327,
      "num_tokens": 12334613.0,
      "reward": 0.9927290678024292,
      "reward_std": 0.023673495277762413,
      "rewards/accuracy_reward_step": 0.23828125,
      "rewards/final_brier_reward_step": 0.7549893856048584,
      "rewards/format_reward_step": 0.9921875,
      "step": 63
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2550811097025871,
      "aux_distill/mean_u": 0.07274823291578757,
      "aux_distill/n_active_tok": 428.1875,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5149980522010129,
      "calib/avg_num_step_conf": 6.66015625,
      "calib/ece": 0.39948616600790515,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00024737047136735467,
      "calib/mean_conf": 0.003675889328063241,
      "calib/mu_c": 0.003823529411764706,
      "calib/mu_w": 0.0035761589403973514,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.006677619669182257,
      "calib/step_conf_rate": 1.0,
      "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": 2485.0,
      "completions/max_terminated_length": 2485.0,
      "completions/mean_length": 324.765625,
      "completions/mean_terminated_length": 324.765625,
      "completions/min_length": 92.0,
      "completions/min_terminated_length": 92.0,
      "epoch": 0.06826666666666667,
      "grad_norm": 0.004222342744469643,
      "learning_rate": 3.777777777777778e-06,
      "loss": 0.0683,
      "num_tokens": 12521529.0,
      "reward": 0.9897758960723877,
      "reward_std": 0.035435110330581665,
      "rewards/accuracy_reward_step": 0.3984375,
      "rewards/final_brier_reward_step": 0.5928331613540649,
      "rewards/format_reward_step": 0.98828125,
      "step": 64
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25398184289224446,
      "aux_distill/mean_u": 0.07289780750063732,
      "aux_distill/n_active_tok": 377.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5176613066002826,
      "calib/avg_num_step_conf": 5.94140625,
      "calib/ece": 0.3434375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.000594765525129516,
      "calib/mean_conf": 0.00421875,
      "calib/mu_c": 0.004606741573033708,
      "calib/mu_w": 0.004011976047904192,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.008441839162025061,
      "calib/step_conf_rate": 1.0,
      "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": 908.0,
      "completions/max_terminated_length": 908.0,
      "completions/mean_length": 286.390625,
      "completions/mean_terminated_length": 287.51373291015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 86.0,
      "epoch": 0.06933333333333333,
      "grad_norm": 0.004306511953473091,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.0023,
      "num_tokens": 12699869.0,
      "reward": 0.9976509809494019,
      "reward_std": 0.013895140960812569,
      "rewards/accuracy_reward_step": 0.34765625,
      "rewards/final_brier_reward_step": 0.6515519618988037,
      "rewards/format_reward_step": 0.99609375,
      "step": 65
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2754818422254175,
      "aux_distill/mean_u": 0.07254377648620623,
      "aux_distill/n_active_tok": 409.5625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4582591093117409,
      "calib/avg_num_step_conf": 6.41015625,
      "calib/ece": 0.25023529411764706,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0011012145748987854,
      "calib/mean_conf": 0.004666666666666666,
      "calib/mu_c": 0.0038461538461538464,
      "calib/mu_w": 0.004947368421052632,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.009403100587231573,
      "calib/step_conf_rate": 1.0,
      "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": 1115.0,
      "completions/max_terminated_length": 1115.0,
      "completions/mean_length": 323.73828125,
      "completions/mean_terminated_length": 325.00787353515625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.0704,
      "grad_norm": 0.00394570454955101,
      "learning_rate": 3.7222222222222225e-06,
      "loss": 0.033,
      "num_tokens": 12889098.0,
      "reward": 0.9931092262268066,
      "reward_std": 0.024321211501955986,
      "rewards/accuracy_reward_step": 0.25390625,
      "rewards/final_brier_reward_step": 0.7401245832443237,
      "rewards/format_reward_step": 0.9921875,
      "step": 66
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2856973884627223,
      "aux_distill/mean_u": 0.08087531891203306,
      "aux_distill/n_active_tok": 387.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4714991181657849,
      "calib/avg_num_step_conf": 6.0546875,
      "calib/ece": 0.311796875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00024126984126984167,
      "calib/mean_conf": 0.004609375,
      "calib/mu_c": 0.0044444444444444444,
      "calib/mu_w": 0.004685714285714286,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.008874326008738637,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 774.0,
      "completions/max_terminated_length": 774.0,
      "completions/mean_length": 308.9609375,
      "completions/mean_terminated_length": 310.1725769042969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 98.0,
      "epoch": 0.07146666666666666,
      "grad_norm": 0.004880924243479967,
      "learning_rate": 3.694444444444445e-06,
      "loss": 0.0259,
      "num_tokens": 13073200.0,
      "reward": 0.9935446977615356,
      "reward_std": 0.02437354251742363,
      "rewards/accuracy_reward_step": 0.31640625,
      "rewards/final_brier_reward_step": 0.6784957051277161,
      "rewards/format_reward_step": 0.9921875,
      "step": 67
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2604788551107049,
      "aux_distill/mean_u": 0.06340758781191137,
      "aux_distill/n_active_tok": 377.90625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.49754325259515564,
      "calib/avg_num_step_conf": 5.98046875,
      "calib/ece": 0.3287058823529412,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.000647058823529412,
      "calib/mean_conf": 0.004627450980392157,
      "calib/mu_c": 0.005058823529411765,
      "calib/mu_w": 0.004411764705882353,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.007705258668569009,
      "calib/step_conf_rate": 1.0,
      "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": 1293.0,
      "completions/max_terminated_length": 1293.0,
      "completions/mean_length": 294.73046875,
      "completions/mean_terminated_length": 295.88629150390625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 98.0,
      "epoch": 0.07253333333333334,
      "grad_norm": 0.004593390505760908,
      "learning_rate": 3.6666666666666666e-06,
      "loss": -0.0051,
      "num_tokens": 13252739.0,
      "reward": 0.9938269853591919,
      "reward_std": 0.02526906132698059,
      "rewards/accuracy_reward_step": 0.33203125,
      "rewards/final_brier_reward_step": 0.6634351015090942,
      "rewards/format_reward_step": 0.9921875,
      "step": 68
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.27652850397862494,
      "aux_distill/mean_u": 0.07943788932679391,
      "aux_distill/n_active_tok": 385.375,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4839379370629371,
      "calib/avg_num_step_conf": 6.01953125,
      "calib/ece": 0.30555118110236223,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.003438228438228437,
      "calib/mean_conf": 0.007125984251968504,
      "calib/mu_c": 0.004743589743589744,
      "calib/mu_w": 0.00818181818181818,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0027952755905511807,
      "calib/std_conf": 0.04486207947935929,
      "calib/step_conf_rate": 1.0,
      "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": 2156.0,
      "completions/max_terminated_length": 2156.0,
      "completions/mean_length": 346.54296875,
      "completions/mean_terminated_length": 346.54296875,
      "completions/min_length": 101.0,
      "completions/min_terminated_length": 101.0,
      "epoch": 0.0736,
      "grad_norm": 0.005000604782253504,
      "learning_rate": 3.638888888888889e-06,
      "loss": 0.0754,
      "num_tokens": 13445950.0,
      "reward": 0.9926091432571411,
      "reward_std": 0.027545515447854996,
      "rewards/accuracy_reward_step": 0.3046875,
      "rewards/final_brier_reward_step": 0.6883434057235718,
      "rewards/format_reward_step": 0.9921875,
      "step": 69
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.27339917747303843,
      "aux_distill/mean_u": 0.07388013492360007,
      "aux_distill/n_active_tok": 392.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.510620479634564,
      "calib/avg_num_step_conf": 6.12890625,
      "calib/ece": 0.2723828125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0009311001141987059,
      "calib/mean_conf": 0.0049609375,
      "calib/mu_c": 0.005633802816901409,
      "calib/mu_w": 0.004702702702702703,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.008052466958708602,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1021.0,
      "completions/max_terminated_length": 1021.0,
      "completions/mean_length": 331.7734375,
      "completions/mean_terminated_length": 333.07452392578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 96.0,
      "epoch": 0.07466666666666667,
      "grad_norm": 0.004021927248686552,
      "learning_rate": 3.6111111111111115e-06,
      "loss": 0.031,
      "num_tokens": 13637876.0,
      "reward": 0.9976115226745605,
      "reward_std": 0.014393772929906845,
      "rewards/accuracy_reward_step": 0.27734375,
      "rewards/final_brier_reward_step": 0.7217855453491211,
      "rewards/format_reward_step": 0.99609375,
      "step": 70
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2658953149802983,
      "aux_distill/mean_u": 0.0833407176913205,
      "aux_distill/n_active_tok": 422.9375,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.48202838827838834,
      "calib/avg_num_step_conf": 6.6484375,
      "calib/ece": 0.27669291338582674,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0011156898656898657,
      "calib/mean_conf": 0.0067716535433070876,
      "calib/mu_c": 0.0059722222222222225,
      "calib/mu_w": 0.007087912087912088,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.008500290848546832,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1047.0,
      "completions/max_terminated_length": 1047.0,
      "completions/mean_length": 354.65234375,
      "completions/mean_terminated_length": 356.04315185546875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 104.0,
      "epoch": 0.07573333333333333,
      "grad_norm": 0.0043555935844779015,
      "learning_rate": 3.5833333333333335e-06,
      "loss": 0.0338,
      "num_tokens": 13833075.0,
      "reward": 0.9899024963378906,
      "reward_std": 0.03595929220318794,
      "rewards/accuracy_reward_step": 0.28125,
      "rewards/final_brier_reward_step": 0.710273802280426,
      "rewards/format_reward_step": 0.98828125,
      "step": 71
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2881365141365677,
      "aux_distill/mean_u": 0.08016156748594756,
      "aux_distill/n_active_tok": 470.875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.48798913800306365,
      "calib/avg_num_step_conf": 7.28125,
      "calib/ece": 0.3341501976284585,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00046302743350508293,
      "calib/mean_conf": 0.005770750988142293,
      "calib/mu_c": 0.005465116279069767,
      "calib/mu_w": 0.00592814371257485,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.007995726021686987,
      "calib/step_conf_rate": 1.0,
      "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": 1039.0,
      "completions/max_terminated_length": 1039.0,
      "completions/mean_length": 360.87109375,
      "completions/mean_terminated_length": 362.2862854003906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 116.0,
      "epoch": 0.0768,
      "grad_norm": 0.004437967669218779,
      "learning_rate": 3.555555555555556e-06,
      "loss": 0.0204,
      "num_tokens": 14029866.0,
      "reward": 0.9920222163200378,
      "reward_std": 0.030886013060808182,
      "rewards/accuracy_reward_step": 0.33984375,
      "rewards/final_brier_reward_step": 0.6559194922447205,
      "rewards/format_reward_step": 0.98828125,
      "step": 72
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2541104690171778,
      "aux_distill/mean_u": 0.061887756224744496,
      "aux_distill/n_active_tok": 391.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5053516317552039,
      "calib/avg_num_step_conf": 6.109375,
      "calib/ece": 0.3766141732283465,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0009481909514741623,
      "calib/mean_conf": 0.006771653543307087,
      "calib/mu_c": 0.006185567010309278,
      "calib/mu_w": 0.00713375796178344,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0007480314960629921,
      "calib/std_conf": 0.01374010107607492,
      "calib/step_conf_rate": 1.0,
      "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": 2095.0,
      "completions/max_terminated_length": 2095.0,
      "completions/mean_length": 340.58203125,
      "completions/mean_terminated_length": 340.58203125,
      "completions/min_length": 87.0,
      "completions/min_terminated_length": 87.0,
      "epoch": 0.07786666666666667,
      "grad_norm": 0.004330141935497522,
      "learning_rate": 3.5277777777777784e-06,
      "loss": 0.0147,
      "num_tokens": 14224087.0,
      "reward": 0.9905085563659668,
      "reward_std": 0.03687606751918793,
      "rewards/accuracy_reward_step": 0.37890625,
      "rewards/final_brier_reward_step": 0.6138296127319336,
      "rewards/format_reward_step": 0.98828125,
      "step": 73
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2642949102446437,
      "aux_distill/mean_u": 0.06173665004754539,
      "aux_distill/n_active_tok": 387.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.512218045112782,
      "calib/avg_num_step_conf": 6.05078125,
      "calib/ece": 0.32298039215686275,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.002257727652464493,
      "calib/mean_conf": 0.009490196078431374,
      "calib/mu_c": 0.007976190476190477,
      "calib/mu_w": 0.01023391812865497,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0015294117647058824,
      "calib/std_conf": 0.01980628989915568,
      "calib/step_conf_rate": 1.0,
      "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": 1834.0,
      "completions/max_terminated_length": 1834.0,
      "completions/mean_length": 340.484375,
      "completions/mean_terminated_length": 340.484375,
      "completions/min_length": 116.0,
      "completions/min_terminated_length": 116.0,
      "epoch": 0.07893333333333333,
      "grad_norm": 0.005016027018427849,
      "learning_rate": 3.5e-06,
      "loss": 0.0484,
      "num_tokens": 14415179.0,
      "reward": 0.998470664024353,
      "reward_std": 0.015169315040111542,
      "rewards/accuracy_reward_step": 0.328125,
      "rewards/final_brier_reward_step": 0.6727226972579956,
      "rewards/format_reward_step": 0.99609375,
      "step": 74
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2712880775798112,
      "aux_distill/mean_u": 0.08906336368210896,
      "aux_distill/n_active_tok": 379.28125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5247783797094311,
      "calib/avg_num_step_conf": 5.90234375,
      "calib/ece": 0.4777254901960784,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00031273085446934315,
      "calib/mean_conf": 0.008549019607843138,
      "calib/mu_c": 0.00870967741935484,
      "calib/mu_w": 0.008396946564885497,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.01226554831670478,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1142.0,
      "completions/max_terminated_length": 1142.0,
      "completions/mean_length": 335.21484375,
      "completions/mean_terminated_length": 336.5294189453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 113.0,
      "epoch": 0.08,
      "grad_norm": 0.00509895384311676,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 0.0256,
      "num_tokens": 14605746.0,
      "reward": 1.0002011060714722,
      "reward_std": 0.017752250656485558,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.5199335813522339,
      "rewards/format_reward_step": 0.99609375,
      "step": 75
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.30028481944464147,
      "aux_distill/mean_u": 0.09071633159242892,
      "aux_distill/n_active_tok": 384.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4678846153846155,
      "calib/avg_num_step_conf": 6.0078125,
      "calib/ece": 0.4003149606299213,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0009756410256410268,
      "calib/mean_conf": 0.009133858267716536,
      "calib/mu_c": 0.008557692307692306,
      "calib/mu_w": 0.009533333333333333,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.011190802047579116,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1724.0,
      "completions/max_terminated_length": 1724.0,
      "completions/mean_length": 338.84375,
      "completions/mean_terminated_length": 340.1725769042969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 117.0,
      "epoch": 0.08106666666666666,
      "grad_norm": 0.004829046316444874,
      "learning_rate": 3.444444444444445e-06,
      "loss": 0.008,
      "num_tokens": 14795546.0,
      "reward": 0.9955216646194458,
      "reward_std": 0.026836896315217018,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.5926058292388916,
      "rewards/format_reward_step": 0.98828125,
      "step": 76
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.28181065688841045,
      "aux_distill/mean_u": 0.08920006079870602,
      "aux_distill/n_active_tok": 400.75,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5287990196078431,
      "calib/avg_num_step_conf": 6.26171875,
      "calib/ece": 0.3913779527559056,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.001964654282765736,
      "calib/mean_conf": 0.010196850393700788,
      "calib/mu_c": 0.011372549019607842,
      "calib/mu_w": 0.009407894736842106,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.016870296950967947,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2167.0,
      "completions/max_terminated_length": 2167.0,
      "completions/mean_length": 337.578125,
      "completions/mean_terminated_length": 337.578125,
      "completions/min_length": 98.0,
      "completions/min_terminated_length": 98.0,
      "epoch": 0.08213333333333334,
      "grad_norm": 0.004952338989824057,
      "learning_rate": 3.416666666666667e-06,
      "loss": 0.1114,
      "num_tokens": 14986630.0,
      "reward": 0.994382381439209,
      "reward_std": 0.0341326966881752,
      "rewards/accuracy_reward_step": 0.3984375,
      "rewards/final_brier_reward_step": 0.6020461320877075,
      "rewards/format_reward_step": 0.98828125,
      "step": 77
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.29722038214094937,
      "aux_distill/mean_u": 0.09135029054194987,
      "aux_distill/n_active_tok": 383.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5404382203162691,
      "calib/avg_num_step_conf": 6.0078125,
      "calib/ece": 0.344156862745098,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0016061377646743504,
      "calib/mean_conf": 0.016,
      "calib/mu_c": 0.017032967032967035,
      "calib/mu_w": 0.015426829268292684,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.001647058823529412,
      "calib/std_conf": 0.04928190217872014,
      "calib/step_conf_rate": 1.0,
      "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": 854.0,
      "completions/max_terminated_length": 854.0,
      "completions/mean_length": 359.78515625,
      "completions/mean_terminated_length": 361.19610595703125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.0832,
      "grad_norm": 0.004928613547235727,
      "learning_rate": 3.3888888888888893e-06,
      "loss": 0.0262,
      "num_tokens": 15186759.0,
      "reward": 1.0008113384246826,
      "reward_std": 0.021877873688936234,
      "rewards/accuracy_reward_step": 0.35546875,
      "rewards/final_brier_reward_step": 0.6500601768493652,
      "rewards/format_reward_step": 0.99609375,
      "step": 78
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.27827800950035453,
      "aux_distill/mean_u": 0.08564135031592453,
      "aux_distill/n_active_tok": 423.78125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5051779325927321,
      "calib/avg_num_step_conf": 6.62109375,
      "calib/ece": 0.43377952755905513,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.001154835875227517,
      "calib/mean_conf": 0.011968503937007874,
      "calib/mu_c": 0.011327433628318584,
      "calib/mu_w": 0.012482269503546101,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.00043307086614173226,
      "calib/std_conf": 0.012488695365610609,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2727.0,
      "completions/max_terminated_length": 2727.0,
      "completions/mean_length": 395.79296875,
      "completions/mean_terminated_length": 395.79296875,
      "completions/min_length": 81.0,
      "completions/min_terminated_length": 81.0,
      "epoch": 0.08426666666666667,
      "grad_norm": 0.004926387686282396,
      "learning_rate": 3.3611111111111117e-06,
      "loss": 0.088,
      "num_tokens": 15394458.0,
      "reward": 0.9950859546661377,
      "reward_std": 0.03354940935969353,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.5604844093322754,
      "rewards/format_reward_step": 0.98828125,
      "step": 79
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31684785219840705,
      "aux_distill/mean_u": 0.10744304575263432,
      "aux_distill/n_active_tok": 405.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5537071078431373,
      "calib/avg_num_step_conf": 6.375,
      "calib/ece": 0.45816406249999997,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0017205882352941178,
      "calib/mean_conf": 0.011835937500000001,
      "calib/mu_c": 0.012750000000000001,
      "calib/mu_w": 0.011029411764705883,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.000625,
      "calib/std_conf": 0.013669810111925248,
      "calib/step_conf_rate": 1.0,
      "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": 1007.0,
      "completions/max_terminated_length": 1007.0,
      "completions/mean_length": 334.9765625,
      "completions/mean_terminated_length": 336.29022216796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 138.0,
      "epoch": 0.08533333333333333,
      "grad_norm": 0.005264780949801207,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.0389,
      "num_tokens": 15582372.0,
      "reward": 1.0019071102142334,
      "reward_std": 0.018337277695536613,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.5389702916145325,
      "rewards/format_reward_step": 0.99609375,
      "step": 80
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3030662953387946,
      "aux_distill/mean_u": 0.09600267836191516,
      "aux_distill/n_active_tok": 392.0,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.571330777656079,
      "calib/avg_num_step_conf": 6.125,
      "calib/ece": 0.3348031496062992,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0016675794085432624,
      "calib/mean_conf": 0.013228346456692915,
      "calib/mu_c": 0.01431818181818182,
      "calib/mu_w": 0.012650602409638558,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0007874015748031496,
      "calib/std_conf": 0.01678403298994395,
      "calib/step_conf_rate": 1.0,
      "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": 2888.0,
