{
  "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": 0.9022029798764449,
      "aux_distill/mean_u": 0.3108363672915008,
      "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": 0.012639827094972134,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.0681,
      "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": 0.5604067397745032,
      "aux_distill/mean_u": 0.29356464889172007,
      "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": 0.008656824938952923,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.0987,
      "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": 0.10000000000000003,
      "aux_distill/loss": 0.7020920283264585,
      "aux_distill/mean_u": 0.2341661230714126,
      "aux_distill/n_active_tok": 45.333333333333336,
      "calib/answer_extract_rate": 0.07421875,
      "calib/auroc": 0.3384615384615385,
      "calib/avg_num_step_conf": 0.3984375,
      "calib/ece": 0.6294444444444445,
      "calib/final_conf_rate": 0.0703125,
      "calib/format_rate": 0.06640625,
      "calib/frac_conf_gt_0.9": 0.7777777777777778,
      "calib/gap": 0.0038461538461534994,
      "calib/mean_conf": 0.9072222222222222,
      "calib/mu_c": 0.9099999999999999,
      "calib/mu_w": 0.9061538461538464,
      "calib/nonempty_final_conf_rate": 0.0703125,
      "calib/nonempty_reasoning_rate": 0.1015625,
      "calib/nonempty_step_conf_rate": 0.09375,
      "calib/pce": 0.6294444444444445,
      "calib/std_conf": 0.12164637070695056,
      "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": 0.0703125,
      "completions/max_length": 3043.0,
      "completions/max_terminated_length": 3043.0,
      "completions/mean_length": 751.33203125,
      "completions/mean_terminated_length": 808.155517578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0032,
      "grad_norm": 0.01135075930505991,
      "learning_rate": 7.5e-07,
      "loss": 0.064,
      "num_tokens": 831064.0,
      "reward": 0.05656171590089798,
      "reward_std": 0.13380812108516693,
      "rewards/accuracy_reward_step": 0.01953125,
      "rewards/final_brier_reward_step": 0.027185939252376556,
      "rewards/format_reward_step": 0.06640625,
      "step": 3
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.6724928220113119,
      "aux_distill/mean_u": 0.2293809413630709,
      "aux_distill/n_active_tok": 56.888888888888886,
      "calib/answer_extract_rate": 0.05859375,
      "calib/auroc": 0.3888888888888889,
      "calib/avg_num_step_conf": 0.25390625,
      "calib/ece": 0.7718181818181816,
      "calib/final_conf_rate": 0.04296875,
      "calib/format_rate": 0.03515625,
      "calib/frac_conf_gt_0.9": 0.8181818181818182,
      "calib/gap": 0.007777777777777661,
      "calib/mean_conf": 0.9536363636363636,
      "calib/mu_c": 0.96,
      "calib/mu_w": 0.9522222222222223,
      "calib/nonempty_final_conf_rate": 0.04296875,
      "calib/nonempty_reasoning_rate": 0.0625,
      "calib/nonempty_step_conf_rate": 0.046875,
      "calib/pce": 0.7718181818181816,
      "calib/std_conf": 0.040959564822068396,
      "calib/step_conf_rate": 0.046875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1015625,
      "completions/max_length": 2996.0,
      "completions/max_terminated_length": 2996.0,
      "completions/mean_length": 606.234375,
      "completions/mean_terminated_length": 674.7651977539062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.004266666666666667,
      "grad_norm": 0.0061745117418468,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0274,
      "num_tokens": 1092428.0,
      "reward": 0.026426563039422035,
      "reward_std": 0.07038375735282898,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/final_brier_reward_step": 0.009884374216198921,
      "rewards/format_reward_step": 0.03515625,
      "step": 4
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.7171915471553802,
      "aux_distill/mean_u": 0.2454572816006418,
      "aux_distill/n_active_tok": 43.42857142857143,
      "calib/answer_extract_rate": 0.0546875,
      "calib/auroc": 0.4285714285714286,
      "calib/avg_num_step_conf": 0.1484375,
      "calib/ece": 0.651375,
      "calib/final_conf_rate": 0.03125,
      "calib/format_rate": 0.02734375,
      "calib/frac_conf_gt_0.9": 0.375,
      "calib/gap": 0.07271428571428562,
      "calib/mean_conf": 0.776375,
      "calib/mu_c": 0.84,
      "calib/mu_w": 0.7672857142857143,
      "calib/nonempty_final_conf_rate": 0.03125,
      "calib/nonempty_reasoning_rate": 0.06640625,
      "calib/nonempty_step_conf_rate": 0.0390625,
      "calib/pce": 0.651375,
      "calib/std_conf": 0.29716491107632476,
      "calib/step_conf_rate": 0.0390625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2979.0,
      "completions/max_terminated_length": 2979.0,
      "completions/mean_length": 718.921875,
      "completions/mean_terminated_length": 789.888427734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.005333333333333333,
      "grad_norm": 0.005417319014668465,
      "learning_rate": 1.25e-06,
      "loss": 0.0285,
      "num_tokens": 1383160.0,
      "reward": 0.021108552813529968,
      "reward_std": 0.05447760224342346,
      "rewards/accuracy_reward_step": 0.00390625,
      "rewards/final_brier_reward_step": 0.010967106558382511,
      "rewards/format_reward_step": 0.02734375,
      "step": 5
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.7201201096177101,
      "aux_distill/mean_u": 0.33045206176182806,
      "aux_distill/n_active_tok": 54.416666666666664,
      "calib/answer_extract_rate": 0.08984375,
      "calib/auroc": 0.45833333333333337,
      "calib/avg_num_step_conf": 0.3046875,
      "calib/ece": 0.8054285714285715,
      "calib/final_conf_rate": 0.0546875,
      "calib/format_rate": 0.05078125,
      "calib/frac_conf_gt_0.9": 0.9285714285714286,
      "calib/gap": 0.013666666666666494,
      "calib/mean_conf": 0.9482857142857143,
      "calib/mu_c": 0.96,
      "calib/mu_w": 0.9463333333333335,
      "calib/nonempty_final_conf_rate": 0.0546875,
      "calib/nonempty_reasoning_rate": 0.09765625,
      "calib/nonempty_step_conf_rate": 0.0625,
      "calib/pce": 0.8054285714285715,
      "calib/std_conf": 0.06116421522826722,
      "calib/step_conf_rate": 0.0625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2988.0,
      "completions/max_terminated_length": 2988.0,
      "completions/mean_length": 630.50390625,
      "completions/mean_terminated_length": 672.5375366210938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0064,
      "grad_norm": 0.006735073868185282,
      "learning_rate": 1.5e-06,
      "loss": 0.0839,
      "num_tokens": 1650521.0,
      "reward": 0.0346597358584404,
      "reward_std": 0.0735585018992424,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/final_brier_reward_step": 0.010725718922913074,
      "rewards/format_reward_step": 0.05078125,
      "step": 6
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.6654665925923515,
      "aux_distill/mean_u": 0.3108273132858048,
      "aux_distill/n_active_tok": 48.05882352941177,
      "calib/answer_extract_rate": 0.08984375,