      "completions/max_terminated_length": 2888.0,
      "completions/mean_length": 359.35546875,
      "completions/mean_terminated_length": 359.35546875,
      "completions/min_length": 113.0,
      "completions/min_terminated_length": 113.0,
      "epoch": 0.0864,
      "grad_norm": 0.004292392637580633,
      "learning_rate": 3.3055555555555558e-06,
      "loss": 0.0761,
      "num_tokens": 15780615.0,
      "reward": 0.9968827962875366,
      "reward_std": 0.028621265664696693,
      "rewards/accuracy_reward_step": 0.34375,
      "rewards/final_brier_reward_step": 0.6578280925750732,
      "rewards/format_reward_step": 0.9921875,
      "step": 81
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2719600473064929,
      "aux_distill/mean_u": 0.06987277391391887,
      "aux_distill/n_active_tok": 367.625,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.527185331380913,
      "calib/avg_num_step_conf": 5.7421875,
      "calib/ece": 0.4353174603174602,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0024269434010313895,
      "calib/mean_conf": 0.016825396825396827,
      "calib/mu_c": 0.015486725663716814,
      "calib/mu_w": 0.017913669064748203,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.001865079365079365,
      "calib/std_conf": 0.03316985611969488,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2329.0,
      "completions/max_terminated_length": 2329.0,
      "completions/mean_length": 350.4140625,
      "completions/mean_terminated_length": 350.4140625,
      "completions/min_length": 28.0,
      "completions/min_terminated_length": 28.0,
      "epoch": 0.08746666666666666,
      "grad_norm": 0.004946700297296047,
      "learning_rate": 3.277777777777778e-06,
      "loss": 0.0753,
      "num_tokens": 15975873.0,
      "reward": 0.9825640320777893,
      "reward_std": 0.04624699801206589,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.5510656237602234,
      "rewards/format_reward_step": 0.97265625,
      "step": 82
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.29490388510748744,
      "aux_distill/mean_u": 0.10091545887367892,
      "aux_distill/n_active_tok": 411.9375,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.542888678650268,
      "calib/avg_num_step_conf": 6.41796875,
      "calib/ece": 0.39640625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0015540838852097123,
      "calib/mean_conf": 0.013750000000000002,
      "calib/mu_c": 0.014666666666666666,
      "calib/mu_w": 0.013112582781456954,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.011353688827865594,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1210.0,
      "completions/max_terminated_length": 1210.0,
      "completions/mean_length": 383.4453125,
      "completions/mean_terminated_length": 384.94903564453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 119.0,
      "epoch": 0.08853333333333334,
      "grad_norm": 0.004362845327705145,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.0247,
      "num_tokens": 16181299.0,
      "reward": 1.0038261413574219,
      "reward_std": 0.013152113184332848,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.6014023423194885,
      "rewards/format_reward_step": 0.99609375,
      "step": 83
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.30570834246464074,
      "aux_distill/mean_u": 0.09895186440162848,
      "aux_distill/n_active_tok": 367.53125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5347222222222222,
      "calib/avg_num_step_conf": 5.7421875,
      "calib/ece": 0.31615079365079357,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.006130952380952372,
      "calib/mean_conf": 0.02257936507936508,
      "calib/mu_c": 0.026666666666666658,
      "calib/mu_w": 0.020535714285714286,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.002698412698412698,
      "calib/std_conf": 0.04605997697526515,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 973.0,
      "completions/max_terminated_length": 973.0,
      "completions/mean_length": 322.72265625,
      "completions/mean_terminated_length": 323.9882507324219,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 44.0,
      "epoch": 0.0896,
      "grad_norm": 0.005511968396604061,
      "learning_rate": 3.2222222222222227e-06,
      "loss": 0.0122,
      "num_tokens": 16369836.0,
      "reward": 0.9879236221313477,
      "reward_std": 0.0691649466753006,
      "rewards/accuracy_reward_step": 0.328125,
      "rewards/final_brier_reward_step": 0.6672534942626953,
      "rewards/format_reward_step": 0.98046875,
      "step": 84
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2954780561849475,
      "aux_distill/mean_u": 0.10002237214826427,
      "aux_distill/n_active_tok": 384.125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5420512820512821,
      "calib/avg_num_step_conf": 6.00390625,
      "calib/ece": 0.395092125984252,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0002288461538461524,
      "calib/mean_conf": 0.017506299212598424,
      "calib/mu_c": 0.017371153846153845,
      "calib/mu_w": 0.017599999999999998,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0015748031496062992,
      "calib/std_conf": 0.01828600966029217,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2014.0,
      "completions/max_terminated_length": 2014.0,
      "completions/mean_length": 366.33984375,
      "completions/mean_terminated_length": 366.33984375,
      "completions/min_length": 97.0,
      "completions/min_terminated_length": 97.0,
      "epoch": 0.09066666666666667,
      "grad_norm": 0.0047206999734044075,
      "learning_rate": 3.1944444444444443e-06,
      "loss": 0.0609,
      "num_tokens": 16571443.0,
      "reward": 0.99502032995224,
      "reward_std": 0.04041251540184021,
      "rewards/accuracy_reward_step": 0.40625,
      "rewards/final_brier_reward_step": 0.5955094695091248,
      "rewards/format_reward_step": 0.98828125,
      "step": 85
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2898980306927115,
      "aux_distill/mean_u": 0.09469991303484078,
      "aux_distill/n_active_tok": 381.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4839870820157438,
      "calib/avg_num_step_conf": 5.984375,
      "calib/ece": 0.3355859375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.005860862544573766,
      "calib/mean_conf": 0.0237109375,
      "calib/mu_c": 0.019887640449438203,
      "calib/mu_w": 0.02574850299401197,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0058203125,
      "calib/std_conf": 0.059770871399629874,
      "calib/step_conf_rate": 1.0,
      "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": 978.0,
      "completions/max_terminated_length": 978.0,
      "completions/mean_length": 349.9609375,
      "completions/mean_terminated_length": 351.3333435058594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 97.0,
      "epoch": 0.09173333333333333,
      "grad_norm": 0.005566046107560396,
      "learning_rate": 3.1666666666666667e-06,
      "loss": 0.0274,
      "num_tokens": 16766545.0,
      "reward": 0.9970375299453735,
      "reward_std": 0.03623506426811218,
      "rewards/accuracy_reward_step": 0.34765625,
      "rewards/final_brier_reward_step": 0.6542313098907471,
      "rewards/format_reward_step": 0.9921875,
      "step": 86
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.28562748548574746,
      "aux_distill/mean_u": 0.09414752872878823,
      "aux_distill/n_active_tok": 366.3125,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5269451920724249,
      "calib/avg_num_step_conf": 5.69921875,
      "calib/ece": 0.4659375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0078125,
      "calib/gap": -0.01422069977978957,
      "calib/mean_conf": 0.025312500000000005,
      "calib/mu_c": 0.017868852459016392,
      "calib/mu_w": 0.03208955223880596,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.00734375,
      "calib/std_conf": 0.08236198208973604,
      "calib/step_conf_rate": 1.0,
      "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": 935.0,
      "completions/max_terminated_length": 935.0,
      "completions/mean_length": 334.71484375,
      "completions/mean_terminated_length": 336.0274658203125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 71.0,
      "epoch": 0.0928,
      "grad_norm": 0.0058571891859173775,
      "learning_rate": 3.138888888888889e-06,
      "loss": -0.0098,
      "num_tokens": 16957728.0,
      "reward": 1.0048034191131592,
      "reward_std": 0.018815290182828903,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.5330445170402527,
      "rewards/format_reward_step": 1.0,
      "step": 87
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.28512857877649367,
      "aux_distill/mean_u": 0.090462066060402,
      "aux_distill/n_active_tok": 408.3125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5339267285861713,
      "calib/avg_num_step_conf": 6.3671875,
      "calib/ece": 0.3851968503937008,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0019569143446852445,
      "calib/mean_conf": 0.016377952755905513,
      "calib/mu_c": 0.01754901960784314,
      "calib/mu_w": 0.015592105263157895,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.012653423260850814,
      "calib/step_conf_rate": 1.0,
      "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": 1426.0,
      "completions/max_terminated_length": 1426.0,
      "completions/mean_length": 391.703125,
      "completions/mean_terminated_length": 393.2392272949219,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 127.0,
      "epoch": 0.09386666666666667,
      "grad_norm": 0.005111593287438154,
      "learning_rate": 3.1111111111111116e-06,
      "loss": 0.0367,
      "num_tokens": 17167852.0,
      "reward": 0.9911565780639648,
      "reward_std": 0.05165533721446991,
      "rewards/accuracy_reward_step": 0.3984375,
      "rewards/final_brier_reward_step": 0.5995007753372192,
      "rewards/format_reward_step": 0.984375,
      "step": 88
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3024716544896364,
      "aux_distill/mean_u": 0.09634288435249655,
      "aux_distill/n_active_tok": 387.75,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5104699703743604,
      "calib/avg_num_step_conf": 6.0703125,
      "calib/ece": 0.3505952380952381,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.002077834635065976,
      "calib/mean_conf": 0.02242063492063492,
      "calib/mu_c": 0.02372340425531914,
      "calib/mu_w": 0.021645569620253165,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.02292356283238295,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1989.0,
      "completions/max_terminated_length": 1989.0,
      "completions/mean_length": 395.765625,
      "completions/mean_terminated_length": 397.3176574707031,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 159.0,
      "epoch": 0.09493333333333333,
      "grad_norm": 0.004692970775067806,
      "learning_rate": 3.0833333333333336e-06,
      "loss": 0.0499,
      "num_tokens": 17378056.0,
      "reward": 0.9808880090713501,
      "reward_std": 0.07970473170280457,
      "rewards/accuracy_reward_step": 0.3671875,
      "rewards/final_brier_reward_step": 0.6219324469566345,
      "rewards/format_reward_step": 0.97265625,
      "step": 89
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31103678117506206,
      "aux_distill/mean_u": 0.12257401174085307,
      "aux_distill/n_active_tok": 406.15625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4879192546583851,
      "calib/avg_num_step_conf": 6.3671875,
      "calib/ece": 0.4323137254901961,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0016366459627329237,
      "calib/mean_conf": 0.018666666666666665,
      "calib/mu_c": 0.01956521739130435,
      "calib/mu_w": 0.017928571428571426,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.01844405070845197,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 984.0,
      "completions/max_terminated_length": 984.0,
      "completions/mean_length": 382.62890625,
      "completions/mean_terminated_length": 384.1294250488281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 136.0,
      "epoch": 0.096,
      "grad_norm": 0.004809149540960789,
      "learning_rate": 3.055555555555556e-06,
      "loss": 0.0212,
      "num_tokens": 17579329.0,
      "reward": 0.9986032843589783,
      "reward_std": 0.038281530141830444,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.5597066283226013,
      "rewards/format_reward_step": 0.98828125,
      "step": 90
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2773177553899586,
      "aux_distill/mean_u": 0.07881918878088494,
      "aux_distill/n_active_tok": 398.0,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5500914768784304,
      "calib/avg_num_step_conf": 6.21875,
      "calib/ece": 0.46099206349206345,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0019210144470380441,
      "calib/mean_conf": 0.01916666666666667,
      "calib/mu_c": 0.02016528925619835,
      "calib/mu_w": 0.018244274809160306,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.018827221932996,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2924.0,
      "completions/max_terminated_length": 2924.0,
      "completions/mean_length": 424.81640625,
      "completions/mean_terminated_length": 424.81640625,
      "completions/min_length": 132.0,
      "completions/min_terminated_length": 132.0,
      "epoch": 0.09706666666666666,
      "grad_norm": 0.004709687549620867,
      "learning_rate": 3.0277777777777776e-06,
      "loss": 0.0682,
      "num_tokens": 17795794.0,
      "reward": 0.9818326234817505,
      "reward_std": 0.07131138443946838,
      "rewards/accuracy_reward_step": 0.47265625,
      "rewards/final_brier_reward_step": 0.518352746963501,
      "rewards/format_reward_step": 0.97265625,
      "step": 91
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24812287255190313,
      "aux_distill/mean_u": 0.06410933965058271,
      "aux_distill/n_active_tok": 388.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5300342549547346,
      "calib/avg_num_step_conf": 6.140625,
      "calib/ece": 0.45496093750000005,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0021079031074137493,
      "calib/mean_conf": 0.021601562500000004,
      "calib/mu_c": 0.022704918032786885,
      "calib/mu_w": 0.020597014925373136,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.022329506769263707,
      "calib/step_conf_rate": 1.0,
      "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": 1385.0,
      "completions/max_terminated_length": 1385.0,
      "completions/mean_length": 381.68359375,
      "completions/mean_terminated_length": 383.180419921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 120.0,
      "epoch": 0.09813333333333334,
      "grad_norm": 0.005170757882297039,
      "learning_rate": 3e-06,
      "loss": 0.0344,
      "num_tokens": 18000225.0,
      "reward": 1.0103375911712646,
      "reward_std": 0.01272611878812313,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.5441129207611084,
      "rewards/format_reward_step": 1.0,
      "step": 92
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2716197760310024,
      "aux_distill/mean_u": 0.08429833158499851,
      "aux_distill/n_active_tok": 430.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5044732232232232,
      "calib/avg_num_step_conf": 6.72265625,
      "calib/ece": 0.41529411764705887,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0028697447447447386,
      "calib/mean_conf": 0.024235294117647063,
      "calib/mu_c": 0.02585585585585585,
      "calib/mu_w": 0.02298611111111111,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0021176470588235297,
      "calib/std_conf": 0.055906991472460144,
      "calib/step_conf_rate": 1.0,
      "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": 1958.0,
      "completions/max_terminated_length": 1958.0,
      "completions/mean_length": 417.24609375,
      "completions/mean_terminated_length": 417.24609375,
      "completions/min_length": 171.0,
      "completions/min_terminated_length": 171.0,
      "epoch": 0.0992,
      "grad_norm": 0.004604802001267672,
      "learning_rate": 2.9722222222222225e-06,
      "loss": 0.0576,
      "num_tokens": 18212816.0,
      "reward": 1.005455493927002,
      "reward_std": 0.02824069932103157,
      "rewards/accuracy_reward_step": 0.43359375,
      "rewards/final_brier_reward_step": 0.5812234282493591,
      "rewards/format_reward_step": 0.99609375,
      "step": 93
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2312735291197896,
      "aux_distill/mean_u": 0.0755610747205075,
      "aux_distill/n_active_tok": 407.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5466537966537968,
      "calib/avg_num_step_conf": 6.36328125,
      "calib/ece": 0.3968379446640316,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.003857142857142861,
      "calib/mean_conf": 0.023399209486166007,
      "calib/mu_c": 0.02114285714285714,
      "calib/mu_w": 0.025,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.002608695652173913,
      "calib/std_conf": 0.0438681405386164,
      "calib/step_conf_rate": 1.0,
      "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": 2575.0,
      "completions/max_terminated_length": 2575.0,
      "completions/mean_length": 424.89453125,
      "completions/mean_terminated_length": 426.5608215332031,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 121.0,
      "epoch": 0.10026666666666667,
      "grad_norm": 0.004604880232363939,
      "learning_rate": 2.944444444444445e-06,
      "loss": 0.0519,
      "num_tokens": 18430269.0,
      "reward": 0.9957316517829895,
      "reward_std": 0.044136419892311096,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.593025803565979,
      "rewards/format_reward_step": 0.98828125,
      "step": 94
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2925974018871784,
      "aux_distill/mean_u": 0.10130479070489674,
      "aux_distill/n_active_tok": 403.0,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5033488372093023,
      "calib/avg_num_step_conf": 6.296875,
      "calib/ece": 0.47114173228346456,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00084403100775194,
      "calib/mean_conf": 0.025708661417322837,
      "calib/mu_c": 0.02528,
      "calib/mu_w": 0.02612403100775194,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.002362204724409449,
      "calib/std_conf": 0.04316341594638858,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2479.0,
      "completions/max_terminated_length": 2479.0,
      "completions/mean_length": 414.8828125,
      "completions/mean_terminated_length": 414.8828125,
      "completions/min_length": 117.0,
      "completions/min_terminated_length": 117.0,
      "epoch": 0.10133333333333333,
      "grad_norm": 0.004823321010917425,
      "learning_rate": 2.916666666666667e-06,
      "loss": 0.0705,
      "num_tokens": 18642607.0,
      "reward": 0.9992181062698364,
      "reward_std": 0.04958602786064148,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/final_brier_reward_step": 0.5257800817489624,
      "rewards/format_reward_step": 0.984375,
      "step": 95
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.26682930160313845,
      "aux_distill/mean_u": 0.09679726153203617,
      "aux_distill/n_active_tok": 406.75,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5002578150177248,
      "calib/avg_num_step_conf": 6.35546875,
      "calib/ece": 0.5546825396825397,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00042926200451176913,
      "calib/mean_conf": 0.020714285714285716,
      "calib/mu_c": 0.020896551724137933,
      "calib/mu_w": 0.020467289719626164,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.01755619419602455,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2514.0,
      "completions/max_terminated_length": 2514.0,
      "completions/mean_length": 430.94921875,
      "completions/mean_terminated_length": 430.94921875,
      "completions/min_length": 141.0,
      "completions/min_terminated_length": 141.0,
      "epoch": 0.1024,
      "grad_norm": 0.004958832170814276,
      "learning_rate": 2.888888888888889e-06,
      "loss": 0.1083,
      "num_tokens": 18858746.0,
      "reward": 0.9919416904449463,
      "reward_std": 0.06605525314807892,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/final_brier_reward_step": 0.4370085895061493,
      "rewards/format_reward_step": 0.98046875,
      "step": 96
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2626429311931133,
      "aux_distill/mean_u": 0.07569661717053826,
      "aux_distill/n_active_tok": 415.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5216194482246691,
      "calib/avg_num_step_conf": 6.5390625,
      "calib/ece": 0.437921568627451,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.00392156862745098,
      "calib/gap": 0.009319559569466784,
      "calib/mean_conf": 0.024823529411764703,
      "calib/mu_c": 0.02983050847457627,
      "calib/mu_w": 0.020510948905109485,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.06323833098042919,
      "calib/step_conf_rate": 1.0,
      "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": 1214.0,
      "completions/max_terminated_length": 1214.0,
      "completions/mean_length": 411.3984375,
      "completions/mean_terminated_length": 413.01177978515625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 164.0,
      "epoch": 0.10346666666666667,
      "grad_norm": 0.004902747459709644,
      "learning_rate": 2.861111111111111e-06,
      "loss": 0.0518,
      "num_tokens": 19069136.0,
      "reward": 1.0075451135635376,
      "reward_std": 0.026664312928915024,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/final_brier_reward_step": 0.5580589771270752,
      "rewards/format_reward_step": 0.99609375,
      "step": 97
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3062682868912816,
      "aux_distill/mean_u": 0.12505230247901689,
      "aux_distill/n_active_tok": 402.5,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.49301054771889696,
      "calib/avg_num_step_conf": 6.2890625,
      "calib/ece": 0.4651394422310757,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0001880798068369556,
      "calib/mean_conf": 0.02091633466135458,
      "calib/mu_c": 0.02081967213114754,
      "calib/mu_w": 0.021007751937984497,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.01900820152488685,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2644.0,
      "completions/max_terminated_length": 2644.0,
      "completions/mean_length": 452.80078125,
      "completions/mean_terminated_length": 454.5765075683594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 163.0,
      "epoch": 0.10453333333333334,
      "grad_norm": 0.0052138082683086395,
      "learning_rate": 2.8333333333333335e-06,
      "loss": 0.0887,
      "num_tokens": 19291237.0,
      "reward": 0.9880070090293884,
      "reward_std": 0.0714532807469368,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.5228890180587769,
      "rewards/format_reward_step": 0.9765625,
      "step": 98
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24276379751972854,
      "aux_distill/mean_u": 0.0592441827702998,
      "aux_distill/n_active_tok": 411.90625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.49762761780104714,
      "calib/avg_num_step_conf": 6.4765625,
      "calib/ece": 0.2354509803921569,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.004008507853403141,
      "calib/mean_conf": 0.02862745098039216,
      "calib/mu_c": 0.025625000000000002,