      "calib/auroc": 0.4181818181818182,
      "calib/avg_num_step_conf": 0.40234375,
      "calib/ece": 0.6393749999999999,
      "calib/final_conf_rate": 0.0625,
      "calib/format_rate": 0.05078125,
      "calib/frac_conf_gt_0.9": 0.9375,
      "calib/gap": 0.008909090909090867,
      "calib/mean_conf": 0.951875,
      "calib/mu_c": 0.958,
      "calib/mu_w": 0.9490909090909091,
      "calib/nonempty_final_conf_rate": 0.0625,
      "calib/nonempty_reasoning_rate": 0.1171875,
      "calib/nonempty_step_conf_rate": 0.08203125,
      "calib/pce": 0.6393749999999999,
      "calib/std_conf": 0.042014692370645756,
      "calib/step_conf_rate": 0.08203125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0859375,
      "completions/max_length": 3053.0,
      "completions/max_terminated_length": 3053.0,
      "completions/mean_length": 708.05078125,
      "completions/mean_terminated_length": 774.6196899414062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.007466666666666667,
      "grad_norm": 0.009294859133660793,
      "learning_rate": 1.75e-06,
      "loss": 0.0773,
      "num_tokens": 1939206.0,
      "reward": 0.044801369309425354,
      "reward_std": 0.10697519779205322,
      "rewards/accuracy_reward_step": 0.01953125,
      "rewards/final_brier_reward_step": 0.01929023489356041,
      "rewards/format_reward_step": 0.05078125,
      "step": 7
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.7723357169067159,
      "aux_distill/mean_u": 0.3379113503404071,
      "aux_distill/n_active_tok": 44.11764705882353,
      "calib/answer_extract_rate": 0.08203125,
      "calib/auroc": 0.0,
      "calib/avg_num_step_conf": 0.2890625,
      "calib/ece": 0.905,
      "calib/final_conf_rate": 0.046875,
      "calib/format_rate": 0.02734375,
      "calib/frac_conf_gt_0.9": 0.8333333333333334,
      "calib/gap": -0.39636363636363636,
      "calib/mean_conf": 0.9133333333333334,
      "calib/mu_c": 0.55,
      "calib/mu_w": 0.9463636363636364,
      "calib/nonempty_final_conf_rate": 0.046875,
      "calib/nonempty_reasoning_rate": 0.1171875,
      "calib/nonempty_step_conf_rate": 0.0703125,
      "calib/pce": 0.8675,
      "calib/std_conf": 0.12638125740085915,
      "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": 0.0625,
      "completions/max_length": 3037.0,
      "completions/max_terminated_length": 3037.0,
      "completions/mean_length": 646.2109375,
      "completions/mean_terminated_length": 689.2916870117188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.008533333333333334,
      "grad_norm": 0.008389068767428398,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.058,
      "num_tokens": 2211148.0,
      "reward": 0.01809394732117653,
      "reward_std": 0.047178227454423904,
      "rewards/accuracy_reward_step": 0.00390625,
      "rewards/final_brier_reward_step": 0.004937890917062759,
      "rewards/format_reward_step": 0.02734375,
      "step": 8
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.6544934511184692,
      "aux_distill/mean_u": 0.21143247752578329,
      "aux_distill/n_active_tok": 58.4,
      "calib/answer_extract_rate": 0.06640625,
      "calib/auroc": 0.8333333333333334,
      "calib/avg_num_step_conf": 0.2890625,
      "calib/ece": 0.5545454545454547,
      "calib/final_conf_rate": 0.04296875,
      "calib/format_rate": 0.03515625,
      "calib/frac_conf_gt_0.9": 0.45454545454545453,
      "calib/gap": 0.28555555555555556,
      "calib/mean_conf": 0.7363636363636363,
      "calib/mu_c": 0.97,
      "calib/mu_w": 0.6844444444444444,
      "calib/nonempty_final_conf_rate": 0.04296875,
      "calib/nonempty_reasoning_rate": 0.08203125,
      "calib/nonempty_step_conf_rate": 0.05859375,
      "calib/pce": 0.5545454545454547,
      "calib/std_conf": 0.32902812057973313,
      "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": 0.12890625,
      "completions/max_length": 3028.0,
      "completions/max_terminated_length": 3028.0,
      "completions/mean_length": 568.79296875,
      "completions/mean_terminated_length": 652.9641723632812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0096,
      "grad_norm": 0.007296205498278141,
      "learning_rate": 2.25e-06,
      "loss": 0.0066,
      "num_tokens": 2464295.0,
      "reward": 0.030635546892881393,
      "reward_std": 0.06952265650033951,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/final_brier_reward_step": 0.018302343785762787,
      "rewards/format_reward_step": 0.03515625,
      "step": 9
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.657138936659869,
      "aux_distill/mean_u": 0.2260673486435751,
      "aux_distill/n_active_tok": 42.35294117647059,
      "calib/answer_extract_rate": 0.10546875,
      "calib/auroc": 0.9444444444444444,
      "calib/avg_num_step_conf": 0.36328125,
      "calib/ece": 0.7887199999999999,
      "calib/final_conf_rate": 0.078125,
      "calib/format_rate": 0.0625,
      "calib/frac_conf_gt_0.9": 0.8,
      "calib/gap": 0.12364444444444445,
      "calib/mean_conf": 0.88872,
      "calib/mu_c": 1.0,
      "calib/mu_w": 0.8763555555555556,
      "calib/nonempty_final_conf_rate": 0.078125,
      "calib/nonempty_reasoning_rate": 0.1171875,
      "calib/nonempty_step_conf_rate": 0.0859375,
      "calib/pce": 0.7887199999999999,
      "calib/std_conf": 0.21795175980019063,
      "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": 0.125,
      "completions/max_length": 2933.0,
      "completions/max_terminated_length": 2933.0,
      "completions/mean_length": 597.015625,
      "completions/mean_terminated_length": 682.3035888671875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.010666666666666666,
      "grad_norm": 0.009703119285404682,
      "learning_rate": 2.5e-06,
      "loss": 0.051,
      "num_tokens": 2723931.0,
      "reward": 0.04364839196205139,
      "reward_std": 0.09204517304897308,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/final_brier_reward_step": 0.016984282061457634,
      "rewards/format_reward_step": 0.0625,
      "step": 10
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.6334928455560104,
      "aux_distill/mean_u": 0.2929954140568702,
      "aux_distill/n_active_tok": 67.95652173913044,
      "calib/answer_extract_rate": 0.15234375,
      "calib/auroc": 0.44545454545454544,
      "calib/avg_num_step_conf": 0.7890625,
      "calib/ece": 0.7625925925925927,
      "calib/final_conf_rate": 0.10546875,
      "calib/format_rate": 0.078125,
      "calib/frac_conf_gt_0.9": 0.9259259259259259,
      "calib/gap": 0.0027272727272726893,
      "calib/mean_conf": 0.9477777777777777,
      "calib/mu_c": 0.95,
      "calib/mu_w": 0.9472727272727273,
      "calib/nonempty_final_conf_rate": 0.10546875,
      "calib/nonempty_reasoning_rate": 0.2109375,
      "calib/nonempty_step_conf_rate": 0.15625,
      "calib/pce": 0.7625925925925927,
      "calib/std_conf": 0.07435317788587381,
      "calib/step_conf_rate": 0.15625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2890.0,
      "completions/max_terminated_length": 2890.0,
      "completions/mean_length": 564.7890625,
      "completions/mean_terminated_length": 615.2595825195312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.011733333333333333,
      "grad_norm": 0.010903569869697094,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.0657,
      "num_tokens": 2972997.0,
      "reward": 0.0605287104845047,
      "reward_std": 0.11411545425653458,
      "rewards/accuracy_reward_step": 0.01953125,
      "rewards/final_brier_reward_step": 0.02340117283165455,
      "rewards/format_reward_step": 0.078125,
      "step": 11
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.6554603688418865,
      "aux_distill/mean_u": 0.3278240132703401,
      "aux_distill/n_active_tok": 63.666666666666664,
      "calib/answer_extract_rate": 0.19921875,
      "calib/auroc": 0.5538461538461538,
      "calib/avg_num_step_conf": 0.74609375,
      "calib/ece": 0.5900000000000001,
      "calib/final_conf_rate": 0.16015625,
      "calib/format_rate": 0.12109375,
      "calib/frac_conf_gt_0.9": 0.6829268292682927,
      "calib/gap": -0.05848717948717963,
      "calib/mean_conf": 0.8797560975609756,
      "calib/mu_c": 0.8426666666666667,
      "calib/mu_w": 0.9011538461538463,
      "calib/nonempty_final_conf_rate": 0.16015625,
      "calib/nonempty_reasoning_rate": 0.23828125,
      "calib/nonempty_step_conf_rate": 0.1796875,
      "calib/pce": 0.5519512195121952,
      "calib/std_conf": 0.17420884305742093,
      "calib/step_conf_rate": 0.1796875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2925.0,
      "completions/max_terminated_length": 2925.0,
      "completions/mean_length": 566.5,
      "completions/mean_terminated_length": 614.5084838867188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.0128,
      "grad_norm": 0.012244653888046741,
      "learning_rate": 3e-06,
      "loss": 0.0742,
      "num_tokens": 3222197.0,
      "reward": 0.12596894800662994,
      "reward_std": 0.22686685621738434,
      "rewards/accuracy_reward_step": 0.06640625,
      "rewards/final_brier_reward_step": 0.06443789601325989,
      "rewards/format_reward_step": 0.12109375,
      "step": 12
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.6795370479424795,
      "aux_distill/mean_u": 0.2659909197732691,
      "aux_distill/n_active_tok": 60.875,
      "calib/answer_extract_rate": 0.1875,
      "calib/auroc": 0.9396551724137931,
      "calib/avg_num_step_conf": 0.7265625,
      "calib/ece": 0.7548387096774192,
      "calib/final_conf_rate": 0.12109375,
      "calib/format_rate": 0.1015625,
      "calib/frac_conf_gt_0.9": 0.5483870967741935,
      "calib/gap": 0.18775862068965532,
      "calib/mean_conf": 0.8193548387096772,
      "calib/mu_c": 0.995,
      "calib/mu_w": 0.8072413793103447,
      "calib/nonempty_final_conf_rate": 0.12109375,
      "calib/nonempty_reasoning_rate": 0.234375,
      "calib/nonempty_step_conf_rate": 0.171875,
      "calib/pce": 0.7548387096774192,
      "calib/std_conf": 0.26929939549146925,
      "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": 0.06640625,
      "completions/max_length": 2820.0,
      "completions/max_terminated_length": 2820.0,
      "completions/mean_length": 665.91796875,
      "completions/mean_terminated_length": 713.2844848632812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.013866666666666666,
      "grad_norm": 0.01278610061854124,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.0775,
      "num_tokens": 3497264.0,
      "reward": 0.07057617604732513,
      "reward_std": 0.13284853100776672,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/final_brier_reward_step": 0.03177734464406967,
      "rewards/format_reward_step": 0.1015625,
      "step": 13
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5881305737742062,
      "aux_distill/mean_u": 0.2588630859384443,
      "aux_distill/n_active_tok": 77.58620689655173,
      "calib/answer_extract_rate": 0.234375,
      "calib/auroc": 0.6453125,
      "calib/avg_num_step_conf": 1.10546875,
      "calib/ece": 0.6635190433333333,
      "calib/final_conf_rate": 0.1875,
      "calib/format_rate": 0.15625,
      "calib/frac_conf_gt_0.9": 0.5833333333333334,
      "calib/gap": 0.05577714799999989,
      "calib/mean_conf": 0.8022690433333333,
      "calib/mu_c": 0.8487499999999999,
      "calib/mu_w": 0.792972852,
      "calib/nonempty_final_conf_rate": 0.1875,
      "calib/nonempty_reasoning_rate": 0.28515625,
      "calib/nonempty_step_conf_rate": 0.22265625,
      "calib/pce": 0.64956071,
      "calib/std_conf": 0.24954569363040827,
      "calib/step_conf_rate": 0.22265625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3034.0,
      "completions/max_terminated_length": 3034.0,
      "completions/mean_length": 621.8984375,
      "completions/mean_terminated_length": 655.168701171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.014933333333333333,
      "grad_norm": 0.013642070814967155,
      "learning_rate": 3.5e-06,
      "loss": 0.1353,
      "num_tokens": 3761870.0,
      "reward": 0.12654060125350952,
      "reward_std": 0.2307237684726715,
      "rewards/accuracy_reward_step": 0.03125,
      "rewards/final_brier_reward_step": 0.06558118760585785,
      "rewards/format_reward_step": 0.15625,
      "step": 14
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5574666220566322,
      "aux_distill/mean_u": 0.33510179894522996,
      "aux_distill/n_active_tok": 96.48275862068965,
      "calib/answer_extract_rate": 0.28515625,
      "calib/auroc": 0.21071428571428574,
      "calib/avg_num_step_conf": 1.35546875,
      "calib/ece": 0.7250874316939889,
      "calib/final_conf_rate": 0.23828125,
      "calib/format_rate": 0.20703125,
      "calib/frac_conf_gt_0.9": 0.5737704918032787,
      "calib/gap": -0.28800595238095245,
      "calib/mean_conf": 0.746398907103825,
      "calib/mu_c": 0.48199999999999993,
      "calib/mu_w": 0.7700059523809524,
      "calib/nonempty_final_conf_rate": 0.23828125,
      "calib/nonempty_reasoning_rate": 0.34765625,
      "calib/nonempty_step_conf_rate": 0.27734375,
      "calib/pce": 0.69475956284153,
      "calib/std_conf": 0.3272540766327567,
      "calib/step_conf_rate": 0.27734375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3035.0,
      "completions/max_terminated_length": 3035.0,
      "completions/mean_length": 550.6953125,
      "completions/mean_terminated_length": 589.8660888671875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 5.0,
      "epoch": 0.016,
      "grad_norm": 0.01197971310466528,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.1051,
      "num_tokens": 4010728.0,
      "reward": 0.1487540900707245,
      "reward_std": 0.23280642926692963,
      "rewards/accuracy_reward_step": 0.0234375,
      "rewards/final_brier_reward_step": 0.06703945249319077,
      "rewards/format_reward_step": 0.20703125,
      "step": 15
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5236153088766953,
      "aux_distill/mean_u": 0.40382603766356184,
      "aux_distill/n_active_tok": 92.96551724137932,
      "calib/answer_extract_rate": 0.25,
      "calib/auroc": 0.4451754385964912,