      "calib/mu_w": 0.029633507853403143,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0065490196078431375,
      "calib/std_conf": 0.04644286423769036,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1407.0,
      "completions/max_terminated_length": 1407.0,
      "completions/mean_length": 477.36328125,
      "completions/mean_terminated_length": 479.2353210449219,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 167.0,
      "epoch": 0.1056,
      "grad_norm": 0.00454300781711936,
      "learning_rate": 2.805555555555556e-06,
      "loss": -0.0003,
      "num_tokens": 19519242.0,
      "reward": 0.9911375045776367,
      "reward_std": 0.040163010358810425,
      "rewards/accuracy_reward_step": 0.25,
      "rewards/final_brier_reward_step": 0.7479000091552734,
      "rewards/format_reward_step": 0.984375,
      "step": 99
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.28997284499928355,
      "aux_distill/mean_u": 0.10573104992759977,
      "aux_distill/n_active_tok": 403.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.44573888818538693,
      "calib/avg_num_step_conf": 6.31640625,
      "calib/ece": 0.39447058823529413,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.009021780422945425,
      "calib/mean_conf": 0.028196078431372548,
      "calib/mu_c": 0.02292452830188679,
      "calib/mu_w": 0.031946308724832215,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0034901960784313726,
      "calib/std_conf": 0.05780868365730124,
      "calib/step_conf_rate": 1.0,
      "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": 1251.0,
      "completions/max_terminated_length": 1251.0,
      "completions/mean_length": 438.28515625,
      "completions/mean_terminated_length": 440.0039367675781,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 193.0,
      "epoch": 0.10666666666666667,
      "grad_norm": 0.0051354048773646355,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 0.012,
      "num_tokens": 19738851.0,
      "reward": 1.0054786205291748,
      "reward_std": 0.020914018154144287,
      "rewards/accuracy_reward_step": 0.41796875,
      "rewards/final_brier_reward_step": 0.5968949198722839,
      "rewards/format_reward_step": 0.99609375,
      "step": 100
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2594314431771636,
      "aux_distill/mean_u": 0.0661123765385953,
      "aux_distill/n_active_tok": 425.03125,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.5487693389592124,
      "calib/avg_num_step_conf": 6.6171875,
      "calib/ece": 0.34370967741935476,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.004032258064516129,
      "calib/gap": 0.0009760900140646997,
      "calib/mean_conf": 0.03282258064516129,
      "calib/mu_c": 0.03344444444444445,
      "calib/mu_w": 0.03246835443037975,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0068145161290322585,
      "calib/std_conf": 0.08555379824029012,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2706.0,
      "completions/max_terminated_length": 2706.0,
      "completions/mean_length": 521.08203125,
      "completions/mean_terminated_length": 521.08203125,
      "completions/min_length": 159.0,
      "completions/min_terminated_length": 159.0,
      "epoch": 0.10773333333333333,
      "grad_norm": 0.004390342626720667,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.1169,
      "num_tokens": 19979240.0,
      "reward": 0.9725391864776611,
      "reward_std": 0.10312092304229736,
      "rewards/accuracy_reward_step": 0.3515625,
      "rewards/final_brier_reward_step": 0.6286722421646118,
      "rewards/format_reward_step": 0.96484375,
      "step": 101
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24633567919954658,
      "aux_distill/mean_u": 0.07292928626486234,
      "aux_distill/n_active_tok": 420.9375,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5528360528360529,
      "calib/avg_num_step_conf": 6.5234375,
      "calib/ece": 0.4944578313253012,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.012048192771084338,
      "calib/gap": 0.02542540792540793,
      "calib/mean_conf": 0.04040160642570282,
      "calib/mu_c": 0.05234848484848485,
      "calib/mu_w": 0.02692307692307692,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0023694779116465864,
      "calib/std_conf": 0.10869728495438853,
      "calib/step_conf_rate": 1.0,
      "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": 2759.0,
      "completions/max_terminated_length": 2759.0,
      "completions/mean_length": 456.921875,
      "completions/mean_terminated_length": 456.921875,
      "completions/min_length": 163.0,
      "completions/min_terminated_length": 163.0,
      "epoch": 0.1088,
      "grad_norm": 0.004803146701306105,
      "learning_rate": 2.7222222222222224e-06,
      "loss": 0.0851,
      "num_tokens": 20202908.0,
      "reward": 0.9931085705757141,
      "reward_std": 0.07233564555644989,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/final_brier_reward_step": 0.497935950756073,
      "rewards/format_reward_step": 0.97265625,
      "step": 102
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2805954380892217,
      "aux_distill/mean_u": 0.0894842856011007,
      "aux_distill/n_active_tok": 416.28125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.4522317188983856,
      "calib/avg_num_step_conf": 6.50390625,
      "calib/ece": 0.44277777777777777,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": -0.012991452991452997,
      "calib/mean_conf": 0.030634920634920636,
      "calib/mu_c": 0.02367521367521367,
      "calib/mu_w": 0.03666666666666667,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.004563492063492063,
      "calib/std_conf": 0.061608673267804566,
      "calib/step_conf_rate": 1.0,
      "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": 2636.0,
      "completions/max_terminated_length": 2636.0,
      "completions/mean_length": 532.38671875,
      "completions/mean_terminated_length": 532.38671875,
      "completions/min_length": 174.0,
      "completions/min_terminated_length": 174.0,
      "epoch": 0.10986666666666667,
      "grad_norm": 0.004372611176222563,
      "learning_rate": 2.6944444444444444e-06,
      "loss": 0.0946,
      "num_tokens": 20443751.0,
      "reward": 0.9928652048110962,
      "reward_std": 0.06021931394934654,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/final_brier_reward_step": 0.5443242192268372,
      "rewards/format_reward_step": 0.984375,
      "step": 103
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24913631053641438,
      "aux_distill/mean_u": 0.07466481954948698,
      "aux_distill/n_active_tok": 433.46875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.528455790784558,
      "calib/avg_num_step_conf": 6.83203125,
      "calib/ece": 0.401640234375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.002628555417185558,
      "calib/mean_conf": 0.028047265625000005,
      "calib/mu_c": 0.029546363636363638,
      "calib/mu_w": 0.02691780821917808,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.021453059048144936,
      "calib/step_conf_rate": 1.0,
      "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": 1032.0,
      "completions/max_terminated_length": 1032.0,
      "completions/mean_length": 441.9140625,
      "completions/mean_terminated_length": 443.6470947265625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 114.0,
      "epoch": 0.11093333333333333,
      "grad_norm": 0.005111818201839924,
      "learning_rate": 2.666666666666667e-06,
      "loss": 0.0324,
      "num_tokens": 20663561.0,
      "reward": 1.008166790008545,
      "reward_std": 0.025309395045042038,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.5905523300170898,
      "rewards/format_reward_step": 0.99609375,
      "step": 104
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25391951971687376,
      "aux_distill/mean_u": 0.07310204909175883,
      "aux_distill/n_active_tok": 430.5,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5330645161290323,
      "calib/avg_num_step_conf": 6.7265625,
      "calib/ece": 0.4659055118110236,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0021339950372208476,
      "calib/mean_conf": 0.02818897637795276,
      "calib/mu_c": 0.027096774193548386,
      "calib/mu_w": 0.029230769230769234,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.002952755905511811,
      "calib/std_conf": 0.025669532871671733,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2520.0,
      "completions/max_terminated_length": 2520.0,
      "completions/mean_length": 485.95703125,
      "completions/mean_terminated_length": 487.8627624511719,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 175.0,
      "epoch": 0.112,
      "grad_norm": 0.004727492108941078,
      "learning_rate": 2.6388888888888893e-06,
      "loss": 0.0597,
      "num_tokens": 20893726.0,
      "reward": 1.0006881952285767,
      "reward_std": 0.04665821045637131,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.5287203192710876,
      "rewards/format_reward_step": 0.98828125,
      "step": 105
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.28392570302821696,
      "aux_distill/mean_u": 0.09882959122606673,
      "aux_distill/n_active_tok": 381.78125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5003563330093943,
      "calib/avg_num_step_conf": 5.96875,
      "calib/ece": 0.38511904761904764,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00031292517006803466,
      "calib/mean_conf": 0.03694444444444445,
      "calib/mu_c": 0.03676190476190476,
      "calib/mu_w": 0.037074829931972794,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0026984126984126986,
      "calib/std_conf": 0.0537316680522342,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2148.0,
      "completions/max_terminated_length": 2148.0,
      "completions/mean_length": 448.796875,
      "completions/mean_terminated_length": 450.556884765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 133.0,
      "epoch": 0.11306666666666666,
      "grad_norm": 0.005242538638412952,
      "learning_rate": 2.6111111111111113e-06,
      "loss": 0.0689,
      "num_tokens": 21113202.0,
      "reward": 0.9895493984222412,
      "reward_std": 0.08381446450948715,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.5923800468444824,
      "rewards/format_reward_step": 0.9765625,
      "step": 106
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2789200625848025,
      "aux_distill/mean_u": 0.08994341773107507,
      "aux_distill/n_active_tok": 420.375,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.4842135847107438,
      "calib/avg_num_step_conf": 6.5625,
      "calib/ece": 0.45875502008032126,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.010996900826446276,
      "calib/mean_conf": 0.040281124497991976,
      "calib/mu_c": 0.03462809917355372,
      "calib/mu_w": 0.045625,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.006546184738955823,
      "calib/std_conf": 0.06681149894070164,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2999.0,
      "completions/max_terminated_length": 2999.0,
      "completions/mean_length": 493.15234375,
      "completions/mean_terminated_length": 495.0863037109375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 137.0,
      "epoch": 0.11413333333333334,
      "grad_norm": 0.00451470073312521,
      "learning_rate": 2.5833333333333337e-06,
      "loss": 0.0982,
      "num_tokens": 21344065.0,
      "reward": 0.9860634803771973,
      "reward_std": 0.08944529294967651,
      "rewards/accuracy_reward_step": 0.47265625,
      "rewards/final_brier_reward_step": 0.5268144607543945,
      "rewards/format_reward_step": 0.97265625,
      "step": 107
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2768033891916275,
      "aux_distill/mean_u": 0.1026058405370248,
      "aux_distill/n_active_tok": 409.3125,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5298049599586669,
      "calib/avg_num_step_conf": 6.390625,
      "calib/ece": 0.584296875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0078125,
      "calib/gap": 0.0021234823043141393,
      "calib/mean_conf": 0.04171875,
      "calib/mu_c": 0.04253164556962026,
      "calib/mu_w": 0.04040816326530612,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0044140625,
      "calib/std_conf": 0.09350475735724625,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1357.0,
      "completions/max_terminated_length": 1357.0,
      "completions/mean_length": 453.48046875,
      "completions/mean_terminated_length": 455.25885009765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 150.0,
      "epoch": 0.1152,
      "grad_norm": 0.004895161371678114,
      "learning_rate": 2.5555555555555557e-06,
      "loss": 0.013,
      "num_tokens": 21563388.0,
      "reward": 1.0190551280975342,
      "reward_std": 0.03934864327311516,
      "rewards/accuracy_reward_step": 0.6171875,
      "rewards/final_brier_reward_step": 0.4248288869857788,
      "rewards/format_reward_step": 0.99609375,
      "step": 108
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25648779910989106,
      "aux_distill/mean_u": 0.06899389784792351,
      "aux_distill/n_active_tok": 466.3125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5520887558563248,
      "calib/avg_num_step_conf": 7.1953125,
      "calib/ece": 0.4136546184738957,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.004016064257028112,
      "calib/gap": 0.009102030192608024,
      "calib/mean_conf": 0.049718875502008035,
      "calib/mu_c": 0.054690265486725675,
      "calib/mu_w": 0.04558823529411765,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.004779116465863454,
      "calib/std_conf": 0.09181963114604492,
      "calib/step_conf_rate": 1.0,
      "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": 2755.0,
      "completions/max_terminated_length": 2755.0,
      "completions/mean_length": 523.25390625,
      "completions/mean_terminated_length": 523.25390625,
      "completions/min_length": 151.0,
      "completions/min_terminated_length": 151.0,
      "epoch": 0.11626666666666667,
      "grad_norm": 0.004565491806715727,
      "learning_rate": 2.5277777777777778e-06,
      "loss": 0.1106,
      "num_tokens": 21801941.0,
      "reward": 0.9934476613998413,
      "reward_std": 0.08000709116458893,
      "rewards/accuracy_reward_step": 0.4453125,
      "rewards/final_brier_reward_step": 0.5689265727996826,
      "rewards/format_reward_step": 0.97265625,
      "step": 109
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2622479344718158,
      "aux_distill/mean_u": 0.07462715782623361,
      "aux_distill/n_active_tok": 385.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.47438095238095235,
      "calib/avg_num_step_conf": 6.01953125,
      "calib/ece": 0.38908549019607847,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.00392156862745098,
      "calib/gap": -0.021064761904761906,
      "calib/mean_conf": 0.049659607843137256,
      "calib/mu_c": 0.03726857142857143,
      "calib/mu_w": 0.058333333333333334,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.013490196078431372,
      "calib/std_conf": 0.09347799770399542,
      "calib/step_conf_rate": 1.0,
      "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": 2308.0,
      "completions/max_terminated_length": 2308.0,
      "completions/mean_length": 444.89453125,
      "completions/mean_terminated_length": 444.89453125,
      "completions/min_length": 161.0,
      "completions/min_terminated_length": 161.0,
      "epoch": 0.11733333333333333,
      "grad_norm": 0.0050817145965993404,
      "learning_rate": 2.5e-06,
      "loss": 0.061,
      "num_tokens": 22020754.0,
      "reward": 1.005799412727356,
      "reward_std": 0.04130372405052185,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.6053489446640015,
      "rewards/format_reward_step": 0.99609375,
      "step": 110
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.27884344570338726,
      "aux_distill/mean_u": 0.09474326124953433,
      "aux_distill/n_active_tok": 379.75,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5250907270679516,
      "calib/avg_num_step_conf": 5.93359375,
      "calib/ece": 0.43691699604743095,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.012169315479914913,
      "calib/mean_conf": 0.05083003952569171,
      "calib/mu_c": 0.05713114754098362,
      "calib/mu_w": 0.044961832061068706,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0027667984189723325,
      "calib/std_conf": 0.10036545261126484,
      "calib/step_conf_rate": 1.0,
      "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": 2514.0,
      "completions/max_terminated_length": 2514.0,
      "completions/mean_length": 461.01953125,
      "completions/mean_terminated_length": 462.8274841308594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 164.0,
      "epoch": 0.1184,
      "grad_norm": 0.00502881919965148,
      "learning_rate": 2.4722222222222226e-06,
      "loss": 0.0708,
      "num_tokens": 22246183.0,
      "reward": 1.0092535018920898,
      "reward_std": 0.06828673183917999,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.5536632537841797,
      "rewards/format_reward_step": 0.98828125,
      "step": 111
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2952077784575522,
      "aux_distill/mean_u": 0.12138369561467421,
      "aux_distill/n_active_tok": 394.25,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5290826284970722,
      "calib/avg_num_step_conf": 6.22265625,
      "calib/ece": 0.37752988047808766,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.00398406374501992,
      "calib/gap": 0.02160897852960314,
      "calib/mean_conf": 0.0447808764940239,
      "calib/mu_c": 0.05726415094339624,
      "calib/mu_w": 0.0356551724137931,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.08238629610369641,
      "calib/step_conf_rate": 1.0,
      "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": 3039.0,
      "completions/max_terminated_length": 3039.0,
      "completions/mean_length": 531.4375,
      "completions/mean_terminated_length": 533.5216064453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 147.0,
      "epoch": 0.11946666666666667,
      "grad_norm": 0.004346077796071768,
      "learning_rate": 2.4444444444444447e-06,
      "loss": 0.0749,
      "num_tokens": 22490151.0,
      "reward": 0.999869167804718,
      "reward_std": 0.07355723530054092,
      "rewards/accuracy_reward_step": 0.4140625,
      "rewards/final_brier_reward_step": 0.6052069664001465,
      "rewards/format_reward_step": 0.98046875,
      "step": 112
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2909727890510112,
      "aux_distill/mean_u": 0.09329004138584782,
      "aux_distill/n_active_tok": 421.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5154938271604937,
      "calib/avg_num_step_conf": 6.59375,
      "calib/ece": 0.42960784313725486,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.00392156862745098,
      "calib/gap": 0.00021296296296297035,
      "calib/mean_conf": 0.05180392156862745,
      "calib/mu_c": 0.05191666666666667,
      "calib/mu_w": 0.0517037037037037,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.005411764705882353,
      "calib/std_conf": 0.08042948917977372,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2453.0,
      "completions/max_terminated_length": 2453.0,
      "completions/mean_length": 429.10546875,
      "completions/mean_terminated_length": 429.10546875,
      "completions/min_length": 207.0,
      "completions/min_terminated_length": 207.0,
      "epoch": 0.12053333333333334,
      "grad_norm": 0.005415992345660925,
      "learning_rate": 2.4166666666666667e-06,
      "loss": 0.0887,
      "num_tokens": 22705202.0,
      "reward": 1.0138792991638184,
      "reward_std": 0.0524032860994339,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.5668210983276367,
      "rewards/format_reward_step": 0.9921875,
      "step": 113
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2658761255443096,
      "aux_distill/mean_u": 0.08234790336442177,
      "aux_distill/n_active_tok": 377.46875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4665414922432899,
      "calib/avg_num_step_conf": 5.88671875,
      "calib/ece": 0.46392156862745093,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.00784313725490196,
      "calib/gap": -0.0002327013050972826,
      "calib/mean_conf": 0.06423529411764707,
      "calib/mu_c": 0.06412213740458016,
      "calib/mu_w": 0.06435483870967744,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.007215686274509805,
      "calib/std_conf": 0.12498197332530019,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1709.0,
      "completions/max_terminated_length": 1709.0,
      "completions/mean_length": 406.2265625,
      "completions/mean_terminated_length": 406.2265625,
      "completions/min_length": 128.0,
      "completions/min_terminated_length": 128.0,
      "epoch": 0.1216,
      "grad_norm": 0.005368839483708143,
      "learning_rate": 2.388888888888889e-06,
      "loss": 0.0373,
      "num_tokens": 22914220.0,
      "reward": 1.0151684284210205,
      "reward_std": 0.0704091489315033,
      "rewards/accuracy_reward_step": 0.51171875,
      "rewards/final_brier_reward_step": 0.5264304876327515,
      "rewards/format_reward_step": 0.9921875,
      "step": 114
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24501192453317344,
      "aux_distill/mean_u": 0.07385531746068447,
      "aux_distill/n_active_tok": 413.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.519644762442729,
      "calib/avg_num_step_conf": 6.45703125,
      "calib/ece": 0.4083858267716536,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": -0.0037977781961965745,
      "calib/mean_conf": 0.04405511811023623,
      "calib/mu_c": 0.04194690265486725,
      "calib/mu_w": 0.045744680851063826,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0037795275590551177,
      "calib/std_conf": 0.06914333020328298,
      "calib/step_conf_rate": 1.0,
      "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": 990.0,
      "completions/max_terminated_length": 990.0,
      "completions/mean_length": 409.91015625,
      "completions/mean_terminated_length": 411.5176696777344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 153.0,
      "epoch": 0.12266666666666666,
      "grad_norm": 0.005021527409553528,
      "learning_rate": 2.361111111111111e-06,
      "loss": 0.017,
      "num_tokens": 23124421.0,
      "reward": 1.007368564605713,
      "reward_std": 0.04787181690335274,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.5811433792114258,
      "rewards/format_reward_step": 0.9921875,
      "step": 115
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25968706305138767,
      "aux_distill/mean_u": 0.07294114098545383,
      "aux_distill/n_active_tok": 454.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.48019192913385833,
      "calib/avg_num_step_conf": 7.1015625,
      "calib/ece": 0.4547450980392157,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.00784313725490196,
      "calib/gap": -0.004753321850393698,
      "calib/mean_conf": 0.05396078431372548,
      "calib/mu_c": 0.05157480314960631,
      "calib/mu_w": 0.05632812500000001,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.005333333333333333,
      "calib/std_conf": 0.09392319782450408,
      "calib/step_conf_rate": 1.0,
      "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": 2546.0,
      "completions/max_terminated_length": 2546.0,
      "completions/mean_length": 480.7265625,