      "calib/avg_num_step_conf": 1.32421875,
      "calib/ece": 0.5738000000000001,
      "calib/final_conf_rate": 0.1953125,
      "calib/format_rate": 0.16796875,
      "calib/frac_conf_gt_0.9": 0.54,
      "calib/gap": -0.004473684210526407,
      "calib/mean_conf": 0.7534000000000001,
      "calib/mu_c": 0.75,
      "calib/mu_w": 0.7544736842105264,
      "calib/nonempty_final_conf_rate": 0.1953125,
      "calib/nonempty_reasoning_rate": 0.33984375,
      "calib/nonempty_step_conf_rate": 0.28125,
      "calib/pce": 0.5436000000000001,
      "calib/std_conf": 0.3135258203083121,
      "calib/step_conf_rate": 0.28125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3026.0,
      "completions/max_terminated_length": 3026.0,
      "completions/mean_length": 630.98828125,
      "completions/mean_terminated_length": 687.3744506835938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.017066666666666667,
      "grad_norm": 0.01098259724676609,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.1117,
      "num_tokens": 4281109.0,
      "reward": 0.14574824273586273,
      "reward_std": 0.261897474527359,
      "rewards/accuracy_reward_step": 0.046875,
      "rewards/final_brier_reward_step": 0.07665273547172546,
      "rewards/format_reward_step": 0.16796875,
      "step": 16
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5484382337139498,
      "aux_distill/mean_u": 0.31876979990318755,
      "aux_distill/n_active_tok": 132.38709677419354,
      "calib/answer_extract_rate": 0.3671875,
      "calib/auroc": 0.5265654648956357,
      "calib/avg_num_step_conf": 1.99609375,
      "calib/ece": 0.5748101265822785,
      "calib/final_conf_rate": 0.30859375,
      "calib/format_rate": 0.28125,
      "calib/frac_conf_gt_0.9": 0.43037974683544306,
      "calib/gap": 0.01680265654648949,
      "calib/mean_conf": 0.7115189873417722,
      "calib/mu_c": 0.7247058823529412,
      "calib/mu_w": 0.7079032258064517,
      "calib/nonempty_final_conf_rate": 0.30859375,
      "calib/nonempty_reasoning_rate": 0.48828125,
      "calib/nonempty_step_conf_rate": 0.421875,
      "calib/pce": 0.5355696202531646,
      "calib/std_conf": 0.3216456443441397,
      "calib/step_conf_rate": 0.421875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2972.0,
      "completions/max_terminated_length": 2972.0,
      "completions/mean_length": 609.50390625,
      "completions/mean_terminated_length": 629.165283203125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.018133333333333335,
      "grad_norm": 0.011893536895513535,
      "learning_rate": 4.25e-06,
      "loss": 0.2036,
      "num_tokens": 4540670.0,
      "reward": 0.24156543612480164,
      "reward_std": 0.3197880983352661,
      "rewards/accuracy_reward_step": 0.06640625,
      "rewards/final_brier_reward_step": 0.13547460734844208,
      "rewards/format_reward_step": 0.28125,
      "step": 17
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5134281227665562,
      "aux_distill/mean_u": 0.2940568334667696,
      "aux_distill/n_active_tok": 126.51612903225806,
      "calib/answer_extract_rate": 0.375,
      "calib/auroc": 0.7677556818181819,
      "calib/avg_num_step_conf": 1.8828125,
      "calib/ece": 0.5039999999999999,
      "calib/final_conf_rate": 0.29296875,
      "calib/format_rate": 0.24609375,
      "calib/frac_conf_gt_0.9": 0.3466666666666667,
      "calib/gap": 0.2730113636363637,
      "calib/mean_conf": 0.6506666666666666,
      "calib/mu_c": 0.8836363636363637,
      "calib/mu_w": 0.610625,
      "calib/nonempty_final_conf_rate": 0.29296875,
      "calib/nonempty_reasoning_rate": 0.46484375,
      "calib/nonempty_step_conf_rate": 0.37890625,
      "calib/pce": 0.5039999999999999,
      "calib/std_conf": 0.32455234126750987,
      "calib/step_conf_rate": 0.37890625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2599.0,
      "completions/max_terminated_length": 2599.0,
      "completions/mean_length": 535.1015625,
      "completions/mean_terminated_length": 563.7283935546875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.0192,
      "grad_norm": 0.012328559532761574,
      "learning_rate": 4.5e-06,
      "loss": 0.1924,
      "num_tokens": 4788376.0,
      "reward": 0.21568945050239563,
      "reward_std": 0.31785470247268677,
      "rewards/accuracy_reward_step": 0.04296875,
      "rewards/final_brier_reward_step": 0.14231640100479126,
      "rewards/format_reward_step": 0.24609375,
      "step": 18
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5287263132631779,
      "aux_distill/mean_u": 0.39094491986338004,
      "aux_distill/n_active_tok": 264.46875,
      "calib/answer_extract_rate": 0.67578125,
      "calib/auroc": 0.49503816793893135,
      "calib/avg_num_step_conf": 3.9140625,
      "calib/ece": 0.4305167770419426,
      "calib/final_conf_rate": 0.58984375,
      "calib/format_rate": 0.51953125,
      "calib/frac_conf_gt_0.9": 0.17880794701986755,
      "calib/gap": 0.004201272264630962,
      "calib/mean_conf": 0.5153551876379692,
      "calib/mu_c": 0.5189999999999999,
      "calib/mu_w": 0.5147987277353689,
      "calib/nonempty_final_conf_rate": 0.58984375,
      "calib/nonempty_reasoning_rate": 0.82421875,
      "calib/nonempty_step_conf_rate": 0.734375,
      "calib/pce": 0.40671081677704196,
      "calib/std_conf": 0.34326462305766253,
      "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": 0.01953125,
      "completions/max_length": 2774.0,
      "completions/max_terminated_length": 2774.0,
      "completions/mean_length": 455.41015625,
      "completions/mean_terminated_length": 464.4820861816406,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 13.0,
      "epoch": 0.020266666666666665,
      "grad_norm": 0.015135838650166988,
      "learning_rate": 4.75e-06,
      "loss": 0.2171,
      "num_tokens": 5009721.0,
      "reward": 0.4618377685546875,
      "reward_std": 0.4449642300605774,
      "rewards/accuracy_reward_step": 0.08203125,
      "rewards/final_brier_reward_step": 0.3221130669116974,
      "rewards/format_reward_step": 0.51953125,
      "step": 19
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5356345996260643,
      "aux_distill/mean_u": 0.3635441304973222,
      "aux_distill/n_active_tok": 273.34375,
      "calib/answer_extract_rate": 0.734375,
      "calib/auroc": 0.5462904911180774,
      "calib/avg_num_step_conf": 4.15234375,
      "calib/ece": 0.3078365168539326,
      "calib/final_conf_rate": 0.6953125,
      "calib/format_rate": 0.6171875,
      "calib/frac_conf_gt_0.9": 0.1348314606741573,
      "calib/gap": 0.03933297805642627,
      "calib/mean_conf": 0.46705,
      "calib/mu_c": 0.49909090909090903,
      "calib/mu_w": 0.45975793103448276,
      "calib/nonempty_final_conf_rate": 0.6953125,
      "calib/nonempty_reasoning_rate": 0.89453125,
      "calib/nonempty_step_conf_rate": 0.8046875,
      "calib/pce": 0.29474662921348316,
      "calib/std_conf": 0.32703286167044365,
      "calib/step_conf_rate": 0.8046875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2788.0,
      "completions/max_terminated_length": 2788.0,
      "completions/mean_length": 405.89453125,
      "completions/mean_terminated_length": 412.3373107910156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 24.0,
      "epoch": 0.021333333333333333,
      "grad_norm": 0.015294047072529793,
      "learning_rate": 5e-06,
      "loss": 0.1992,
      "num_tokens": 5218502.0,
      "reward": 0.586089015007019,
      "reward_std": 0.45785677433013916,
      "rewards/accuracy_reward_step": 0.12890625,
      "rewards/final_brier_reward_step": 0.4260842204093933,
      "rewards/format_reward_step": 0.6171875,
      "step": 20
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5090240705758333,
      "aux_distill/mean_u": 0.3329150960047672,
      "aux_distill/n_active_tok": 251.65625,
      "calib/answer_extract_rate": 0.78125,
      "calib/auroc": 0.44311613242249087,
      "calib/avg_num_step_conf": 3.89453125,
      "calib/ece": 0.3569728205128205,
      "calib/final_conf_rate": 0.76171875,
      "calib/format_rate": 0.66796875,
      "calib/frac_conf_gt_0.9": 0.14871794871794872,
      "calib/gap": -0.06790578034682071,
      "calib/mean_conf": 0.4042446153846154,
      "calib/mu_c": 0.3440000000000001,
      "calib/mu_w": 0.4119057803468208,
      "calib/nonempty_final_conf_rate": 0.76171875,
      "calib/nonempty_reasoning_rate": 0.9140625,
      "calib/nonempty_step_conf_rate": 0.859375,
      "calib/pce": 0.3241984615384615,
      "calib/std_conf": 0.3538810433323326,
      "calib/step_conf_rate": 0.859375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2406.0,
      "completions/max_terminated_length": 2406.0,
      "completions/mean_length": 329.9296875,
      "completions/mean_terminated_length": 333.8419189453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 12.0,
      "epoch": 0.0224,
      "grad_norm": 0.016006583347916603,
      "learning_rate": 4.9722222222222224e-06,
      "loss": 0.172,
      "num_tokens": 5405924.0,
      "reward": 0.6004676222801208,
      "reward_std": 0.45191651582717896,
      "rewards/accuracy_reward_step": 0.0859375,
      "rewards/final_brier_reward_step": 0.4470290243625641,
      "rewards/format_reward_step": 0.66796875,
      "step": 21
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5112969875335693,
      "aux_distill/mean_u": 0.3821755846262978,
      "aux_distill/n_active_tok": 295.46875,
      "calib/answer_extract_rate": 0.8515625,
      "calib/auroc": 0.619299571680524,
      "calib/avg_num_step_conf": 4.54296875,
      "calib/ece": 0.29235460317460316,
      "calib/final_conf_rate": 0.8203125,
      "calib/format_rate": 0.765625,
      "calib/frac_conf_gt_0.9": 0.07142857142857142,
      "calib/gap": 0.11863703703703693,
      "calib/mean_conf": 0.3673536507936508,
      "calib/mu_c": 0.47412698412698406,
      "calib/mu_w": 0.35548994708994713,
      "calib/nonempty_final_conf_rate": 0.8203125,
      "calib/nonempty_reasoning_rate": 0.9765625,
      "calib/nonempty_step_conf_rate": 0.94140625,
      "calib/pce": 0.279854126984127,
      "calib/std_conf": 0.31691713012233746,
      "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": 0.0,
      "completions/max_length": 2508.0,
      "completions/max_terminated_length": 2508.0,
      "completions/mean_length": 329.28125,
      "completions/mean_terminated_length": 329.28125,
      "completions/min_length": 29.0,
      "completions/min_terminated_length": 29.0,
      "epoch": 0.023466666666666667,
      "grad_norm": 0.013608487322926521,
      "learning_rate": 4.944444444444445e-06,
      "loss": 0.1555,
      "num_tokens": 5592036.0,
      "reward": 0.7152169346809387,
      "reward_std": 0.41016140580177307,
      "rewards/accuracy_reward_step": 0.08984375,
      "rewards/final_brier_reward_step": 0.5749650001525879,
      "rewards/format_reward_step": 0.765625,
      "step": 22
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5156875886023045,
      "aux_distill/mean_u": 0.31364435556478476,
      "aux_distill/n_active_tok": 292.28125,
      "calib/answer_extract_rate": 0.86328125,
      "calib/auroc": 0.5870047814207651,
      "calib/avg_num_step_conf": 4.46484375,
      "calib/ece": 0.2904984186046512,
      "calib/final_conf_rate": 0.83984375,
      "calib/format_rate": 0.7890625,
      "calib/frac_conf_gt_0.9": 0.13953488372093023,
      "calib/gap": 0.10023129781420764,
      "calib/mean_conf": 0.38249934883720926,
      "calib/mu_c": 0.46781249999999996,
      "calib/mu_w": 0.3675812021857923,
      "calib/nonempty_final_conf_rate": 0.83984375,
      "calib/nonempty_reasoning_rate": 0.9765625,
      "calib/nonempty_step_conf_rate": 0.9296875,
      "calib/pce": 0.26208027906976744,
      "calib/std_conf": 0.33951695417822614,
      "calib/step_conf_rate": 0.9296875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015625,
      "completions/max_length": 2779.0,
      "completions/max_terminated_length": 2779.0,
      "completions/mean_length": 342.59375,
      "completions/mean_terminated_length": 348.0317687988281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 21.0,
      "epoch": 0.024533333333333334,
      "grad_norm": 0.012036532163619995,
      "learning_rate": 4.9166666666666665e-06,
      "loss": 0.2276,
      "num_tokens": 5783676.0,
      "reward": 0.7419977188110352,
      "reward_std": 0.4112620949745178,
      "rewards/accuracy_reward_step": 0.125,
      "rewards/final_brier_reward_step": 0.5699329376220703,
      "rewards/format_reward_step": 0.7890625,
      "step": 23
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5309049636125565,
      "aux_distill/mean_u": 0.38915524999249224,
      "aux_distill/n_active_tok": 273.8125,
      "calib/answer_extract_rate": 0.875,
      "calib/auroc": 0.560290404040404,
      "calib/avg_num_step_conf": 4.14453125,
      "calib/ece": 0.25765810810810813,
      "calib/final_conf_rate": 0.8671875,
      "calib/format_rate": 0.78125,
      "calib/frac_conf_gt_0.9": 0.06306306306306306,
      "calib/gap": 0.04373686868686877,
      "calib/mean_conf": 0.33765810810810815,
      "calib/mu_c": 0.3766666666666667,
      "calib/mu_w": 0.33292979797979794,
      "calib/nonempty_final_conf_rate": 0.8671875,
      "calib/nonempty_reasoning_rate": 0.9765625,
      "calib/nonempty_step_conf_rate": 0.91796875,
      "calib/pce": 0.24360405405405405,
      "calib/std_conf": 0.3121383562224444,
      "calib/step_conf_rate": 0.91796875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2976.0,
      "completions/max_terminated_length": 2976.0,
      "completions/mean_length": 271.21875,
      "completions/mean_terminated_length": 271.21875,
      "completions/min_length": 13.0,
      "completions/min_terminated_length": 13.0,
      "epoch": 0.0256,
      "grad_norm": 0.011990250088274479,
      "learning_rate": 4.888888888888889e-06,
      "loss": 0.1158,
      "num_tokens": 5957620.0,
      "reward": 0.7375925183296204,
      "reward_std": 0.40519455075263977,
      "rewards/accuracy_reward_step": 0.09375,
      "rewards/final_brier_reward_step": 0.6001850366592407,
      "rewards/format_reward_step": 0.78125,
      "step": 24