      "completions/mean_terminated_length": 480.7265625,
      "completions/min_length": 137.0,
      "completions/min_terminated_length": 137.0,
      "epoch": 0.12373333333333333,
      "grad_norm": 0.004678206518292427,
      "learning_rate": 2.3333333333333336e-06,
      "loss": 0.0504,
      "num_tokens": 23352007.0,
      "reward": 1.0158360004425049,
      "reward_std": 0.043127551674842834,
      "rewards/accuracy_reward_step": 0.49609375,
      "rewards/final_brier_reward_step": 0.5394843816757202,
      "rewards/format_reward_step": 0.99609375,
      "step": 116
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2503124848008156,
      "aux_distill/mean_u": 0.06315268753420264,
      "aux_distill/n_active_tok": 422.5625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5624794799395889,
      "calib/avg_num_step_conf": 6.6015625,
      "calib/ece": 0.3496062992125984,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.007874015748031496,
      "calib/gap": -0.0007735241972552376,
      "calib/mean_conf": 0.07078740157480315,
      "calib/mu_c": 0.07030927835051547,
      "calib/mu_w": 0.0710828025477707,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.019251968503937008,
      "calib/std_conf": 0.13701316851024462,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 429.171875,
      "completions/mean_terminated_length": 430.85491943359375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 182.0,
      "epoch": 0.1248,
      "grad_norm": 0.004880519118160009,
      "learning_rate": 2.305555555555556e-06,
      "loss": 0.0006,
      "num_tokens": 23568475.0,
      "reward": 1.0031300783157349,
      "reward_std": 0.07905666530132294,
      "rewards/accuracy_reward_step": 0.37890625,
      "rewards/final_brier_reward_step": 0.6390726566314697,
      "rewards/format_reward_step": 0.98828125,
      "step": 117
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2309753114823252,
      "aux_distill/mean_u": 0.0686600635114206,
      "aux_distill/n_active_tok": 419.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.4429446319210099,
      "calib/avg_num_step_conf": 6.55078125,
      "calib/ece": 0.4574703557312253,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.003952569169960474,
      "calib/gap": -0.013482689663792007,
      "calib/mean_conf": 0.05446640316205534,
      "calib/mu_c": 0.0476984126984127,
      "calib/mu_w": 0.061181102362204705,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.006956521739130434,
      "calib/std_conf": 0.09215584335319088,
      "calib/step_conf_rate": 1.0,
      "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": 2754.0,
      "completions/max_terminated_length": 2754.0,
      "completions/mean_length": 433.08203125,
      "completions/mean_terminated_length": 434.7804260253906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 172.0,
      "epoch": 0.12586666666666665,
      "grad_norm": 0.004958142526447773,
      "learning_rate": 2.277777777777778e-06,
      "loss": 0.0466,
      "num_tokens": 23783352.0,
      "reward": 1.006095290184021,
      "reward_std": 0.06497341394424438,
      "rewards/accuracy_reward_step": 0.4921875,
      "rewards/final_brier_reward_step": 0.531721830368042,
      "rewards/format_reward_step": 0.98828125,
      "step": 118
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.29260341147892177,
      "aux_distill/mean_u": 0.08997504566866737,
      "aux_distill/n_active_tok": 431.09375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5216897856242118,
      "calib/avg_num_step_conf": 6.73828125,
      "calib/ece": 0.4380952380952381,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": 0.012141235813366957,
      "calib/mean_conf": 0.046031746031746035,
      "calib/mu_c": 0.052295081967213115,
      "calib/mu_w": 0.04015384615384616,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0731174242048241,
      "calib/step_conf_rate": 1.0,
      "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": 2578.0,
      "completions/max_terminated_length": 2578.0,
      "completions/mean_length": 465.40625,
      "completions/mean_terminated_length": 467.2314147949219,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 103.0,
      "epoch": 0.12693333333333334,
      "grad_norm": 0.004794496577233076,
      "learning_rate": 2.25e-06,
      "loss": 0.0309,
      "num_tokens": 24007560.0,
      "reward": 1.0056226253509521,
      "reward_std": 0.05668995529413223,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.5503078103065491,
      "rewards/format_reward_step": 0.984375,
      "step": 119
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.28152807522565126,
      "aux_distill/mean_u": 0.08463958399267932,
      "aux_distill/n_active_tok": 378.46875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5635561860120882,
      "calib/avg_num_step_conf": 5.90234375,
      "calib/ece": 0.482078431372549,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.00784313725490196,
      "calib/gap": 0.016905143702972733,
      "calib/mean_conf": 0.0507843137254902,
      "calib/mu_c": 0.05880597014925373,
      "calib/mu_w": 0.041900826446281,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.003686274509803921,
      "calib/std_conf": 0.09609723705511551,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2741.0,
      "completions/max_terminated_length": 2741.0,
      "completions/mean_length": 394.36328125,
      "completions/mean_terminated_length": 394.36328125,
      "completions/min_length": 139.0,
      "completions/min_terminated_length": 139.0,
      "epoch": 0.128,
      "grad_norm": 0.005581455770879984,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.0293,
      "num_tokens": 24215205.0,
      "reward": 1.01896071434021,
      "reward_std": 0.050968606024980545,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/final_brier_reward_step": 0.5222964286804199,
      "rewards/format_reward_step": 0.9921875,
      "step": 120
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2638569069094956,
      "aux_distill/mean_u": 0.08030497207501669,
      "aux_distill/n_active_tok": 430.40625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5537249733258018,
      "calib/avg_num_step_conf": 6.72265625,
      "calib/ece": 0.39425196850393696,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": 0.028054980229711923,
      "calib/mean_conf": 0.061417322834645675,
      "calib/mu_c": 0.07699115044247788,
      "calib/mu_w": 0.04893617021276596,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.005393700787401575,
      "calib/std_conf": 0.12124176672719679,
      "calib/step_conf_rate": 1.0,
      "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": 2032.0,
      "completions/max_terminated_length": 2032.0,
      "completions/mean_length": 477.2421875,
      "completions/mean_terminated_length": 477.2421875,
      "completions/min_length": 170.0,
      "completions/min_terminated_length": 170.0,
      "epoch": 0.12906666666666666,
      "grad_norm": 0.005236073397099972,
      "learning_rate": 2.1944444444444445e-06,
      "loss": 0.0612,
      "num_tokens": 24442435.0,
      "reward": 1.013102650642395,
      "reward_std": 0.06841275840997696,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.5965180397033691,
      "rewards/format_reward_step": 0.98828125,
      "step": 121
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.30170248029753566,
      "aux_distill/mean_u": 0.09285741823046123,
      "aux_distill/n_active_tok": 392.25,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5370165065287016,
      "calib/avg_num_step_conf": 6.12890625,
      "calib/ece": 0.4724313725490196,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.00392156862745098,
      "calib/gap": -0.0039042867701404374,
      "calib/mean_conf": 0.05635294117647059,
      "calib/mu_c": 0.05446969696969697,
      "calib/mu_w": 0.058373983739837404,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.005568627450980392,
      "calib/std_conf": 0.08888653783895817,
      "calib/step_conf_rate": 1.0,
      "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": 2746.0,
      "completions/max_terminated_length": 2746.0,
      "completions/mean_length": 416.80078125,
      "completions/mean_terminated_length": 416.80078125,
      "completions/min_length": 173.0,
      "completions/min_terminated_length": 173.0,
      "epoch": 0.13013333333333332,
      "grad_norm": 0.006182767450809479,
      "learning_rate": 2.166666666666667e-06,
      "loss": 0.0311,
      "num_tokens": 24656480.0,
      "reward": 1.014758586883545,
      "reward_std": 0.05790330469608307,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/final_brier_reward_step": 0.5217046737670898,
      "rewards/format_reward_step": 0.9921875,
      "step": 122
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.29011322488076985,
      "aux_distill/mean_u": 0.09992779375887095,
      "aux_distill/n_active_tok": 414.84375,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4108954845256215,
      "calib/avg_num_step_conf": 6.4609375,
      "calib/ece": 0.39342519685039384,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.007874015748031496,
      "calib/gap": -0.022611618467782847,
      "calib/mean_conf": 0.059015748031496056,
      "calib/mu_c": 0.04601851851851852,
      "calib/mu_w": 0.06863013698630137,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.013622047244094488,
      "calib/std_conf": 0.11863485195530105,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2660.0,
      "completions/max_terminated_length": 2660.0,
      "completions/mean_length": 486.0546875,
      "completions/mean_terminated_length": 486.0546875,
      "completions/min_length": 187.0,
      "completions/min_terminated_length": 187.0,
      "epoch": 0.1312,
      "grad_norm": 0.005104098469018936,
      "learning_rate": 2.138888888888889e-06,
      "loss": 0.0494,
      "num_tokens": 24886198.0,
      "reward": 0.9989855289459229,
      "reward_std": 0.05920081585645676,
      "rewards/accuracy_reward_step": 0.421875,
      "rewards/final_brier_reward_step": 0.5878148078918457,
      "rewards/format_reward_step": 0.98828125,
      "step": 123
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.27182958112098277,
      "aux_distill/mean_u": 0.07772999730943052,
      "aux_distill/n_active_tok": 394.6875,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5088645910551025,
      "calib/avg_num_step_conf": 6.1796875,
      "calib/ece": 0.5028515625000001,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.00390625,
      "calib/gap": -0.0032406721027922053,
      "calib/mean_conf": 0.06785156250000002,
      "calib/mu_c": 0.06640845070422535,
      "calib/mu_w": 0.06964912280701756,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0080078125,
      "calib/std_conf": 0.11065003882199315,
      "calib/step_conf_rate": 1.0,
      "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": 1031.0,
      "completions/max_terminated_length": 1031.0,
      "completions/mean_length": 432.05078125,
      "completions/mean_terminated_length": 433.7451171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 176.0,
      "epoch": 0.13226666666666667,
      "grad_norm": 0.005296853370964527,
      "learning_rate": 2.1111111111111114e-06,
      "loss": 0.0149,
      "num_tokens": 25103619.0,
      "reward": 1.0284123420715332,
      "reward_std": 0.04734199121594429,
      "rewards/accuracy_reward_step": 0.5546875,
      "rewards/final_brier_reward_step": 0.5021370649337769,
      "rewards/format_reward_step": 1.0,
      "step": 124
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2565637626685202,
      "aux_distill/mean_u": 0.08125035877125407,
      "aux_distill/n_active_tok": 428.6875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.519600428562425,
      "calib/avg_num_step_conf": 6.6796875,
      "calib/ece": 0.4577380952380951,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.009307997731140091,
      "calib/mean_conf": 0.07035714285714285,
      "calib/mu_c": 0.0658139534883721,
      "calib/mu_w": 0.07512195121951219,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.008095238095238095,
      "calib/std_conf": 0.10090203989676719,
      "calib/step_conf_rate": 1.0,
      "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": 2476.0,
      "completions/max_terminated_length": 2476.0,
      "completions/mean_length": 464.0078125,
      "completions/mean_terminated_length": 464.0078125,
      "completions/min_length": 151.0,
      "completions/min_terminated_length": 151.0,
      "epoch": 0.13333333333333333,
      "grad_norm": 0.004538561217486858,
      "learning_rate": 2.0833333333333334e-06,
      "loss": 0.1154,
      "num_tokens": 25327213.0,
      "reward": 1.0061924457550049,
      "reward_std": 0.09807506203651428,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.5280097723007202,
      "rewards/format_reward_step": 0.98046875,
      "step": 125
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25415724446065724,
      "aux_distill/mean_u": 0.09618354704585452,
      "aux_distill/n_active_tok": 453.0,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5506628787878788,
      "calib/avg_num_step_conf": 7.078125,
      "calib/ece": 0.3693700787401575,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.015748031496062992,
      "calib/gap": 0.03522474747474749,
      "calib/mean_conf": 0.07039370078740156,
      "calib/mu_c": 0.09036363636363638,
      "calib/mu_w": 0.05513888888888889,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.003346456692913386,
      "calib/std_conf": 0.13549211311447404,
      "calib/step_conf_rate": 1.0,
      "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": 2794.0,
      "completions/max_terminated_length": 2794.0,
      "completions/mean_length": 458.890625,
      "completions/mean_terminated_length": 458.890625,
      "completions/min_length": 156.0,
      "completions/min_terminated_length": 156.0,
      "epoch": 0.1344,
      "grad_norm": 0.00495866360142827,
      "learning_rate": 2.0555555555555555e-06,
      "loss": 0.067,
      "num_tokens": 25550153.0,
      "reward": 1.0194499492645264,
      "reward_std": 0.06901095807552338,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.6170250177383423,
      "rewards/format_reward_step": 0.9921875,
      "step": 126
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24842926231212914,
      "aux_distill/mean_u": 0.08070918253257382,
      "aux_distill/n_active_tok": 426.5,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.4438401214267645,
      "calib/avg_num_step_conf": 6.6640625,
      "calib/ece": 0.412579365079365,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.008068555527447507,
      "calib/mean_conf": 0.06734126984126985,
      "calib/mu_c": 0.06305084745762712,
      "calib/mu_w": 0.07111940298507463,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.005833333333333333,
      "calib/std_conf": 0.09446420861627229,
      "calib/step_conf_rate": 1.0,
      "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": 2899.0,
      "completions/max_terminated_length": 2899.0,
      "completions/mean_length": 445.41796875,
      "completions/mean_terminated_length": 448.9252014160156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.13546666666666668,
      "grad_norm": 0.0055075036361813545,
      "learning_rate": 2.027777777777778e-06,
      "loss": 0.0387,
      "num_tokens": 25767852.0,
      "reward": 1.0068135261535645,
      "reward_std": 0.07916107773780823,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/final_brier_reward_step": 0.5683144330978394,
      "rewards/format_reward_step": 0.984375,
      "step": 127
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.28774312208406627,
      "aux_distill/mean_u": 0.08850660982137487,
      "aux_distill/n_active_tok": 385.75,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5455390910781821,
      "calib/avg_num_step_conf": 6.02734375,
      "calib/ece": 0.432244094488189,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": 0.007952755905511838,
      "calib/mean_conf": 0.08216535433070866,
      "calib/mu_c": 0.08614173228346458,
      "calib/mu_w": 0.07818897637795275,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.007204724409448818,
      "calib/std_conf": 0.13285343672660974,
      "calib/step_conf_rate": 1.0,
      "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": 2503.0,
      "completions/max_terminated_length": 2503.0,
      "completions/mean_length": 465.015625,
      "completions/mean_terminated_length": 466.8392333984375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 183.0,
      "epoch": 0.13653333333333334,
      "grad_norm": 0.005524276290088892,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0462,
      "num_tokens": 25993560.0,
      "reward": 1.0189135074615479,
      "reward_std": 0.0828317403793335,
      "rewards/accuracy_reward_step": 0.49609375,
      "rewards/final_brier_reward_step": 0.5534519553184509,
      "rewards/format_reward_step": 0.98828125,
      "step": 128
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.286829563556239,
      "aux_distill/mean_u": 0.10107326307054654,
      "aux_distill/n_active_tok": 427.4375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5765517241379311,
      "calib/avg_num_step_conf": 6.625,
      "calib/ece": 0.47545098039215694,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.00784313725490196,
      "calib/gap": 0.06378056426332289,
      "calib/mean_conf": 0.0931764705882353,
      "calib/mu_c": 0.1206896551724138,
      "calib/mu_w": 0.05690909090909091,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.15115148950539692,
      "calib/step_conf_rate": 1.0,
      "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": 1065.0,
      "completions/max_terminated_length": 1065.0,
      "completions/mean_length": 413.0859375,
      "completions/mean_terminated_length": 414.7059020996094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 163.0,
      "epoch": 0.1376,
      "grad_norm": 0.005374370142817497,
      "learning_rate": 1.9722222222222224e-06,
      "loss": 0.0417,
      "num_tokens": 26201694.0,
      "reward": 1.048750400543213,
      "reward_std": 0.07723085582256317,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/final_brier_reward_step": 0.5350008010864258,
      "rewards/format_reward_step": 0.99609375,
      "step": 129
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25664816494099796,
      "aux_distill/mean_u": 0.08382097004231381,
      "aux_distill/n_active_tok": 418.5625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5630448717948718,
      "calib/avg_num_step_conf": 6.55859375,
      "calib/ece": 0.5329296875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.01171875,
      "calib/gap": 0.011325641025641028,
      "calib/mean_conf": 0.0841015625,
      "calib/mu_c": 0.08852564102564103,
      "calib/mu_w": 0.0772,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0038281250000000017,
      "calib/std_conf": 0.13308821908816196,
      "calib/step_conf_rate": 1.0,
      "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": 1410.0,
      "completions/max_terminated_length": 1410.0,
      "completions/mean_length": 425.19921875,
      "completions/mean_terminated_length": 426.86669921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 171.0,
      "epoch": 0.13866666666666666,
      "grad_norm": 0.00602386612445116,
      "learning_rate": 1.944444444444445e-06,
      "loss": 0.0334,
      "num_tokens": 26415833.0,
      "reward": 1.0415525436401367,
      "reward_std": 0.060056086629629135,
      "rewards/accuracy_reward_step": 0.609375,
      "rewards/final_brier_reward_step": 0.47373008728027344,
      "rewards/format_reward_step": 1.0,
      "step": 130
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2656695644836873,
      "aux_distill/mean_u": 0.07452905320152445,
      "aux_distill/n_active_tok": 410.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.47207547169811315,
      "calib/avg_num_step_conf": 6.515625,
      "calib/ece": 0.3513671875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.01171875,
      "calib/gap": -0.010069182389937092,
      "calib/mean_conf": 0.09816406250000001,
      "calib/mu_c": 0.09226415094339624,
      "calib/mu_w": 0.10233333333333333,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.017734375,
      "calib/std_conf": 0.15105148529059917,
      "calib/step_conf_rate": 1.0,
      "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": 1161.0,
      "completions/max_terminated_length": 1161.0,
      "completions/mean_length": 406.83984375,
      "completions/mean_terminated_length": 408.4353332519531,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 204.0,
      "epoch": 0.13973333333333332,
      "grad_norm": 0.006129199638962746,
      "learning_rate": 1.916666666666667e-06,
      "loss": 0.0029,
      "num_tokens": 26626192.0,
      "reward": 1.0219767093658447,
      "reward_std": 0.0614263117313385,
      "rewards/accuracy_reward_step": 0.4140625,
      "rewards/final_brier_reward_step": 0.629891037940979,
      "rewards/format_reward_step": 1.0,
      "step": 131
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.29841618961654603,
      "aux_distill/mean_u": 0.14273641680605606,
      "aux_distill/n_active_tok": 422.15625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4860747365059087,
      "calib/avg_num_step_conf": 6.625,
      "calib/ece": 0.5340625000000001,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0078125,
      "calib/gap": 0.0047716384541680035,
      "calib/mean_conf": 0.075859375,
      "calib/mu_c": 0.07774193548387097,
      "calib/mu_w": 0.07297029702970297,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0022265625,
      "calib/std_conf": 0.10301340434433459,
      "calib/step_conf_rate": 1.0,
      "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": 1196.0,
      "completions/max_terminated_length": 1196.0,
      "completions/mean_length": 446.05078125,
      "completions/mean_terminated_length": 447.8000183105469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 112.0,
      "epoch": 0.1408,
      "grad_norm": 0.005947123281657696,
      "learning_rate": 1.888888888888889e-06,
      "loss": 0.056,
      "num_tokens": 26845973.0,
      "reward": 1.0388870239257812,
      "reward_std": 0.04800880327820778,
      "rewards/accuracy_reward_step": 0.60546875,
      "rewards/final_brier_reward_step": 0.4723054766654968,
      "rewards/format_reward_step": 1.0,
      "step": 132
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2982191611081362,
      "aux_distill/mean_u": 0.09348481597065017,
      "aux_distill/n_active_tok": 451.40625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5446681792399319,
      "calib/avg_num_step_conf": 7.0390625,
      "calib/ece": 0.2591338582677165,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.007874015748031496,
      "calib/gap": 0.02002977878615997,
      "calib/mean_conf": 0.09582677165354332,
      "calib/mu_c": 0.10939024390243904,
      "calib/mu_w": 0.08936046511627907,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.016062992125984252,
      "calib/std_conf": 0.14255233018907015,
      "calib/step_conf_rate": 1.0,
      "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": 2528.0,
      "completions/max_terminated_length": 2528.0,
      "completions/mean_length": 500.64453125,
      "completions/mean_terminated_length": 502.6078796386719,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 148.0,