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.49400834552943707,
      "aux_distill/mean_u": 0.33734021095642364,
      "aux_distill/n_active_tok": 262.9375,
      "calib/answer_extract_rate": 0.92578125,
      "calib/auroc": 0.6188197767145135,
      "calib/avg_num_step_conf": 4.078125,
      "calib/ece": 0.2676514644351464,
      "calib/final_conf_rate": 0.93359375,
      "calib/format_rate": 0.87109375,
      "calib/frac_conf_gt_0.9": 0.09205020920502092,
      "calib/gap": 0.15417878787878786,
      "calib/mean_conf": 0.3525075313807531,
      "calib/mu_c": 0.48733333333333334,
      "calib/mu_w": 0.3331545454545455,
      "calib/nonempty_final_conf_rate": 0.93359375,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.95703125,
      "calib/pce": 0.24731799163179916,
      "calib/std_conf": 0.3220192977821595,
      "calib/step_conf_rate": 0.95703125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1409.0,
      "completions/max_terminated_length": 1409.0,
      "completions/mean_length": 241.72265625,
      "completions/mean_terminated_length": 242.6706085205078,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 24.0,
      "epoch": 0.02666666666666667,
      "grad_norm": 0.01171167753636837,
      "learning_rate": 4.861111111111111e-06,
      "loss": 0.024,
      "num_tokens": 6122725.0,
      "reward": 0.8318536281585693,
      "reward_std": 0.3171781897544861,
      "rewards/accuracy_reward_step": 0.12109375,
      "rewards/final_brier_reward_step": 0.6715196967124939,
      "rewards/format_reward_step": 0.87109375,
      "step": 25
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5315526509657502,
      "aux_distill/mean_u": 0.4173970698103785,
      "aux_distill/n_active_tok": 267.8125,
      "calib/answer_extract_rate": 0.9296875,
      "calib/auroc": 0.4739819004524887,
      "calib/avg_num_step_conf": 4.06640625,
      "calib/ece": 0.2542253112033195,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.89453125,
      "calib/frac_conf_gt_0.9": 0.04564315352697095,
      "calib/gap": -0.04812986425339366,
      "calib/mean_conf": 0.29713568464730294,
      "calib/mu_c": 0.253,
      "calib/mu_w": 0.30112986425339366,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.23418672199170124,
      "calib/std_conf": 0.2775307865477138,
      "calib/step_conf_rate": 0.98046875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2361.0,
      "completions/max_terminated_length": 2361.0,
      "completions/mean_length": 244.53125,
      "completions/mean_terminated_length": 244.53125,
      "completions/min_length": 25.0,
      "completions/min_terminated_length": 25.0,
      "epoch": 0.027733333333333332,
      "grad_norm": 0.012768019922077656,
      "learning_rate": 4.833333333333333e-06,
      "loss": 0.171,
      "num_tokens": 6290565.0,
      "reward": 0.8395206332206726,
      "reward_std": 0.28634095191955566,
      "rewards/accuracy_reward_step": 0.078125,
      "rewards/final_brier_reward_step": 0.70638507604599,
      "rewards/format_reward_step": 0.89453125,
      "step": 26
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5077687716111541,
      "aux_distill/mean_u": 0.36229705136659734,
      "aux_distill/n_active_tok": 250.0625,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.48710407239819004,
      "calib/avg_num_step_conf": 3.8828125,
      "calib/ece": 0.1784360995850622,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.93359375,
      "calib/frac_conf_gt_0.9": 0.029045643153526972,
      "calib/gap": -0.0015764705882352625,
      "calib/mean_conf": 0.22994564315352697,
      "calib/mu_c": 0.2285,
      "calib/mu_w": 0.23007647058823527,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.16269709543568464,
      "calib/std_conf": 0.2365592422526031,
      "calib/step_conf_rate": 0.98046875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1575.0,
      "completions/max_terminated_length": 1575.0,
      "completions/mean_length": 214.62109375,
      "completions/mean_terminated_length": 215.46275329589844,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 49.0,
      "epoch": 0.0288,
      "grad_norm": 0.011952613480389118,
      "learning_rate": 4.805555555555556e-06,
      "loss": 0.0934,
      "num_tokens": 6450724.0,
      "reward": 0.9002429842948914,
      "reward_std": 0.21553806960582733,
      "rewards/accuracy_reward_step": 0.078125,
      "rewards/final_brier_reward_step": 0.7887672185897827,
      "rewards/format_reward_step": 0.93359375,
      "step": 27
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4802441941574216,
      "aux_distill/mean_u": 0.33220112003744817,
      "aux_distill/n_active_tok": 234.8125,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.47613561553653716,
      "calib/avg_num_step_conf": 3.67578125,
      "calib/ece": 0.18265755102040815,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 0.012244897959183673,
      "calib/gap": -0.00416255760368664,
      "calib/mean_conf": 0.22349755102040816,
      "calib/mu_c": 0.21981071428571428,
      "calib/mu_w": 0.22397327188940092,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.14593469387755104,
      "calib/std_conf": 0.23264285704224408,
      "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": 2264.0,
      "completions/max_terminated_length": 2264.0,
      "completions/mean_length": 211.421875,
      "completions/mean_terminated_length": 212.25099182128906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 15.0,
      "epoch": 0.029866666666666666,
      "grad_norm": 0.011538458988070488,
      "learning_rate": 4.777777777777778e-06,
      "loss": 0.0221,
      "num_tokens": 6611792.0,
      "reward": 0.9147476553916931,
      "reward_std": 0.2037634551525116,
      "rewards/accuracy_reward_step": 0.11328125,
      "rewards/final_brier_reward_step": 0.7787140607833862,
      "rewards/format_reward_step": 0.9375,
      "step": 28
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4926240276545286,
      "aux_distill/mean_u": 0.36958355035024315,
      "aux_distill/n_active_tok": 216.5,
      "calib/answer_extract_rate": 0.95703125,
      "calib/auroc": 0.5128104575163399,
      "calib/avg_num_step_conf": 3.359375,
      "calib/ece": 0.20314049586776856,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.90625,
      "calib/frac_conf_gt_0.9": 0.01652892561983471,
      "calib/gap": 0.034379084967320256,
      "calib/mean_conf": 0.24685950413223137,
      "calib/mu_c": 0.2788235294117647,
      "calib/mu_w": 0.24444444444444444,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.96484375,
      "calib/pce": 0.18987603305785122,
      "calib/std_conf": 0.23114709658767324,
      "calib/step_conf_rate": 0.96484375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2303.0,
      "completions/max_terminated_length": 2303.0,
      "completions/mean_length": 199.52734375,
      "completions/mean_terminated_length": 200.309814453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 28.0,
      "epoch": 0.030933333333333334,
      "grad_norm": 0.012967920862138271,
      "learning_rate": 4.75e-06,