      "epoch": 0.14186666666666667,
      "grad_norm": 0.004796723835170269,
      "learning_rate": 1.8611111111111113e-06,
      "loss": 0.0458,
      "num_tokens": 27080482.0,
      "reward": 1.0125898122787476,
      "reward_std": 0.08151218295097351,
      "rewards/accuracy_reward_step": 0.3203125,
      "rewards/final_brier_reward_step": 0.7126797437667847,
      "rewards/format_reward_step": 0.9921875,
      "step": 133
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25477377604693174,
      "aux_distill/mean_u": 0.07858637495460441,
      "aux_distill/n_active_tok": 447.0,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5150949268596328,
      "calib/avg_num_step_conf": 6.96875,
      "calib/ece": 0.3933464566929134,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.015748031496062992,
      "calib/gap": 0.01750575785869904,
      "calib/mean_conf": 0.09027559055118112,
      "calib/mu_c": 0.09957983193277312,
      "calib/mu_w": 0.08207407407407408,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.007559055118110235,
      "calib/std_conf": 0.149685769402636,
      "calib/step_conf_rate": 1.0,
      "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": 1522.0,
      "completions/max_terminated_length": 1522.0,
      "completions/mean_length": 510.84375,
      "completions/mean_terminated_length": 512.8471069335938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 199.0,
      "epoch": 0.14293333333333333,
      "grad_norm": 0.004951619543135166,
      "learning_rate": 1.8333333333333333e-06,
      "loss": 0.0196,
      "num_tokens": 27320210.0,
      "reward": 1.0233181715011597,
      "reward_std": 0.0827612429857254,
      "rewards/accuracy_reward_step": 0.46484375,
      "rewards/final_brier_reward_step": 0.5896050930023193,
      "rewards/format_reward_step": 0.9921875,
      "step": 134
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.26247306214645505,
      "aux_distill/mean_u": 0.07126044264829093,
      "aux_distill/n_active_tok": 443.28125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5324651162790698,
      "calib/avg_num_step_conf": 6.8984375,
      "calib/ece": 0.4218503937007875,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.01968503937007874,
      "calib/gap": 0.01186914728682173,
      "calib/mean_conf": 0.10925196850393701,
      "calib/mu_c": 0.11528000000000002,
      "calib/mu_w": 0.10341085271317829,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.019488188976377954,
      "calib/std_conf": 0.16813424840956995,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2301.0,
      "completions/max_terminated_length": 2301.0,
      "completions/mean_length": 519.9921875,
      "completions/mean_terminated_length": 519.9921875,
      "completions/min_length": 178.0,
      "completions/min_terminated_length": 178.0,
      "epoch": 0.144,
      "grad_norm": 0.004932735580950975,
      "learning_rate": 1.8055555555555557e-06,
      "loss": 0.0677,
      "num_tokens": 27559208.0,
      "reward": 1.0246297121047974,
      "reward_std": 0.10806107521057129,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/final_brier_reward_step": 0.57269686460495,
      "rewards/format_reward_step": 0.98828125,
      "step": 135
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2540166104445234,
      "aux_distill/mean_u": 0.07143832899791504,
      "aux_distill/n_active_tok": 462.75,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5743910256410257,
      "calib/avg_num_step_conf": 7.23046875,
      "calib/ece": 0.3616929133858268,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": -0.0005217948717948367,
      "calib/mean_conf": 0.09751968503937009,
      "calib/mu_c": 0.09721153846153847,
      "calib/mu_w": 0.09773333333333331,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.024881889763779527,
      "calib/std_conf": 0.1430177760216139,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2921.0,
      "completions/max_terminated_length": 2921.0,
      "completions/mean_length": 493.8828125,
      "completions/mean_terminated_length": 493.8828125,
      "completions/min_length": 131.0,
      "completions/min_terminated_length": 131.0,
      "epoch": 0.14506666666666668,
      "grad_norm": 0.005398144014179707,
      "learning_rate": 1.777777777777778e-06,
      "loss": 0.0419,
      "num_tokens": 27794130.0,
      "reward": 1.0051078796386719,
      "reward_std": 0.10781945288181305,
      "rewards/accuracy_reward_step": 0.40625,
      "rewards/final_brier_reward_step": 0.6234968900680542,
      "rewards/format_reward_step": 0.98046875,
      "step": 136
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2740769935771823,
      "aux_distill/mean_u": 0.098901419560321,
      "aux_distill/n_active_tok": 467.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5006140997297961,
      "calib/avg_num_step_conf": 7.328125,
      "calib/ece": 0.391484375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0234375,
      "calib/gap": 0.010186686317858018,
      "calib/mean_conf": 0.10179687500000001,
      "calib/mu_c": 0.10728813559322034,
      "calib/mu_w": 0.09710144927536232,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.016171875000000002,
      "calib/std_conf": 0.16214592961352553,
      "calib/step_conf_rate": 1.0,
      "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": 1113.0,
      "completions/max_terminated_length": 1113.0,
      "completions/mean_length": 458.484375,
      "completions/mean_terminated_length": 460.2823791503906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 148.0,
      "epoch": 0.14613333333333334,
      "grad_norm": 0.005362810101360083,
      "learning_rate": 1.75e-06,
      "loss": 0.031,
      "num_tokens": 28018486.0,
      "reward": 1.0311261415481567,
      "reward_std": 0.06680846959352493,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/final_brier_reward_step": 0.6013148427009583,
      "rewards/format_reward_step": 1.0,
      "step": 137
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.23668501828797162,
      "aux_distill/mean_u": 0.07159552976812263,
      "aux_distill/n_active_tok": 435.0,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4650963189013923,
      "calib/avg_num_step_conf": 6.796875,
      "calib/ece": 0.49929133858267705,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.015748031496062992,
      "calib/gap": -0.016700362387945825,
      "calib/mean_conf": 0.10070866141732283,
      "calib/mu_c": 0.0936734693877551,
      "calib/mu_w": 0.11037383177570093,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.01062992125984252,
      "calib/std_conf": 0.147430384491924,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2362.0,
      "completions/max_terminated_length": 2362.0,
      "completions/mean_length": 479.22265625,
      "completions/mean_terminated_length": 481.10198974609375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 156.0,
      "epoch": 0.1472,
      "grad_norm": 0.006265109404921532,
      "learning_rate": 1.7222222222222224e-06,
      "loss": 0.0322,
      "num_tokens": 28245503.0,
      "reward": 1.0223547220230103,
      "reward_std": 0.1057506576180458,
      "rewards/accuracy_reward_step": 0.57421875,
      "rewards/final_brier_reward_step": 0.48611563444137573,
      "rewards/format_reward_step": 0.984375,
      "step": 138
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2570781602989882,
      "aux_distill/mean_u": 0.08331463642739878,
      "aux_distill/n_active_tok": 434.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.47682389937106917,
      "calib/avg_num_step_conf": 6.81640625,
      "calib/ece": 0.500546875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.00390625,
      "calib/gap": 0.01027672955974844,
      "calib/mean_conf": 0.100078125,
      "calib/mu_c": 0.10433333333333333,
      "calib/mu_w": 0.0940566037735849,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.00734375,
      "calib/std_conf": 0.1285435000164706,
      "calib/step_conf_rate": 1.0,
      "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": 1121.0,
      "completions/max_terminated_length": 1121.0,
      "completions/mean_length": 434.38671875,
      "completions/mean_terminated_length": 436.0902099609375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 166.0,
      "epoch": 0.14826666666666666,
      "grad_norm": 0.005986573174595833,
      "learning_rate": 1.6944444444444446e-06,
      "loss": -0.0022,
      "num_tokens": 28459802.0,
      "reward": 1.0478633642196655,
      "reward_std": 0.07045100629329681,
      "rewards/accuracy_reward_step": 0.5859375,
      "rewards/final_brier_reward_step": 0.5097890496253967,
      "rewards/format_reward_step": 1.0,
      "step": 139
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24574917717836797,
      "aux_distill/mean_u": 0.0754423186694742,
      "aux_distill/n_active_tok": 439.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5346385542168675,
      "calib/avg_num_step_conf": 6.8984375,
      "calib/ece": 0.5513671875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.01171875,
      "calib/gap": 0.032148594377510026,
      "calib/mean_conf": 0.1041796875,
      "calib/mu_c": 0.11548192771084338,
      "calib/mu_w": 0.08333333333333336,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0035546875000000014,
      "calib/std_conf": 0.15045539899718569,
      "calib/step_conf_rate": 1.0,
      "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": 1043.0,
      "completions/max_terminated_length": 1043.0,
      "completions/mean_length": 466.796875,
      "completions/mean_terminated_length": 468.6274719238281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 172.0,
      "epoch": 0.14933333333333335,
      "grad_norm": 0.005342083051800728,
      "learning_rate": 1.6666666666666667e-06,
      "loss": 0.055,
      "num_tokens": 28684318.0,
      "reward": 1.0581376552581787,
      "reward_std": 0.07809313386678696,
      "rewards/accuracy_reward_step": 0.6484375,
      "rewards/final_brier_reward_step": 0.4678378999233246,
      "rewards/format_reward_step": 1.0,
      "step": 140
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25117631489410996,
      "aux_distill/mean_u": 0.09753673918011269,
      "aux_distill/n_active_tok": 453.15625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.49602630857769253,
      "calib/avg_num_step_conf": 7.06640625,
      "calib/ece": 0.5469960474308301,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.02766798418972332,
      "calib/gap": -0.007095094546451075,
      "calib/mean_conf": 0.13371541501976283,
      "calib/mu_c": 0.13121951219512196,
      "calib/mu_w": 0.13831460674157303,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.016245059288537547,
      "calib/std_conf": 0.19704243548771552,
      "calib/step_conf_rate": 1.0,
      "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": 2834.0,
      "completions/max_terminated_length": 2834.0,
      "completions/mean_length": 520.3203125,
      "completions/mean_terminated_length": 522.36083984375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 217.0,
      "epoch": 0.1504,
      "grad_norm": 0.00543976342305541,
      "learning_rate": 1.638888888888889e-06,
      "loss": 0.0539,
      "num_tokens": 28924616.0,
      "reward": 1.044323205947876,
      "reward_std": 0.12294139713048935,
      "rewards/accuracy_reward_step": 0.640625,
      "rewards/final_brier_reward_step": 0.45974022150039673,
      "rewards/format_reward_step": 0.98828125,
      "step": 141
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24302505189552903,
      "aux_distill/mean_u": 0.08167430237365858,
      "aux_distill/n_active_tok": 460.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5374961832061069,
      "calib/avg_num_step_conf": 7.24609375,
      "calib/ece": 0.40203124999999995,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.015625,
      "calib/gap": 0.02609099236641224,
      "calib/mean_conf": 0.12703125,
      "calib/mu_c": 0.13977099236641224,
      "calib/mu_w": 0.11368,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.008671875,
      "calib/std_conf": 0.1699418989638444,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1123.0,
      "completions/max_terminated_length": 1123.0,
      "completions/mean_length": 481.52734375,
      "completions/mean_terminated_length": 483.41571044921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 181.0,
      "epoch": 0.15146666666666667,
      "grad_norm": 0.0055787647143006325,
      "learning_rate": 1.6111111111111113e-06,
      "loss": 0.0218,
      "num_tokens": 29153047.0,
      "reward": 1.0430495738983154,
      "reward_std": 0.0907321497797966,
      "rewards/accuracy_reward_step": 0.51171875,
      "rewards/final_brier_reward_step": 0.5821930170059204,
      "rewards/format_reward_step": 0.9921875,
      "step": 142
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25827985629439354,
      "aux_distill/mean_u": 0.08566694257432725,
      "aux_distill/n_active_tok": 480.1875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5286016949152543,
      "calib/avg_num_step_conf": 7.45703125,
      "calib/ece": 0.4213779527559056,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.007874015748031496,
      "calib/gap": 0.03622133599202389,
      "calib/mean_conf": 0.11405511811023622,
      "calib/mu_c": 0.13088235294117648,
      "calib/mu_w": 0.09466101694915259,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.13521815602166537,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2658.0,
      "completions/max_terminated_length": 2658.0,
      "completions/mean_length": 508.015625,
      "completions/mean_terminated_length": 508.015625,
      "completions/min_length": 195.0,
      "completions/min_terminated_length": 195.0,
      "epoch": 0.15253333333333333,
      "grad_norm": 0.005299092270433903,
      "learning_rate": 1.5833333333333333e-06,
      "loss": 0.0571,
      "num_tokens": 29390435.0,
      "reward": 1.0422933101654053,
      "reward_std": 0.10097577422857285,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/final_brier_reward_step": 0.5650554895401001,
      "rewards/format_reward_step": 0.98828125,
      "step": 143
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.22756909485906363,
      "aux_distill/mean_u": 0.05125861934791381,
      "aux_distill/n_active_tok": 478.5,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5264581296839361,
      "calib/avg_num_step_conf": 7.51953125,
      "calib/ece": 0.46848307086614177,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.007874015748031496,
      "calib/gap": 0.019926014988595614,
      "calib/mean_conf": 0.15781614173228348,
      "calib/mu_c": 0.1655825806451613,
      "calib/mu_w": 0.14565656565656568,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.008031496062992128,
      "calib/std_conf": 0.16857469551445417,
      "calib/step_conf_rate": 1.0,
      "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": 1621.0,
      "completions/max_terminated_length": 1621.0,
      "completions/mean_length": 495.6484375,
      "completions/mean_terminated_length": 497.5921936035156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 135.0,
      "epoch": 0.1536,
      "grad_norm": 0.006199431139975786,
      "learning_rate": 1.5555555555555558e-06,
      "loss": 0.0044,
      "num_tokens": 29621449.0,
      "reward": 1.0659891366958618,
      "reward_std": 0.11490050703287125,
      "rewards/accuracy_reward_step": 0.60546875,
      "rewards/final_brier_reward_step": 0.5343221426010132,
      "rewards/format_reward_step": 0.9921875,
      "step": 144
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24254822148941457,
      "aux_distill/mean_u": 0.0902089617046771,
      "aux_distill/n_active_tok": 475.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5481939935064934,
      "calib/avg_num_step_conf": 7.484375,
      "calib/ece": 0.48573671875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.01953125,
      "calib/gap": 0.05436439393939396,
      "calib/mean_conf": 0.18340390625,
      "calib/mu_c": 0.2020916666666667,
      "calib/mu_w": 0.14772727272727273,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.006445312500000001,
      "calib/std_conf": 0.20833097724172275,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1139.0,
      "completions/max_terminated_length": 1139.0,
      "completions/mean_length": 463.28125,
      "completions/mean_terminated_length": 465.0980529785156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 187.0,
      "epoch": 0.15466666666666667,
      "grad_norm": 0.0066559393890202045,
      "learning_rate": 1.527777777777778e-06,
      "loss": 0.0377,
      "num_tokens": 29842753.0,
      "reward": 1.0902000665664673,
      "reward_std": 0.12901276350021362,
      "rewards/accuracy_reward_step": 0.65625,
      "rewards/final_brier_reward_step": 0.5280565023422241,
      "rewards/format_reward_step": 0.99609375,
      "step": 145
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24017924163490534,
      "aux_distill/mean_u": 0.055623048994550915,
      "aux_distill/n_active_tok": 482.0,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5867346938775511,
      "calib/avg_num_step_conf": 7.69921875,
      "calib/ece": 0.27204724409448816,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.015748031496062992,
      "calib/gap": 0.007849293563579302,
      "calib/mean_conf": 0.18456692913385828,
      "calib/mu_c": 0.18938775510204084,
      "calib/mu_w": 0.18153846153846154,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.035393700787401575,
      "calib/std_conf": 0.1884405790155179,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2372.0,
      "completions/max_terminated_length": 2372.0,
      "completions/mean_length": 501.0,
      "completions/mean_terminated_length": 502.9647216796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 215.0,
      "epoch": 0.15573333333333333,
      "grad_norm": 0.006187766790390015,
      "learning_rate": 1.5e-06,
      "loss": 0.0536,
      "num_tokens": 30078225.0,
      "reward": 1.0223695039749146,
      "reward_std": 0.12816308438777924,
      "rewards/accuracy_reward_step": 0.3828125,
      "rewards/final_brier_reward_step": 0.6775516271591187,
      "rewards/format_reward_step": 0.984375,
      "step": 146
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.23265002365224063,
      "aux_distill/mean_u": 0.07974558114182709,
      "aux_distill/n_active_tok": 506.25,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.48316215375038907,
      "calib/avg_num_step_conf": 7.91015625,
      "calib/ece": 0.3200393700787401,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.007874015748031496,
      "calib/gap": 0.001661375661375697,
      "calib/mean_conf": 0.18625984251968505,
      "calib/mu_c": 0.18714285714285717,
      "calib/mu_w": 0.18548148148148147,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.018897637795275597,
      "calib/std_conf": 0.17227229465438432,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2598.0,
      "completions/max_terminated_length": 2598.0,
      "completions/mean_length": 526.20703125,
      "completions/mean_terminated_length": 526.20703125,
      "completions/min_length": 180.0,
      "completions/min_terminated_length": 180.0,
      "epoch": 0.1568,
      "grad_norm": 0.005905737169086933,
      "learning_rate": 1.4722222222222225e-06,
      "loss": 0.0701,
      "num_tokens": 30316614.0,
      "reward": 1.0446211099624634,
      "reward_std": 0.11209353804588318,
      "rewards/accuracy_reward_step": 0.46484375,
      "rewards/final_brier_reward_step": 0.6361172199249268,
      "rewards/format_reward_step": 0.98828125,
      "step": 147
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2535096355713904,
      "aux_distill/mean_u": 0.08370382600464303,
      "aux_distill/n_active_tok": 458.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5175713523606296,
      "calib/avg_num_step_conf": 7.16796875,
      "calib/ece": 0.43325490196078437,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.023529411764705882,
      "calib/gap": 0.008258869031741778,
      "calib/mean_conf": 0.2384313725490196,
      "calib/mu_c": 0.2414110429447853,
      "calib/mu_w": 0.23315217391304352,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.01623529411764706,
      "calib/std_conf": 0.21298046075361465,
      "calib/step_conf_rate": 1.0,
      "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": 2498.0,
      "completions/max_terminated_length": 2498.0,
      "completions/mean_length": 476.1640625,
      "completions/mean_terminated_length": 476.1640625,
      "completions/min_length": 135.0,
      "completions/min_terminated_length": 135.0,
      "epoch": 0.15786666666666666,
      "grad_norm": 0.006926925852894783,
      "learning_rate": 1.4444444444444445e-06,
      "loss": 0.0402,
      "num_tokens": 30543624.0,
      "reward": 1.098899245262146,
      "reward_std": 0.13413414359092712,
      "rewards/accuracy_reward_step": 0.63671875,
      "rewards/final_brier_reward_step": 0.564985990524292,
      "rewards/format_reward_step": 0.99609375,
      "step": 148
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.23205461748875678,
      "aux_distill/mean_u": 0.06465849815899094,
      "aux_distill/n_active_tok": 519.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5840522630657663,
      "calib/avg_num_step_conf": 8.1171875,
      "calib/ece": 0.2754877470355731,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.007905138339920948,
      "calib/gap": 0.05102485621405345,
      "calib/mean_conf": 0.24925533596837945,
      "calib/mu_c": 0.2742635658914728,
      "calib/mu_w": 0.22323870967741938,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.007430830039525693,
      "calib/std_conf": 0.18631744623604815,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2487.0,
      "completions/max_terminated_length": 2487.0,
      "completions/mean_length": 577.2734375,
      "completions/mean_terminated_length": 577.2734375,
      "completions/min_length": 192.0,
      "completions/min_terminated_length": 192.0,
      "epoch": 0.15893333333333334,
      "grad_norm": 0.005738920997828245,
      "learning_rate": 1.4166666666666667e-06,
      "loss": 0.077,
      "num_tokens": 30795862.0,
      "reward": 1.0786305665969849,
      "reward_std": 0.14628009498119354,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.6650736927986145,
      "rewards/format_reward_step": 0.98828125,
      "step": 149
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.26180982729420066,
      "aux_distill/mean_u": 0.08911000792044878,
      "aux_distill/n_active_tok": 464.90625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5819143446852425,
      "calib/avg_num_step_conf": 7.26171875,
      "calib/ece": 0.3128740157480314,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.01968503937007874,
      "calib/gap": 0.061225490196078425,
      "calib/mean_conf": 0.2979133858267717,
      "calib/mu_c": 0.32249999999999995,
      "calib/mu_w": 0.26127450980392153,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.006181102362204721,
      "calib/std_conf": 0.21724649535051074,
      "calib/step_conf_rate": 1.0,
      "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": 2426.0,
      "completions/max_terminated_length": 2426.0,
      "completions/mean_length": 479.67578125,
      "completions/mean_terminated_length": 479.67578125,
      "completions/min_length": 182.0,
      "completions/min_terminated_length": 182.0,
      "epoch": 0.16,
      "grad_norm": 0.006487760692834854,
      "learning_rate": 1.3888888888888892e-06,
      "loss": 0.0327,
      "num_tokens": 31023619.0,
      "reward": 1.116228699684143,
      "reward_std": 0.15358006954193115,
      "rewards/accuracy_reward_step": 0.59375,
      "rewards/final_brier_reward_step": 0.6465198993682861,
      "rewards/format_reward_step": 0.9921875,
      "step": 150
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2628675187006593,
      "aux_distill/mean_u": 0.08701564027971015,
      "aux_distill/n_active_tok": 504.78125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.532805645622735,
      "calib/avg_num_step_conf": 7.93359375,
      "calib/ece": 0.19424685039370077,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.015748031496062992,
      "calib/gap": 0.02318017038591136,
      "calib/mean_conf": 0.3063043307086614,
      "calib/mu_c": 0.31971962616822425,
      "calib/mu_w": 0.2965394557823129,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.03964566929133858,
      "calib/std_conf": 0.2062846493465961,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1406.0,
      "completions/max_terminated_length": 1406.0,
      "completions/mean_length": 506.50390625,
      "completions/mean_terminated_length": 508.490234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 187.0,
      "epoch": 0.16106666666666666,
      "grad_norm": 0.00684371730312705,
      "learning_rate": 1.3611111111111112e-06,
      "loss": 0.0143,
      "num_tokens": 31260308.0,
      "reward": 1.0545872449874878,
      "reward_std": 0.16151979565620422,
      "rewards/accuracy_reward_step": 0.41796875,
      "rewards/final_brier_reward_step": 0.7029244899749756,
      "rewards/format_reward_step": 0.98828125,
      "step": 151
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24851590069010854,
      "aux_distill/mean_u": 0.06157251443173721,
      "aux_distill/n_active_tok": 468.65625,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5187788906009245,
      "calib/avg_num_step_conf": 7.51171875,
      "calib/ece": 0.3058164,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.02,
      "calib/gap": -0.0011815485362095535,
      "calib/mean_conf": 0.3052236,
      "calib/mu_c": 0.3046659090909091,
      "calib/mu_w": 0.3058474576271187,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.04151999999999999,
      "calib/std_conf": 0.2093360103829248,
      "calib/step_conf_rate": 1.0,
      "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": 2837.0,
      "completions/max_terminated_length": 2837.0,
      "completions/mean_length": 511.64453125,
      "completions/mean_terminated_length": 515.6732177734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 199.0,
      "epoch": 0.16213333333333332,
      "grad_norm": 0.0067654200829565525,
      "learning_rate": 1.3333333333333334e-06,
      "loss": 0.0823,
      "num_tokens": 31496681.0,
      "reward": 1.0633323192596436,
      "reward_std": 0.19225209951400757,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/final_brier_reward_step": 0.6383833885192871,
      "rewards/format_reward_step": 0.97265625,
      "step": 152
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2772968136705458,
      "aux_distill/mean_u": 0.10214465292753641,
      "aux_distill/n_active_tok": 470.53125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5065810907669251,
      "calib/avg_num_step_conf": 7.21484375,
      "calib/ece": 0.3198753968253968,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.015873015873015872,
      "calib/gap": -0.00302903499055196,
      "calib/mean_conf": 0.32607698412698416,
      "calib/mu_c": 0.3248389261744966,
      "calib/mu_w": 0.32786796116504857,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.02734126984126984,
      "calib/std_conf": 0.21308511290863438,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2955.0,
      "completions/max_terminated_length": 2955.0,
      "completions/mean_length": 526.5625,
      "completions/mean_terminated_length": 526.5625,
      "completions/min_length": 183.0,
      "completions/min_terminated_length": 183.0,
      "epoch": 0.1632,
      "grad_norm": 0.00628744438290596,
      "learning_rate": 1.3055555555555556e-06,
      "loss": 0.1055,
      "num_tokens": 31738801.0,
      "reward": 1.0987610816955566,
      "reward_std": 0.1717906892299652,
      "rewards/accuracy_reward_step": 0.58203125,
      "rewards/final_brier_reward_step": 0.6311159133911133,
      "rewards/format_reward_step": 0.984375,
      "step": 153
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25234381505288184,
      "aux_distill/mean_u": 0.08080222799912855,
      "aux_distill/n_active_tok": 431.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.48318968152087544,
      "calib/avg_num_step_conf": 6.77734375,
      "calib/ece": 0.229622265625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.03125,
      "calib/gap": -0.022947973592517823,
      "calib/mean_conf": 0.372955859375,
      "calib/mu_c": 0.36193007518796994,
      "calib/mu_w": 0.38487804878048776,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.041523437499999996,
      "calib/std_conf": 0.21237021947984544,
      "calib/step_conf_rate": 1.0,
      "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": 1556.0,
      "completions/max_terminated_length": 1556.0,
      "completions/mean_length": 476.73046875,
      "completions/mean_terminated_length": 478.60003662109375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 199.0,
      "epoch": 0.16426666666666667,
      "grad_norm": 0.006957808043807745,
      "learning_rate": 1.2777777777777779e-06,
      "loss": 0.034,
      "num_tokens": 31965284.0,
      "reward": 1.095935344696045,
      "reward_std": 0.15885654091835022,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/final_brier_reward_step": 0.6723395586013794,
      "rewards/format_reward_step": 1.0,
      "step": 154
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24342874670401216,
      "aux_distill/mean_u": 0.0742800901495456,
      "aux_distill/n_active_tok": 441.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4895392278953922,
      "calib/avg_num_step_conf": 6.921875,
      "calib/ece": 0.18644921875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.015625,
      "calib/gap": -0.0059505603985056665,
      "calib/mean_conf": 0.37456640625000004,
      "calib/mu_c": 0.3711727272727272,
      "calib/mu_w": 0.3771232876712329,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0656640625,
      "calib/std_conf": 0.20673298440309412,
      "calib/step_conf_rate": 1.0,
      "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": 1042.0,
      "completions/max_terminated_length": 1042.0,
      "completions/mean_length": 457.39453125,
      "completions/mean_terminated_length": 459.1882629394531,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 185.0,
      "epoch": 0.16533333333333333,
      "grad_norm": 0.008058539591729641,
      "learning_rate": 1.25e-06,
      "loss": 0.0351,
      "num_tokens": 32189593.0,
      "reward": 1.0679690837860107,
      "reward_std": 0.1546422243118286,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.7062505483627319,
      "rewards/format_reward_step": 1.0,
      "step": 155
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24015296366997063,
      "aux_distill/mean_u": 0.07636279268520298,
      "aux_distill/n_active_tok": 467.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.53173828125,
      "calib/avg_num_step_conf": 7.375,
      "calib/ece": 0.21515625000000005,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.02734375,
      "calib/gap": 0.016718750000000004,
      "calib/mean_conf": 0.39187500000000003,
      "calib/mu_c": 0.40023437500000003,
      "calib/mu_w": 0.383515625,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.05351562500000002,
      "calib/std_conf": 0.21984280605696427,
      "calib/step_conf_rate": 1.0,
      "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": 1665.0,
      "completions/max_terminated_length": 1665.0,
      "completions/mean_length": 470.7890625,
      "completions/mean_terminated_length": 472.63531494140625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 188.0,
      "epoch": 0.1664,
      "grad_norm": 0.0072465743869543076,
      "learning_rate": 1.2222222222222223e-06,
      "loss": 0.0173,
      "num_tokens": 32414875.0,
      "reward": 1.0991687774658203,
      "reward_std": 0.1531982123851776,
      "rewards/accuracy_reward_step": 0.5,
      "rewards/final_brier_reward_step": 0.6983374953269958,
      "rewards/format_reward_step": 1.0,
      "step": 156
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2766073530074209,
      "aux_distill/mean_u": 0.09899811663300234,
      "aux_distill/n_active_tok": 461.625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5205314009661836,
      "calib/avg_num_step_conf": 7.17578125,
      "calib/ece": 0.18030277777777773,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.027777777777777776,
      "calib/gap": 0.01946998474446987,
      "calib/mean_conf": 0.41803055555555557,
      "calib/mu_c": 0.4268384057971014,
      "calib/mu_w": 0.40736842105263155,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.025357142857142856,
      "calib/std_conf": 0.20073677321784758,
      "calib/step_conf_rate": 1.0,
      "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": 2958.0,
      "completions/max_terminated_length": 2958.0,
      "completions/mean_length": 498.0625,
      "completions/mean_terminated_length": 500.0157165527344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 156.0,
      "epoch": 0.16746666666666668,
      "grad_norm": 0.006882913876324892,
      "learning_rate": 1.1944444444444446e-06,
      "loss": 0.04,
      "num_tokens": 32646107.0,
      "reward": 1.1105782985687256,
      "reward_std": 0.1789938509464264,
      "rewards/accuracy_reward_step": 0.54296875,
      "rewards/final_brier_reward_step": 0.6938129663467407,
      "rewards/format_reward_step": 0.984375,
      "step": 157
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2523889895528555,
      "aux_distill/mean_u": 0.08844801318115658,
      "aux_distill/n_active_tok": 426.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.44285187007874016,
      "calib/avg_num_step_conf": 6.66796875,
      "calib/ece": 0.22184588235294125,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.047058823529411764,
      "calib/gap": -0.04047578740157476,
      "calib/mean_conf": 0.4496835294117647,
      "calib/mu_c": 0.42952500000000005,
      "calib/mu_w": 0.4700007874015748,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.08478431372549022,
      "calib/std_conf": 0.2085711531463499,
      "calib/step_conf_rate": 1.0,
      "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": 2191.0,
      "completions/max_terminated_length": 2191.0,
      "completions/mean_length": 474.046875,
      "completions/mean_terminated_length": 474.046875,
      "completions/min_length": 170.0,
      "completions/min_terminated_length": 170.0,
      "epoch": 0.16853333333333334,
      "grad_norm": 0.007169242948293686,
      "learning_rate": 1.1666666666666668e-06,
      "loss": 0.0458,
      "num_tokens": 32872703.0,
      "reward": 1.088477611541748,
      "reward_std": 0.19478802382946014,
      "rewards/accuracy_reward_step": 0.5,
      "rewards/final_brier_reward_step": 0.6808613538742065,
      "rewards/format_reward_step": 0.99609375,
      "step": 158
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2364857776556164,
      "aux_distill/mean_u": 0.07033274390036429,
      "aux_distill/n_active_tok": 448.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5201918377321604,
      "calib/avg_num_step_conf": 7.0859375,
      "calib/ece": 0.182526953125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.046875,
      "calib/gap": 0.00289982893450641,
      "calib/mean_conf": 0.43778554687500004,
      "calib/mu_c": 0.4391901515151515,
      "calib/mu_w": 0.4362903225806451,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.05234375000000001,
      "calib/std_conf": 0.2133529488140946,
      "calib/step_conf_rate": 1.0,
      "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": 1494.0,
      "completions/max_terminated_length": 1494.0,
      "completions/mean_length": 444.61328125,
      "completions/mean_terminated_length": 446.3569030761719,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 179.0,
      "epoch": 0.1696,
      "grad_norm": 0.007922896184027195,
      "learning_rate": 1.138888888888889e-06,
      "loss": 0.0248,
      "num_tokens": 33091308.0,
      "reward": 1.1078696250915527,
      "reward_std": 0.1769241839647293,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/final_brier_reward_step": 0.7001141309738159,
      "rewards/format_reward_step": 1.0,
      "step": 159
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24150769133120775,
      "aux_distill/mean_u": 0.06557536838335866,
      "aux_distill/n_active_tok": 482.0,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.49964607464607463,
      "calib/avg_num_step_conf": 7.5390625,
      "calib/ece": 0.20342390438247007,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0398406374501992,
      "calib/gap": 0.00861539897039898,
      "calib/mean_conf": 0.44542071713147413,
      "calib/mu_c": 0.44923071428571426,
      "calib/mu_w": 0.4406153153153153,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.04553784860557768,
      "calib/std_conf": 0.20457986052335805,
      "calib/step_conf_rate": 1.0,
      "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": 2763.0,
      "completions/max_terminated_length": 2763.0,
      "completions/mean_length": 508.8671875,
      "completions/mean_terminated_length": 508.8671875,
      "completions/min_length": 181.0,
      "completions/min_terminated_length": 181.0,
      "epoch": 0.17066666666666666,
      "grad_norm": 0.007158962078392506,
      "learning_rate": 1.111111111111111e-06,
      "loss": 0.0566,
      "num_tokens": 33326418.0,
      "reward": 1.1083617210388184,
      "reward_std": 0.17632320523262024,
      "rewards/accuracy_reward_step": 0.546875,
      "rewards/final_brier_reward_step": 0.6893797516822815,
      "rewards/format_reward_step": 0.98046875,
      "step": 160
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24881407851353288,
      "aux_distill/mean_u": 0.07915090880055803,
      "aux_distill/n_active_tok": 419.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5378493229616825,
      "calib/avg_num_step_conf": 6.57421875,
      "calib/ece": 0.27948046875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.02734375,
      "calib/gap": 0.020856669547680617,
      "calib/mean_conf": 0.45270703125,
      "calib/mu_c": 0.459061797752809,
      "calib/mu_w": 0.43820512820512836,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.018437499999999992,
      "calib/std_conf": 0.20979708413860648,
      "calib/step_conf_rate": 1.0,
      "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": 1202.0,
      "completions/max_terminated_length": 1202.0,
      "completions/mean_length": 453.00390625,
      "completions/mean_terminated_length": 454.7804260253906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 174.0,
      "epoch": 0.17173333333333332,
      "grad_norm": 0.007694258354604244,
      "learning_rate": 1.0833333333333335e-06,
      "loss": 0.0477,
      "num_tokens": 33546307.0,
      "reward": 1.1947121620178223,
      "reward_std": 0.16380652785301208,
      "rewards/accuracy_reward_step": 0.6953125,
      "rewards/final_brier_reward_step": 0.6941118240356445,
      "rewards/format_reward_step": 1.0,
      "step": 161
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.23428629664704204,
      "aux_distill/mean_u": 0.07806861418678401,
      "aux_distill/n_active_tok": 445.375,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.563063973063973,
      "calib/avg_num_step_conf": 6.97265625,
      "calib/ece": 0.2307843137254902,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.023529411764705882,
      "calib/gap": 0.0368787878787879,
      "calib/mean_conf": 0.46952941176470586,
      "calib/mu_c": 0.48254545454545456,
      "calib/mu_w": 0.44566666666666666,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.026627450980392167,
      "calib/std_conf": 0.20145493856493812,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1177.0,
      "completions/max_terminated_length": 1177.0,
      "completions/mean_length": 439.0078125,
      "completions/mean_terminated_length": 440.72943115234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 194.0,
      "epoch": 0.1728,
      "grad_norm": 0.008261593990027905,
      "learning_rate": 1.0555555555555557e-06,
      "loss": 0.0115,
      "num_tokens": 33762837.0,
      "reward": 1.173771619796753,
      "reward_std": 0.1813250035047531,
      "rewards/accuracy_reward_step": 0.6484375,
      "rewards/final_brier_reward_step": 0.7108246088027954,
      "rewards/format_reward_step": 0.98828125,
      "step": 162
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.257523697335273,
      "aux_distill/mean_u": 0.08217107906337166,
      "aux_distill/n_active_tok": 482.46875,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5485661424606846,
      "calib/avg_num_step_conf": 7.58984375,
      "calib/ece": 0.144507421875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0390625,
      "calib/gap": 0.040020875732346584,
      "calib/mean_conf": 0.478305078125,
      "calib/mu_c": 0.5003478260869566,
      "calib/mu_w": 0.46032695035461,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.08679687500000001,
      "calib/std_conf": 0.21256887379629488,
      "calib/step_conf_rate": 1.0,
      "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": 1571.0,
      "completions/max_terminated_length": 1571.0,
      "completions/mean_length": 497.17578125,
      "completions/mean_terminated_length": 499.1255187988281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 168.0,
      "epoch": 0.17386666666666667,
      "grad_norm": 0.007228810340166092,
      "learning_rate": 1.0277777777777777e-06,
      "loss": 0.0138,
      "num_tokens": 33994946.0,
      "reward": 1.087785005569458,
      "reward_std": 0.17991483211517334,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.7263512015342712,
      "rewards/format_reward_step": 1.0,
      "step": 163
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24482636875472963,
      "aux_distill/mean_u": 0.06283773442355123,
      "aux_distill/n_active_tok": 508.28125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4689828614008942,
      "calib/avg_num_step_conf": 7.94140625,
      "calib/ece": 0.19609488188976373,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.04330708661417323,
      "calib/gap": -0.03212917287630396,
      "calib/mean_conf": 0.47059803149606294,
      "calib/mu_c": 0.45516590909090915,
      "calib/mu_w": 0.4872950819672131,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.073503937007874,
      "calib/std_conf": 0.1955032348546813,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2961.0,
      "completions/max_terminated_length": 2961.0,
      "completions/mean_length": 545.29296875,
      "completions/mean_terminated_length": 545.29296875,
      "completions/min_length": 193.0,
      "completions/min_terminated_length": 193.0,
      "epoch": 0.17493333333333333,
      "grad_norm": 0.007839228957891464,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0301,
      "num_tokens": 34240677.0,
      "reward": 1.094243049621582,
      "reward_std": 0.19971774518489838,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/final_brier_reward_step": 0.6845799088478088,
      "rewards/format_reward_step": 0.98828125,
      "step": 164
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.27021268755197525,
      "aux_distill/mean_u": 0.07547623212198123,
      "aux_distill/n_active_tok": 471.9375,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.49396135265700486,
      "calib/avg_num_step_conf": 7.296875,
      "calib/ece": 0.15078745098039217,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.03137254901960784,
      "calib/gap": -0.006471794871794889,
      "calib/mean_conf": 0.4786305882352941,
      "calib/mu_c": 0.47512820512820514,
      "calib/mu_w": 0.48160000000000003,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.08529725490196079,
      "calib/std_conf": 0.18865541199797206,
      "calib/step_conf_rate": 1.0,
      "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": 1387.0,
      "completions/max_terminated_length": 1387.0,
      "completions/mean_length": 510.9765625,
      "completions/mean_terminated_length": 512.9804077148438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 185.0,
      "epoch": 0.176,
      "grad_norm": 0.00910587515681982,
      "learning_rate": 9.722222222222224e-07,
      "loss": 0.0356,
      "num_tokens": 34477063.0,
      "reward": 1.0814200639724731,
      "reward_std": 0.16908253729343414,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/final_brier_reward_step": 0.7097151875495911,
      "rewards/format_reward_step": 0.99609375,
      "step": 165
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2651129311416298,
      "aux_distill/mean_u": 0.09152722727924217,
      "aux_distill/n_active_tok": 490.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5931132075471699,
      "calib/avg_num_step_conf": 7.69140625,
      "calib/ece": 0.15345742187500003,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.03125,
      "calib/gap": 0.0671997106918239,
      "calib/mean_conf": 0.501464453125,
      "calib/mu_c": 0.5292893333333333,
      "calib/mu_w": 0.4620896226415094,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03449218750000001,
      "calib/std_conf": 0.19940186717996694,
      "calib/step_conf_rate": 1.0,
      "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": 1826.0,
      "completions/max_terminated_length": 1826.0,
      "completions/mean_length": 508.7421875,
      "completions/mean_terminated_length": 510.7372741699219,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 183.0,
      "epoch": 0.17706666666666668,
      "grad_norm": 0.008176633156836033,
      "learning_rate": 9.444444444444445e-07,
      "loss": 0.0261,
      "num_tokens": 34713485.0,
      "reward": 1.1645166873931885,
      "reward_std": 0.16638818383216858,
      "rewards/accuracy_reward_step": 0.5859375,
      "rewards/final_brier_reward_step": 0.7430957555770874,
      "rewards/format_reward_step": 1.0,
      "step": 166
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2851410242728889,
      "aux_distill/mean_u": 0.10715140228125174,
      "aux_distill/n_active_tok": 474.5,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5361252979230507,
      "calib/avg_num_step_conf": 7.4140625,
      "calib/ece": 0.22204724409448814,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.04330708661417323,
      "calib/gap": 0.01943343547837939,
      "calib/mean_conf": 0.5154330708661418,
      "calib/mu_c": 0.5222424242424243,
      "calib/mu_w": 0.5028089887640449,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.043937007874015735,
      "calib/std_conf": 0.1914254687532988,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3027.0,
      "completions/max_terminated_length": 3027.0,
      "completions/mean_length": 492.58203125,
      "completions/mean_terminated_length": 492.58203125,
      "completions/min_length": 1.0,