      "loss": 0.127,
      "num_tokens": 6769999.0,
      "reward": 0.8791476488113403,
      "reward_std": 0.2470008134841919,
      "rewards/accuracy_reward_step": 0.07421875,
      "rewards/final_brier_reward_step": 0.7778265476226807,
      "rewards/format_reward_step": 0.90625,
      "step": 29
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.49175314232707024,
      "aux_distill/mean_u": 0.32133843266442685,
      "aux_distill/n_active_tok": 209.53125,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.5404340716515534,
      "calib/avg_num_step_conf": 3.2578125,
      "calib/ece": 0.13880524193548385,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 0.012096774193548387,
      "calib/gap": 0.02508617511520736,
      "calib/mean_conf": 0.21417862903225804,
      "calib/mu_c": 0.2361290322580645,
      "calib/mu_w": 0.21104285714285714,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9765625,
      "calib/pce": 0.11399193548387095,
      "calib/std_conf": 0.2115425749942957,
      "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": 0.0,
      "completions/max_length": 2050.0,
      "completions/max_terminated_length": 2050.0,
      "completions/mean_length": 174.0546875,
      "completions/mean_terminated_length": 174.0546875,
      "completions/min_length": 44.0,
      "completions/min_terminated_length": 44.0,
      "epoch": 0.032,
      "grad_norm": 0.012511853128671646,
      "learning_rate": 4.722222222222222e-06,
      "loss": 0.0732,
      "num_tokens": 6921541.0,
      "reward": 0.9234327077865601,
      "reward_std": 0.19281959533691406,
      "rewards/accuracy_reward_step": 0.12109375,
      "rewards/final_brier_reward_step": 0.7882716655731201,
      "rewards/format_reward_step": 0.9375,
      "step": 30
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4863882064819336,
      "aux_distill/mean_u": 0.3262798401698808,
      "aux_distill/n_active_tok": 212.71875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5903232012312427,
      "calib/avg_num_step_conf": 3.33984375,
      "calib/ece": 0.15530120481927714,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.012048192771084338,
      "calib/gap": 0.0482185455944594,
      "calib/mean_conf": 0.22188755020080322,
      "calib/mu_c": 0.26565217391304347,
      "calib/mu_w": 0.21743362831858407,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.1424096385542169,
      "calib/std_conf": 0.2197334351794591,
      "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": 656.0,
      "completions/max_terminated_length": 656.0,
      "completions/mean_length": 163.23828125,
      "completions/mean_terminated_length": 163.87844848632812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 10.0,
      "epoch": 0.03306666666666667,
      "grad_norm": 0.010997715406119823,
      "learning_rate": 4.694444444444445e-06,
      "loss": 0.0457,
      "num_tokens": 7069242.0,
      "reward": 0.9402660131454468,
      "reward_std": 0.17586375772953033,
      "rewards/accuracy_reward_step": 0.09375,
      "rewards/final_brier_reward_step": 0.8258445262908936,
      "rewards/format_reward_step": 0.9609375,
      "step": 31
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5132203064858913,
      "aux_distill/mean_u": 0.34264686314469306,
      "aux_distill/n_active_tok": 211.03125,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5187995860641601,
      "calib/avg_num_step_conf": 3.265625,
      "calib/ece": 0.13620522088353412,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.008032128514056224,
      "calib/gap": 0.047089030700241424,
      "calib/mean_conf": 0.18590481927710845,
      "calib/mu_c": 0.2280769230769231,
      "calib/mu_w": 0.18098789237668167,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.10884618473895584,
      "calib/std_conf": 0.1932826076047128,
      "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": 517.0,
      "completions/max_terminated_length": 517.0,
      "completions/mean_length": 158.86328125,
      "completions/mean_terminated_length": 159.4862823486328,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 20.0,
      "epoch": 0.034133333333333335,
      "grad_norm": 0.012896922416985035,
      "learning_rate": 4.666666666666667e-06,
      "loss": -0.0091,
      "num_tokens": 7216615.0,
      "reward": 0.949363112449646,
      "reward_std": 0.14953142404556274,
      "rewards/accuracy_reward_step": 0.1015625,
      "rewards/final_brier_reward_step": 0.836226224899292,
      "rewards/format_reward_step": 0.9609375,
      "step": 32
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48916500341147184,
      "aux_distill/mean_u": 0.33687785008226384,
      "aux_distill/n_active_tok": 225.0625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5212639783124364,
      "calib/avg_num_step_conf": 3.51953125,
      "calib/ece": 0.13667549407114624,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.011857707509881422,
      "calib/gap": 0.006728464927143352,
      "calib/mean_conf": 0.1870399209486166,
      "calib/mu_c": 0.1930769230769231,
      "calib/mu_w": 0.18634845814977974,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.11047430830039523,
      "calib/std_conf": 0.19247799942857202,
      "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": 526.0,
      "completions/max_terminated_length": 526.0,
      "completions/mean_length": 171.6015625,
      "completions/mean_terminated_length": 172.27452087402344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.0352,
      "grad_norm": 0.010750737972557545,
      "learning_rate": 4.638888888888889e-06,
      "loss": 0.0778,
      "num_tokens": 7367417.0,
      "reward": 0.9722968339920044,
      "reward_std": 0.10676688700914383,
      "rewards/accuracy_reward_step": 0.1015625,
      "rewards/final_brier_reward_step": 0.8547499179840088,
      "rewards/format_reward_step": 0.98828125,
      "step": 33
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47535041719675064,
      "aux_distill/mean_u": 0.33305644552452596,
      "aux_distill/n_active_tok": 255.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.4020833333333334,
      "calib/avg_num_step_conf": 3.99609375,
      "calib/ece": 0.16385826771653544,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": -0.07352898550724639,
      "calib/mean_conf": 0.175748031496063,
      "calib/mu_c": 0.10916666666666668,
      "calib/mu_w": 0.18269565217391306,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.12255905511811026,
      "calib/std_conf": 0.18389067140982998,
      "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": 526.0,
      "completions/max_terminated_length": 526.0,
      "completions/mean_length": 177.60546875,
      "completions/mean_terminated_length": 178.30197143554688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 26.0,
      "epoch": 0.03626666666666667,
      "grad_norm": 0.011679654009640217,
      "learning_rate": 4.611111111111112e-06,
      "loss": 0.0003,
      "num_tokens": 7517996.0,
      "reward": 0.962560772895813,
      "reward_std": 0.10435040295124054,