      "completions/min_terminated_length": 1.0,
      "epoch": 0.17813333333333334,
      "grad_norm": 0.007018796168267727,
      "learning_rate": 9.166666666666666e-07,
      "loss": 0.0462,
      "num_tokens": 34945194.0,
      "reward": 1.1749070882797241,
      "reward_std": 0.17637360095977783,
      "rewards/accuracy_reward_step": 0.64453125,
      "rewards/final_brier_reward_step": 0.7170015573501587,
      "rewards/format_reward_step": 0.98828125,
      "step": 167
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25646522268652916,
      "aux_distill/mean_u": 0.0982307983881972,
      "aux_distill/n_active_tok": 507.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5028800400701228,
      "calib/avg_num_step_conf": 7.93359375,
      "calib/ece": 0.16849169960474308,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.03162055335968379,
      "calib/gap": 0.0014297520661158414,
      "calib/mean_conf": 0.5114292490118577,
      "calib/mu_c": 0.5121752066115703,
      "calib/mu_w": 0.5107454545454545,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.10083003952569172,
      "calib/std_conf": 0.1999903553361993,
      "calib/step_conf_rate": 1.0,
      "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": 2440.0,
      "completions/max_terminated_length": 2440.0,
      "completions/mean_length": 529.5,
      "completions/mean_terminated_length": 529.5,
      "completions/min_length": 172.0,
      "completions/min_terminated_length": 172.0,
      "epoch": 0.1792,
      "grad_norm": 0.00782513152807951,
      "learning_rate": 8.88888888888889e-07,
      "loss": 0.0136,
      "num_tokens": 35185418.0,
      "reward": 1.085259199142456,
      "reward_std": 0.22317655384540558,
      "rewards/accuracy_reward_step": 0.48046875,
      "rewards/final_brier_reward_step": 0.7017684578895569,
      "rewards/format_reward_step": 0.98828125,
      "step": 168
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2343459421535954,
      "aux_distill/mean_u": 0.06730843714362388,
      "aux_distill/n_active_tok": 472.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4648706896551724,
      "calib/avg_num_step_conf": 7.40625,
      "calib/ece": 0.16858476562499997,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0390625,
      "calib/gap": -0.014609975369458073,
      "calib/mean_conf": 0.49805585937499997,
      "calib/mu_c": 0.4914357142857143,
      "calib/mu_w": 0.5060456896551724,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.059882812500000014,
      "calib/std_conf": 0.17959830494061663,
      "calib/step_conf_rate": 1.0,
      "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": 1085.0,
      "completions/max_terminated_length": 1085.0,
      "completions/mean_length": 511.5546875,
      "completions/mean_terminated_length": 513.560791015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 186.0,
      "epoch": 0.18026666666666666,
      "grad_norm": 0.007080877665430307,
      "learning_rate": 8.611111111111112e-07,
      "loss": 0.0235,
      "num_tokens": 35420560.0,
      "reward": 1.128596305847168,
      "reward_std": 0.1757933497428894,
      "rewards/accuracy_reward_step": 0.546875,
      "rewards/final_brier_reward_step": 0.7103176116943359,
      "rewards/format_reward_step": 1.0,
      "step": 169
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24169203732162714,
      "aux_distill/mean_u": 0.07235590006785955,
      "aux_distill/n_active_tok": 500.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.48724048442906565,
      "calib/avg_num_step_conf": 7.82421875,
      "calib/ece": 0.18541176470588233,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.07058823529411765,
      "calib/gap": 0.002457983193277502,
      "calib/mean_conf": 0.5169411764705882,
      "calib/mu_c": 0.5180882352941177,
      "calib/mu_w": 0.5156302521008402,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.08450980392156862,
      "calib/std_conf": 0.20376855496070237,
      "calib/step_conf_rate": 1.0,
      "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": 2780.0,
      "completions/max_terminated_length": 2780.0,
      "completions/mean_length": 520.23828125,
      "completions/mean_terminated_length": 520.23828125,
      "completions/min_length": 186.0,
      "completions/min_terminated_length": 186.0,
      "epoch": 0.18133333333333335,
      "grad_norm": 0.007690028753131628,
      "learning_rate": 8.333333333333333e-07,
      "loss": 0.0464,
      "num_tokens": 35657893.0,
      "reward": 1.1175562143325806,
      "reward_std": 0.18893221020698547,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/final_brier_reward_step": 0.7077687978744507,
      "rewards/format_reward_step": 0.99609375,
      "step": 170
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.23977546906098723,
      "aux_distill/mean_u": 0.07869014971729088,
      "aux_distill/n_active_tok": 492.53125,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5743453579930415,
      "calib/avg_num_step_conf": 7.74609375,
      "calib/ece": 0.11351601562499998,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.06640625,
      "calib/gap": 0.04610310077519375,
      "calib/mean_conf": 0.529531640625,
      "calib/mu_c": 0.5524031007751937,
      "calib/mu_w": 0.5063,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.069570703125,
      "calib/std_conf": 0.19799818352338072,
      "calib/step_conf_rate": 1.0,
      "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": 1500.0,
      "completions/max_terminated_length": 1500.0,
      "completions/mean_length": 494.609375,
      "completions/mean_terminated_length": 496.5490417480469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 148.0,
      "epoch": 0.1824,
      "grad_norm": 0.008125293999910355,
      "learning_rate": 8.055555555555557e-07,
      "loss": 0.0251,
      "num_tokens": 35891409.0,
      "reward": 1.1185557842254639,
      "reward_std": 0.19421759247779846,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.7332054376602173,
      "rewards/format_reward_step": 1.0,
      "step": 171
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2299743457697332,
      "aux_distill/mean_u": 0.06312461743109235,
      "aux_distill/n_active_tok": 474.625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5154698242933536,
      "calib/avg_num_step_conf": 7.43359375,
      "calib/ece": 0.16480390624999997,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.03515625,
      "calib/gap": 0.006018665648077404,
      "calib/mean_conf": 0.53894609375,
      "calib/mu_c": 0.5413441558441558,
      "calib/mu_w": 0.5353254901960784,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.05109375000000001,
      "calib/std_conf": 0.18868068777134087,
      "calib/step_conf_rate": 1.0,
      "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": 1732.0,
      "completions/max_terminated_length": 1732.0,
      "completions/mean_length": 495.3671875,
      "completions/mean_terminated_length": 497.3098449707031,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 161.0,
      "epoch": 0.18346666666666667,
      "grad_norm": 0.00744143221527338,
      "learning_rate": 7.777777777777779e-07,
      "loss": 0.0229,
      "num_tokens": 36121575.0,
      "reward": 1.1626207828521729,
      "reward_std": 0.19035744667053223,
      "rewards/accuracy_reward_step": 0.6015625,
      "rewards/final_brier_reward_step": 0.7236789464950562,
      "rewards/format_reward_step": 1.0,
      "step": 172
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.22252439288422465,
      "aux_distill/mean_u": 0.07546641405015214,
      "aux_distill/n_active_tok": 527.53125,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.5107060755336617,
      "calib/avg_num_step_conf": 8.24609375,
      "calib/ece": 0.1567344,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.036,
      "calib/gap": 0.007666469622331862,
      "calib/mean_conf": 0.5260655999999999,
      "calib/mu_c": 0.5292855172413794,
      "calib/mu_w": 0.5216190476190475,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.05140000000000002,
      "calib/std_conf": 0.18500178468501324,
      "calib/step_conf_rate": 1.0,
      "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": 2847.0,
      "completions/max_terminated_length": 2847.0,
      "completions/mean_length": 577.44140625,
      "completions/mean_terminated_length": 577.44140625,
      "completions/min_length": 158.0,
      "completions/min_terminated_length": 158.0,
      "epoch": 0.18453333333333333,
      "grad_norm": 0.007290961220860481,
      "learning_rate": 7.5e-07,
      "loss": 0.0186,
      "num_tokens": 36372560.0,
      "reward": 1.1172091960906982,
      "reward_std": 0.21927499771118164,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/final_brier_reward_step": 0.699262261390686,
      "rewards/format_reward_step": 0.96875,
      "step": 173
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.26341936527751386,
      "aux_distill/mean_u": 0.08537082247943154,
      "aux_distill/n_active_tok": 507.21875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.48029110444777606,
      "calib/avg_num_step_conf": 7.9453125,
      "calib/ece": 0.1985464566929134,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.047244094488188976,
      "calib/gap": -0.003500662168915536,
      "calib/mean_conf": 0.5084614173228347,
      "calib/mu_c": 0.5065594827586207,
      "calib/mu_w": 0.5100601449275363,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.12515748031496068,
      "calib/std_conf": 0.20398919361035334,
      "calib/step_conf_rate": 1.0,
      "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": 1966.0,
      "completions/max_terminated_length": 1966.0,
      "completions/mean_length": 533.25,
      "completions/mean_terminated_length": 535.3411865234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 186.0,
      "epoch": 0.1856,
      "grad_norm": 0.0074493614956736565,
      "learning_rate": 7.222222222222222e-07,
      "loss": 0.0166,
      "num_tokens": 36613304.0,
      "reward": 1.0728223323822021,
      "reward_std": 0.22512751817703247,
      "rewards/accuracy_reward_step": 0.453125,
      "rewards/final_brier_reward_step": 0.7003322839736938,
      "rewards/format_reward_step": 0.9921875,
      "step": 174
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2397211652714759,
      "aux_distill/mean_u": 0.07551316888469718,
      "aux_distill/n_active_tok": 517.5,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5412217692357472,
      "calib/avg_num_step_conf": 8.15234375,
      "calib/ece": 0.1779350393700788,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.07480314960629922,
      "calib/gap": 0.02992464246424653,
      "calib/mean_conf": 0.5439547244094488,
      "calib/mu_c": 0.561980198019802,
      "calib/mu_w": 0.5320555555555555,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.16212598425196853,
      "calib/std_conf": 0.20303705617070827,
      "calib/step_conf_rate": 1.0,
      "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": 1543.0,
      "completions/max_terminated_length": 1543.0,
      "completions/mean_length": 500.96484375,
      "completions/mean_terminated_length": 504.9094543457031,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 198.0,
      "epoch": 0.18666666666666668,
      "grad_norm": 0.007952606305480003,
      "learning_rate": 6.944444444444446e-07,
      "loss": 0.0231,
      "num_tokens": 36847375.0,
      "reward": 1.0466676950454712,
      "reward_std": 0.202549010515213,
      "rewards/accuracy_reward_step": 0.39453125,
      "rewards/final_brier_reward_step": 0.7066166400909424,
      "rewards/format_reward_step": 0.9921875,
      "step": 175
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.27991568110883236,
      "aux_distill/mean_u": 0.10163999840663654,
      "aux_distill/n_active_tok": 489.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5015845398620518,
      "calib/avg_num_step_conf": 7.6640625,
      "calib/ece": 0.14553464566929142,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.027559055118110236,
      "calib/gap": 0.004988616168520532,
      "calib/mean_conf": 0.5305283464566929,
      "calib/mu_c": 0.5331404958677686,
      "calib/mu_w": 0.528151879699248,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.09984251968503945,
      "calib/std_conf": 0.17561454229133455,
      "calib/step_conf_rate": 1.0,
      "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": 1290.0,
      "completions/max_terminated_length": 1290.0,
      "completions/mean_length": 499.60546875,
      "completions/mean_terminated_length": 501.5647277832031,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 206.0,
      "epoch": 0.18773333333333334,
      "grad_norm": 0.006957260426133871,
      "learning_rate": 6.666666666666667e-07,
      "loss": 0.0035,
      "num_tokens": 37079338.0,
      "reward": 1.0912022590637207,
      "reward_std": 0.2196333408355713,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.713654637336731,
      "rewards/format_reward_step": 0.9921875,
      "step": 176
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.22919973416719586,
      "aux_distill/mean_u": 0.0725726520841592,
      "aux_distill/n_active_tok": 533.5625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5201538461538462,
      "calib/avg_num_step_conf": 8.33984375,
      "calib/ece": 0.14137254901960786,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.050980392156862744,
      "calib/gap": 0.017147692307692197,
      "calib/mean_conf": 0.5409019607843137,
      "calib/mu_c": 0.5493076923076923,
      "calib/mu_w": 0.5321600000000001,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.08623529411764708,
      "calib/std_conf": 0.1974850954814864,
      "calib/step_conf_rate": 1.0,
      "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": 2626.0,
      "completions/max_terminated_length": 2626.0,
      "completions/mean_length": 527.80078125,
      "completions/mean_terminated_length": 527.80078125,
      "completions/min_length": 178.0,
      "completions/min_terminated_length": 178.0,
      "epoch": 0.1888,
      "grad_norm": 0.007587118539959192,
      "learning_rate": 6.388888888888889e-07,
      "loss": 0.0014,
      "num_tokens": 37318287.0,
      "reward": 1.1098990440368652,
      "reward_std": 0.18522927165031433,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.7158917784690857,
      "rewards/format_reward_step": 0.99609375,
      "step": 177
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2627130839973688,
      "aux_distill/mean_u": 0.09110170433559824,
      "aux_distill/n_active_tok": 548.5625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5410313192027839,
      "calib/avg_num_step_conf": 8.53125,
      "calib/ece": 0.11646338582677156,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.031496062992125984,
      "calib/gap": 0.023774419487503917,
      "calib/mean_conf": 0.5405444881889764,
      "calib/mu_c": 0.550746896551724,
      "calib/mu_w": 0.5269724770642201,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.043070866141732264,
      "calib/std_conf": 0.17757437819813585,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2580.0,
      "completions/max_terminated_length": 2580.0,
      "completions/mean_length": 527.83984375,
      "completions/mean_terminated_length": 527.83984375,
      "completions/min_length": 172.0,
      "completions/min_terminated_length": 172.0,
      "epoch": 0.18986666666666666,
      "grad_norm": 0.0069207060150802135,
      "learning_rate": 6.111111111111112e-07,
      "loss": 0.0427,
      "num_tokens": 37559486.0,
      "reward": 1.143537998199463,
      "reward_std": 0.20619063079357147,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/final_brier_reward_step": 0.7284823060035706,
      "rewards/format_reward_step": 0.9921875,
      "step": 178
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.26747484086081386,
      "aux_distill/mean_u": 0.08128448746610327,
      "aux_distill/n_active_tok": 508.625,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5576569477278697,
      "calib/avg_num_step_conf": 7.9453125,
      "calib/ece": 0.12248915662650602,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.05220883534136546,
      "calib/gap": 0.04837458628841618,
      "calib/mean_conf": 0.5564650602409639,
      "calib/mu_c": 0.5774468085106385,
      "calib/mu_w": 0.5290722222222223,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.97265625,
      "calib/nonempty_step_conf_rate": 0.97265625,
      "calib/pce": 0.056344578313253,
      "calib/std_conf": 0.18966318239214708,
      "calib/step_conf_rate": 0.97265625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2277.0,
      "completions/max_terminated_length": 2277.0,
      "completions/mean_length": 523.2265625,
      "completions/mean_terminated_length": 525.2784423828125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 140.0,
      "epoch": 0.19093333333333334,
      "grad_norm": 0.008067484013736248,
      "learning_rate": 5.833333333333334e-07,
      "loss": 0.0513,
      "num_tokens": 37799696.0,
      "reward": 1.1226156949996948,
      "reward_std": 0.21316027641296387,
      "rewards/accuracy_reward_step": 0.55078125,
      "rewards/final_brier_reward_step": 0.7217940092086792,
      "rewards/format_reward_step": 0.97265625,
      "step": 179
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2592920111492276,
      "aux_distill/mean_u": 0.10416253591956139,
      "aux_distill/n_active_tok": 583.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5327018633540372,
      "calib/avg_num_step_conf": 9.11328125,
      "calib/ece": 0.12081411764705885,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.07450980392156863,
      "calib/gap": 0.027076149068323008,
      "calib/mean_conf": 0.5397349019607842,
      "calib/mu_c": 0.5519457142857144,
      "calib/mu_w": 0.5248695652173914,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.05576470588235296,
      "calib/std_conf": 0.18514426637108808,
      "calib/step_conf_rate": 1.0,
      "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": 2586.0,
      "completions/max_terminated_length": 2586.0,
      "completions/mean_length": 607.265625,
      "completions/mean_terminated_length": 607.265625,
      "completions/min_length": 203.0,
      "completions/min_terminated_length": 203.0,
      "epoch": 0.192,
      "grad_norm": 0.007254335097968578,
      "learning_rate": 5.555555555555555e-07,
      "loss": 0.0433,
      "num_tokens": 38059012.0,
      "reward": 1.1357789039611816,
      "reward_std": 0.22384758293628693,
      "rewards/accuracy_reward_step": 0.546875,
      "rewards/final_brier_reward_step": 0.7285890579223633,
      "rewards/format_reward_step": 0.99609375,
      "step": 180
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24254727293737233,
      "aux_distill/mean_u": 0.06503256739429049,
      "aux_distill/n_active_tok": 493.53125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4927691146881288,
      "calib/avg_num_step_conf": 7.78125,
      "calib/ece": 0.17722047244094488,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.031496062992125984,
      "calib/gap": -0.013612927565392341,
      "calib/mean_conf": 0.5388425196850394,
      "calib/mu_c": 0.5312321428571428,
      "calib/mu_w": 0.5448450704225352,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.13755905511811026,
      "calib/std_conf": 0.18144008035438455,
      "calib/step_conf_rate": 1.0,
      "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": 2533.0,
      "completions/max_terminated_length": 2533.0,
      "completions/mean_length": 504.9296875,
      "completions/mean_terminated_length": 506.9098205566406,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 186.0,
      "epoch": 0.19306666666666666,
      "grad_norm": 0.00835175346583128,
      "learning_rate": 5.277777777777779e-07,
      "loss": 0.0239,
      "num_tokens": 38294538.0,
      "reward": 1.0642285346984863,
      "reward_std": 0.19430889189243317,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/final_brier_reward_step": 0.6987694501876831,
      "rewards/format_reward_step": 0.9921875,
      "step": 181
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.23280840320512652,
      "aux_distill/mean_u": 0.07703698157631368,
      "aux_distill/n_active_tok": 552.96875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.541740957717779,
      "calib/avg_num_step_conf": 8.62890625,
      "calib/ece": 0.0938780392156863,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.047058823529411764,
      "calib/gap": 0.024451980387162564,
      "calib/mean_conf": 0.5652592156862745,
      "calib/mu_c": 0.5752317880794703,
      "calib/mu_w": 0.5507798076923077,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03349019607843139,
      "calib/std_conf": 0.17475515730102623,
      "calib/step_conf_rate": 1.0,
      "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": 2817.0,
      "completions/max_terminated_length": 2817.0,
      "completions/mean_length": 543.6015625,
      "completions/mean_terminated_length": 543.6015625,
      "completions/min_length": 233.0,
      "completions/min_terminated_length": 233.0,
      "epoch": 0.19413333333333332,
      "grad_norm": 0.007248417474329472,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.0393,
      "num_tokens": 38539860.0,
      "reward": 1.157365083694458,
      "reward_std": 0.20332470536231995,
      "rewards/accuracy_reward_step": 0.58984375,
      "rewards/final_brier_reward_step": 0.7326991558074951,
      "rewards/format_reward_step": 0.9921875,
      "step": 182
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2438644040375948,
      "aux_distill/mean_u": 0.0821571597139334,
      "aux_distill/n_active_tok": 574.96875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5762962962962963,
      "calib/avg_num_step_conf": 9.0390625,
      "calib/ece": 0.07639215686274511,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.03529411764705882,
      "calib/gap": 0.04816666666666669,
      "calib/mean_conf": 0.5279999999999999,
      "calib/mu_c": 0.5535,
      "calib/mu_w": 0.5053333333333333,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.06690196078431374,
      "calib/std_conf": 0.1661098148461893,
      "calib/step_conf_rate": 1.0,
      "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": 1648.0,
      "completions/max_terminated_length": 1648.0,
      "completions/mean_length": 585.2578125,
      "completions/mean_terminated_length": 587.552978515625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 199.0,
      "epoch": 0.1952,
      "grad_norm": 0.007068790029734373,
      "learning_rate": 4.7222222222222226e-07,
      "loss": 0.0288,
      "num_tokens": 38796366.0,
      "reward": 1.102957010269165,
      "reward_std": 0.2160937488079071,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.7410702705383301,
      "rewards/format_reward_step": 0.99609375,
      "step": 183
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2564887339249253,
      "aux_distill/mean_u": 0.10162648412902427,
      "aux_distill/n_active_tok": 555.75,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5537652531517562,
      "calib/avg_num_step_conf": 8.6875,
      "calib/ece": 0.18566771653543307,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.04330708661417323,
      "calib/gap": 0.025047960628328703,
      "calib/mean_conf": 0.5240960629921261,
      "calib/mu_c": 0.5330699386503067,
      "calib/mu_w": 0.508021978021978,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03401574803149607,
      "calib/std_conf": 0.19249284019881638,
      "calib/step_conf_rate": 1.0,
      "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": 2560.0,
      "completions/max_terminated_length": 2560.0,
      "completions/mean_length": 568.65625,
      "completions/mean_terminated_length": 568.65625,
      "completions/min_length": 214.0,
      "completions/min_terminated_length": 214.0,
      "epoch": 0.19626666666666667,
      "grad_norm": 0.0074840690940618515,
      "learning_rate": 4.444444444444445e-07,
      "loss": 0.0811,
      "num_tokens": 39047222.0,
      "reward": 1.1769556999206543,
      "reward_std": 0.19394519925117493,
      "rewards/accuracy_reward_step": 0.63671875,
      "rewards/final_brier_reward_step": 0.7250052094459534,
      "rewards/format_reward_step": 0.9921875,
      "step": 184
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24720469187013805,
      "aux_distill/mean_u": 0.09512097785666002,
      "aux_distill/n_active_tok": 571.0625,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.4885932780669623,
      "calib/avg_num_step_conf": 8.9609375,
      "calib/ece": 0.15493400000000002,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.064,
      "calib/gap": -0.01469857335646807,
      "calib/mean_conf": 0.536506,
      "calib/mu_c": 0.5296270676691729,
      "calib/mu_w": 0.544325641025641,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.07972,
      "calib/std_conf": 0.20492050518188756,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2768.0,
      "completions/max_terminated_length": 2768.0,
      "completions/mean_length": 580.8984375,
      "completions/mean_terminated_length": 583.176513671875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 163.0,
      "epoch": 0.19733333333333333,
      "grad_norm": 0.008756212890148163,
      "learning_rate": 4.1666666666666667e-07,
      "loss": 0.0359,
      "num_tokens": 39302852.0,
      "reward": 1.0872716903686523,
      "reward_std": 0.23970657587051392,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/final_brier_reward_step": 0.6823559999465942,
      "rewards/format_reward_step": 0.97265625,
      "step": 185
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2736734275240451,
      "aux_distill/mean_u": 0.09042671780625551,
      "aux_distill/n_active_tok": 544.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5295651107158192,
      "calib/avg_num_step_conf": 8.546875,
      "calib/ece": 0.13554687500000007,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.03515625,
      "calib/gap": 0.025905661534686875,
      "calib/mean_conf": 0.5333593750000001,
      "calib/mu_c": 0.5454014598540146,
      "calib/mu_w": 0.5194957983193277,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.06687500000000006,
      "calib/std_conf": 0.18393391027651582,
      "calib/step_conf_rate": 1.0,
      "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": 1582.0,
      "completions/max_terminated_length": 1582.0,
      "completions/mean_length": 558.234375,
      "completions/mean_terminated_length": 560.423583984375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 216.0,
      "epoch": 0.1984,
      "grad_norm": 0.007027934771031141,
      "learning_rate": 3.8888888888888895e-07,
      "loss": 0.0447,
      "num_tokens": 39550800.0,
      "reward": 1.132723093032837,
      "reward_std": 0.186387836933136,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/final_brier_reward_step": 0.7302898168563843,
      "rewards/format_reward_step": 1.0,
      "step": 186
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2511372393928468,
      "aux_distill/mean_u": 0.07868845063013544,
      "aux_distill/n_active_tok": 574.75,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5694851535618708,
      "calib/avg_num_step_conf": 8.96484375,
      "calib/ece": 0.08982310756972113,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.03187250996015936,
      "calib/gap": 0.03919008538294899,
      "calib/mean_conf": 0.5342406374501992,
      "calib/mu_c": 0.5526646616541354,
      "calib/mu_w": 0.5134745762711864,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.047091633466135464,
      "calib/std_conf": 0.1741905540078731,
      "calib/step_conf_rate": 1.0,
      "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": 2665.0,
      "completions/max_terminated_length": 2665.0,
      "completions/mean_length": 596.44140625,
      "completions/mean_terminated_length": 598.7804565429688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 199.0,
      "epoch": 0.19946666666666665,
      "grad_norm": 0.006233526859432459,
      "learning_rate": 3.611111111111111e-07,
      "loss": 0.047,
      "num_tokens": 39805033.0,
      "reward": 1.112801194190979,
      "reward_std": 0.21134722232818604,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/final_brier_reward_step": 0.7256022691726685,
      "rewards/format_reward_step": 0.98046875,
      "step": 187
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2566721090115607,
      "aux_distill/mean_u": 0.09184461434823142,
      "aux_distill/n_active_tok": 534.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5928427799421893,
      "calib/avg_num_step_conf": 8.421875,
      "calib/ece": 0.11929529411764701,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.050980392156862744,
      "calib/gap": 0.057999704662561324,
      "calib/mean_conf": 0.5308223529411765,
      "calib/mu_c": 0.5556143835616438,
      "calib/mu_w": 0.4976146788990825,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.03878431372549018,
      "calib/std_conf": 0.17984997716057607,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1905.0,
      "completions/max_terminated_length": 1905.0,
      "completions/mean_length": 575.453125,
      "completions/mean_terminated_length": 577.7098388671875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 105.0,
      "epoch": 0.20053333333333334,
      "grad_norm": 0.006328468210995197,
      "learning_rate": 3.3333333333333335e-07,
      "loss": 0.0486,
      "num_tokens": 40056421.0,
      "reward": 1.1526392698287964,
      "reward_std": 0.18582162261009216,
      "rewards/accuracy_reward_step": 0.5703125,
      "rewards/final_brier_reward_step": 0.7427785992622375,
      "rewards/format_reward_step": 0.9921875,
      "step": 188
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2664337196620181,
      "aux_distill/mean_u": 0.10317610606850601,
      "aux_distill/n_active_tok": 523.9375,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5631231306081755,
      "calib/avg_num_step_conf": 8.2890625,
      "calib/ece": 0.09929133858267714,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.05511811023622047,
      "calib/gap": 0.03753738783649063,
      "calib/mean_conf": 0.5631496062992126,
      "calib/mu_c": 0.5805882352941177,
      "calib/mu_w": 0.5430508474576271,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.063503937007874,
      "calib/std_conf": 0.17280775713214339,
      "calib/step_conf_rate": 1.0,
      "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": 1308.0,
      "completions/max_terminated_length": 1308.0,
      "completions/mean_length": 524.1328125,
      "completions/mean_terminated_length": 528.2598266601562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 183.0,
      "epoch": 0.2016,
      "grad_norm": 0.007905436679720879,
      "learning_rate": 3.055555555555556e-07,
      "loss": 0.0023,
      "num_tokens": 40298367.0,
      "reward": 1.1284804344177246,
      "reward_std": 0.1881653368473053,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/final_brier_reward_step": 0.733523428440094,
      "rewards/format_reward_step": 0.9921875,
      "step": 189
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24586508085485548,
      "aux_distill/mean_u": 0.08787149806346584,
      "aux_distill/n_active_tok": 583.28125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5551407782347358,
      "calib/avg_num_step_conf": 9.10546875,
      "calib/ece": 0.13894999999999996,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.03543307086614173,
      "calib/gap": 0.02305158494147419,
      "calib/mean_conf": 0.5204988188976377,
      "calib/mu_c": 0.5303910344827586,
      "calib/mu_w": 0.5073394495412844,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.044291338582677184,
      "calib/std_conf": 0.19267354330562245,
      "calib/step_conf_rate": 1.0,
      "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": 2638.0,
      "completions/max_terminated_length": 2638.0,
      "completions/mean_length": 602.95703125,
      "completions/mean_terminated_length": 602.95703125,
      "completions/min_length": 206.0,
      "completions/min_terminated_length": 206.0,
      "epoch": 0.20266666666666666,
      "grad_norm": 0.008090218529105186,
      "learning_rate": 2.7777777777777776e-07,
      "loss": 0.0204,
      "num_tokens": 40558332.0,
      "reward": 1.1397864818572998,
      "reward_std": 0.20814228057861328,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/final_brier_reward_step": 0.7209792733192444,
      "rewards/format_reward_step": 0.9921875,
      "step": 190
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.21869845094624907,
      "aux_distill/mean_u": 0.06494175114544838,
      "aux_distill/n_active_tok": 559.21875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5373224251648909,
      "calib/avg_num_step_conf": 8.71875,
      "calib/ece": 0.16432480314960635,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.047244094488188976,
      "calib/gap": 0.011381253170979266,
      "calib/mean_conf": 0.5445728346456692,
      "calib/mu_c": 0.5511148148148148,
      "calib/mu_w": 0.5397335616438356,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.14185039370078745,
      "calib/std_conf": 0.18207366362671173,
      "calib/step_conf_rate": 1.0,
      "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": 2965.0,
      "completions/max_terminated_length": 2965.0,
      "completions/mean_length": 541.48046875,
      "completions/mean_terminated_length": 541.48046875,
      "completions/min_length": 170.0,
      "completions/min_terminated_length": 170.0,
      "epoch": 0.20373333333333332,
      "grad_norm": 0.007324114441871643,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.0204,
      "num_tokens": 40801119.0,
      "reward": 1.0630748271942139,
      "reward_std": 0.20367825031280518,
      "rewards/accuracy_reward_step": 0.42578125,
      "rewards/final_brier_reward_step": 0.7081811428070068,
      "rewards/format_reward_step": 0.9921875,
      "step": 191
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2494658336509019,
      "aux_distill/mean_u": 0.08475135246744968,
      "aux_distill/n_active_tok": 539.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5712225705329153,
      "calib/avg_num_step_conf": 8.48046875,
      "calib/ece": 0.12419607843137251,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.054901960784313725,
      "calib/gap": 0.052937304075235114,
      "calib/mean_conf": 0.5196470588235294,
      "calib/mu_c": 0.5424827586206896,
      "calib/mu_w": 0.4895454545454545,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.03760784313725488,
      "calib/std_conf": 0.19056436548659947,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 568.83203125,
      "completions/mean_terminated_length": 571.0628051757812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 161.0,
      "epoch": 0.2048,
      "grad_norm": 0.007168214302510023,
      "learning_rate": 2.2222222222222224e-07,
      "loss": 0.0243,
      "num_tokens": 41051716.0,
      "reward": 1.1472058296203613,
      "reward_std": 0.20757043361663818,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/final_brier_reward_step": 0.7358179688453674,
      "rewards/format_reward_step": 0.9921875,
      "step": 192
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24612186197191477,
      "aux_distill/mean_u": 0.06994783112146444,
      "aux_distill/n_active_tok": 584.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6323058939338009,
      "calib/avg_num_step_conf": 9.125,
      "calib/ece": 0.07956862745098042,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0392156862745098,
      "calib/gap": 0.08549095607235152,
      "calib/mean_conf": 0.5361960784313725,
      "calib/mu_c": 0.5794444444444445,
      "calib/mu_w": 0.493953488372093,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.06082352941176473,
      "calib/std_conf": 0.1779558239528534,
      "calib/step_conf_rate": 0.99609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2645.0,
      "completions/max_terminated_length": 2645.0,
      "completions/mean_length": 565.8125,
      "completions/mean_terminated_length": 565.8125,
      "completions/min_length": 230.0,
      "completions/min_terminated_length": 230.0,
      "epoch": 0.20586666666666667,
      "grad_norm": 0.006805724464356899,
      "learning_rate": 1.9444444444444447e-07,
      "loss": 0.0322,
      "num_tokens": 41302276.0,
      "reward": 1.118457317352295,
      "reward_std": 0.22730928659439087,
      "rewards/accuracy_reward_step": 0.4921875,
      "rewards/final_brier_reward_step": 0.7525394558906555,
      "rewards/format_reward_step": 0.9921875,
      "step": 193
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2551674894057214,
      "aux_distill/mean_u": 0.09176961205337196,
      "aux_distill/n_active_tok": 491.65625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5459585459585459,
      "calib/avg_num_step_conf": 7.7578125,
      "calib/ece": 0.1108271653543307,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.03937007874015748,
      "calib/gap": 0.02570521640521639,
      "calib/mean_conf": 0.5436610236220473,
      "calib/mu_c": 0.5548944055944055,
      "calib/mu_w": 0.5291891891891891,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.04574803149606298,
      "calib/std_conf": 0.19019859310579726,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2413.0,
      "completions/max_terminated_length": 2413.0,
      "completions/mean_length": 510.03125,
      "completions/mean_terminated_length": 512.0314331054688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 200.0,
      "epoch": 0.20693333333333333,
      "grad_norm": 0.00830480270087719,
      "learning_rate": 1.6666666666666668e-07,
      "loss": 0.0502,
      "num_tokens": 41538788.0,
      "reward": 1.1304701566696167,
      "reward_std": 0.2282831370830536,
      "rewards/accuracy_reward_step": 0.55859375,
      "rewards/final_brier_reward_step": 0.7179715633392334,
      "rewards/format_reward_step": 0.984375,
      "step": 194
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.26188449282199144,
      "aux_distill/mean_u": 0.09851403474716852,
      "aux_distill/n_active_tok": 497.90625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5571828358208955,
      "calib/avg_num_step_conf": 7.89453125,
      "calib/ece": 0.12019566929133857,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.051181102362204724,
      "calib/gap": 0.04633953980099503,
      "calib/mean_conf": 0.5530326771653543,
      "calib/mu_c": 0.5749253731343283,
      "calib/mu_w": 0.5285858333333333,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.07283464566929132,
      "calib/std_conf": 0.17800778558107633,
      "calib/step_conf_rate": 1.0,
      "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": 1612.0,
      "completions/max_terminated_length": 1612.0,
      "completions/mean_length": 502.91796875,
      "completions/mean_terminated_length": 504.8902282714844,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 208.0,
      "epoch": 0.208,
      "grad_norm": 0.007816116325557232,
      "learning_rate": 1.3888888888888888e-07,
      "loss": 0.0065,
      "num_tokens": 41773519.0,
      "reward": 1.1256775856018066,
      "reward_std": 0.17670682072639465,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/final_brier_reward_step": 0.7357300519943237,
      "rewards/format_reward_step": 0.9921875,
      "step": 195
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.25544829899445176,
      "aux_distill/mean_u": 0.0779318698241067,
      "aux_distill/n_active_tok": 490.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5593128505864355,
      "calib/avg_num_step_conf": 7.67578125,
      "calib/ece": 0.09490984251968505,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.031496062992125984,
      "calib/gap": 0.028060135135135167,
      "calib/mean_conf": 0.55635,
      "calib/mu_c": 0.5680601351351352,
      "calib/mu_w": 0.54,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03429133858267718,
      "calib/std_conf": 0.16964918057127165,
      "calib/step_conf_rate": 1.0,
      "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": 932.0,
      "completions/max_terminated_length": 932.0,
      "completions/mean_length": 454.61328125,
      "completions/mean_terminated_length": 456.3961181640625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 190.0,
      "epoch": 0.20906666666666668,
      "grad_norm": 0.008982392027974129,
      "learning_rate": 1.1111111111111112e-07,
      "loss": 0.0266,
      "num_tokens": 41992444.0,
      "reward": 1.1527657508850098,
      "reward_std": 0.17497286200523376,
      "rewards/accuracy_reward_step": 0.578125,
      "rewards/final_brier_reward_step": 0.73521888256073,
      "rewards/format_reward_step": 0.9921875,
      "step": 196
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.30042441678233445,
      "aux_distill/mean_u": 0.12130369822225158,
      "aux_distill/n_active_tok": 494.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.6012697049622768,
      "calib/avg_num_step_conf": 7.76171875,
      "calib/ece": 0.07515624999999998,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.05859375,
      "calib/gap": 0.06255903821382569,
      "calib/mean_conf": 0.572890625,
      "calib/mu_c": 0.6019708029197081,
      "calib/mu_w": 0.5394117647058824,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.05644531249999998,
      "calib/std_conf": 0.1681013103372766,
      "calib/step_conf_rate": 1.0,
      "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": 1399.0,
      "completions/max_terminated_length": 1399.0,
      "completions/mean_length": 511.00390625,
      "completions/mean_terminated_length": 513.0078735351562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 233.0,
      "epoch": 0.21013333333333334,
      "grad_norm": 0.007946827448904514,
      "learning_rate": 8.333333333333334e-08,
      "loss": 0.0417,
      "num_tokens": 42228317.0,
      "reward": 1.1439175605773926,
      "reward_std": 0.18342596292495728,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/final_brier_reward_step": 0.7526788711547852,
      "rewards/format_reward_step": 1.0,
      "step": 197
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24310559080913663,
      "aux_distill/mean_u": 0.06648958651023255,
      "aux_distill/n_active_tok": 526.03125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.486159169550173,
      "calib/avg_num_step_conf": 8.21875,
      "calib/ece": 0.1758301960784314,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0392156862745098,
      "calib/gap": -0.008017647058823574,
      "calib/mean_conf": 0.5491109803921569,
      "calib/mu_c": 0.5459039215686274,
      "calib/mu_w": 0.553921568627451,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.06247058823529414,
      "calib/std_conf": 0.1798272084503311,
      "calib/step_conf_rate": 1.0,
      "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": 2580.0,
      "completions/max_terminated_length": 2580.0,
      "completions/mean_length": 543.18359375,
      "completions/mean_terminated_length": 543.18359375,
      "completions/min_length": 176.0,
      "completions/min_terminated_length": 176.0,
      "epoch": 0.2112,
      "grad_norm": 0.007474220357835293,
      "learning_rate": 5.555555555555556e-08,
      "loss": 0.0647,
      "num_tokens": 42472756.0,
      "reward": 1.1560783386230469,
      "reward_std": 0.19086578488349915,
      "rewards/accuracy_reward_step": 0.59765625,
      "rewards/final_brier_reward_step": 0.7184067368507385,
      "rewards/format_reward_step": 0.99609375,
      "step": 198
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2779740021796897,
      "aux_distill/mean_u": 0.11147548417562779,
      "aux_distill/n_active_tok": 554.5,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5123243933588761,
      "calib/avg_num_step_conf": 8.8203125,
      "calib/ece": 0.13972766798418973,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.05138339920948617,
      "calib/gap": 0.014945051085568206,
      "calib/mean_conf": 0.552999604743083,
      "calib/mu_c": 0.5593793103448275,
      "calib/mu_w": 0.5444342592592593,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.059802371541502,
      "calib/std_conf": 0.16990868054094213,
      "calib/step_conf_rate": 1.0,
      "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": 2556.0,
      "completions/max_terminated_length": 2556.0,
      "completions/mean_length": 573.48828125,
      "completions/mean_terminated_length": 575.7373046875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 201.0,
      "epoch": 0.21226666666666666,
      "grad_norm": 0.006655022501945496,
      "learning_rate": 2.777777777777778e-08,
      "loss": 0.0542,
      "num_tokens": 42723769.0,
      "reward": 1.1397393941879272,
      "reward_std": 0.22068241238594055,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/final_brier_reward_step": 0.7247913479804993,
      "rewards/format_reward_step": 0.98828125,
      "step": 199
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.24724427005276084,
      "aux_distill/mean_u": 0.07388270953856758,
      "aux_distill/n_active_tok": 515.75,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.561402942694889,
      "calib/avg_num_step_conf": 8.0625,
      "calib/ece": 0.10376403162055337,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.023715415019762844,
      "calib/gap": 0.033141281621063334,
      "calib/mean_conf": 0.5397324110671937,
      "calib/mu_c": 0.5533557046979865,
      "calib/mu_w": 0.5202144230769231,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.027281818181818195,
      "calib/std_conf": 0.16138108452849692,
      "calib/step_conf_rate": 1.0,
      "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": 2466.0,
      "completions/max_terminated_length": 2466.0,
      "completions/mean_length": 568.984375,
      "completions/mean_terminated_length": 568.984375,
      "completions/min_length": 163.0,
      "completions/min_terminated_length": 163.0,
      "epoch": 0.21333333333333335,
      "grad_norm": 0.007749881129711866,
      "learning_rate": 0.0,
      "loss": 0.0315,
      "num_tokens": 42977477.0,
      "reward": 1.1535335779190063,
      "reward_std": 0.178613543510437,
      "rewards/accuracy_reward_step": 0.58203125,
      "rewards/final_brier_reward_step": 0.7367547154426575,
      "rewards/format_reward_step": 0.98828125,
      "step": 200
    },
    {
      "epoch": 0.21333333333333335,
      "step": 200,
      "total_flos": 0.0,
      "train_loss": 0.050044363581109794,
      "train_runtime": 17251.8426,
      "train_samples_per_second": 2.968,
      "train_steps_per_second": 0.012
    }
  ],
  "logging_steps": 1,
  "max_steps": 200,
  "num_input_tokens_seen": 42977477,
  "num_train_epochs": 1,
  "save_steps": 20,
  "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
}