      "rewards/accuracy_reward_step": 0.09375,
      "rewards/final_brier_reward_step": 0.8469964861869812,
      "rewards/format_reward_step": 0.984375,
      "step": 34
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48178594931960106,
      "aux_distill/mean_u": 0.3530250464581981,
      "aux_distill/n_active_tok": 261.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.6274481566820276,
      "calib/avg_num_step_conf": 4.09765625,
      "calib/ece": 0.06537254901960785,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.00392156862745098,
      "calib/gap": 0.06063940092165898,
      "calib/mean_conf": 0.14576470588235294,
      "calib/mu_c": 0.19903225806451613,
      "calib/mu_w": 0.13839285714285715,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.04478431372549019,
      "calib/std_conf": 0.15192394746434412,
      "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": 512.0,
      "completions/max_terminated_length": 512.0,
      "completions/mean_length": 186.82421875,
      "completions/mean_terminated_length": 187.55686950683594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 67.0,
      "epoch": 0.037333333333333336,
      "grad_norm": 0.010054091922938824,
      "learning_rate": 4.583333333333333e-06,
      "loss": 0.0217,
      "num_tokens": 7675079.0,
      "reward": 0.9981177449226379,
      "reward_std": 0.07535213232040405,
      "rewards/accuracy_reward_step": 0.12109375,
      "rewards/final_brier_reward_step": 0.8790480494499207,
      "rewards/format_reward_step": 0.99609375,
      "step": 35
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.49229648895561695,
      "aux_distill/mean_u": 0.33450523767788776,
      "aux_distill/n_active_tok": 244.6875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5527711084433774,
      "calib/avg_num_step_conf": 3.82421875,
      "calib/ece": 0.11885375494071149,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.01883253301320527,
      "calib/mean_conf": 0.11379446640316206,
      "calib/mu_c": 0.12897959183673469,
      "calib/mu_w": 0.11014705882352942,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.019486166007905134,
      "calib/std_conf": 0.13063477301985285,
      "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": 458.0,
      "completions/max_terminated_length": 458.0,
      "completions/mean_length": 173.2734375,
      "completions/mean_terminated_length": 173.9529571533203,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 44.0,
      "epoch": 0.0384,
      "grad_norm": 0.010825212113559246,
      "learning_rate": 4.555555555555556e-06,
      "loss": 0.0231,
      "num_tokens": 7822149.0,
      "reward": 0.998137354850769,
      "reward_std": 0.0790884867310524,
      "rewards/accuracy_reward_step": 0.19140625,
      "rewards/final_brier_reward_step": 0.8165871500968933,
      "rewards/format_reward_step": 0.98828125,
      "step": 36
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4659074544906616,
      "aux_distill/mean_u": 0.2862848587118983,
      "aux_distill/n_active_tok": 283.40625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.3690749235474007,
      "calib/avg_num_step_conf": 4.421875,
      "calib/ece": 0.13330708661417323,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.04467125382262997,
      "calib/mean_conf": 0.08417322834645669,
      "calib/mu_c": 0.04583333333333334,
      "calib/mu_w": 0.09050458715596331,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.03787401574803151,
      "calib/std_conf": 0.10079558283724063,
      "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": 820.0,
      "completions/max_terminated_length": 820.0,
      "completions/mean_length": 203.38671875,
      "completions/mean_terminated_length": 204.184326171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 49.0,
      "epoch": 0.039466666666666664,
      "grad_norm": 0.010696339420974255,
      "learning_rate": 4.527777777777778e-06,
      "loss": 0.0656,
      "num_tokens": 7981312.0,
      "reward": 0.9861996173858643,
      "reward_std": 0.04991428181529045,
      "rewards/accuracy_reward_step": 0.140625,
      "rewards/final_brier_reward_step": 0.8434929847717285,
      "rewards/format_reward_step": 0.98828125,
      "step": 37
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5006300732493401,
      "aux_distill/mean_u": 0.36540829187686324,
      "aux_distill/n_active_tok": 330.375,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.40192603095828905,
      "calib/avg_num_step_conf": 5.15625,
      "calib/ece": 0.135975390625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0076003426680846115,
      "calib/mean_conf": 0.062852734375,
      "calib/mu_c": 0.056410256410256404,
      "calib/mu_w": 0.06401059907834102,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.023242187499999997,
      "calib/std_conf": 0.08558604812204028,
      "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": 678.0,
      "completions/max_terminated_length": 678.0,
      "completions/mean_length": 222.60546875,
      "completions/mean_terminated_length": 223.4784393310547,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 52.0,
      "epoch": 0.04053333333333333,
      "grad_norm": 0.0095800356939435,
      "learning_rate": 4.5e-06,
      "loss": 0.0517,
      "num_tokens": 8145187.0,
      "reward": 0.999052882194519,
      "reward_std": 0.03454170748591423,
      "rewards/accuracy_reward_step": 0.15234375,
      "rewards/final_brier_reward_step": 0.8496682643890381,
      "rewards/format_reward_step": 0.99609375,
      "step": 38
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47253161668777466,
      "aux_distill/mean_u": 0.31778438508438045,
      "aux_distill/n_active_tok": 330.625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.41043385680063876,
      "calib/avg_num_step_conf": 5.1796875,
      "calib/ece": 0.12454901960784313,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.022895927601809953,
      "calib/mean_conf": 0.04925490196078431,
      "calib/mu_c": 0.029411764705882353,
      "calib/mu_w": 0.052307692307692305,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.020235294117647056,
      "calib/std_conf": 0.08329862252874574,
      "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": 604.0,
      "completions/max_terminated_length": 604.0,
      "completions/mean_length": 241.74609375,
      "completions/mean_terminated_length": 242.6941375732422,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 84.0,
      "epoch": 0.0416,
      "grad_norm": 0.010325453244149685,
      "learning_rate": 4.472222222222223e-06,
      "loss": 0.0637,
      "num_tokens": 8313162.0,
      "reward": 0.9914298057556152,
      "reward_std": 0.03904122859239578,
      "rewards/accuracy_reward_step": 0.1328125,
      "rewards/final_brier_reward_step": 0.85785973072052,
      "rewards/format_reward_step": 0.9921875,
      "step": 39
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.46040870528668165,
      "aux_distill/mean_u": 0.2964483882269135,
      "aux_distill/n_active_tok": 388.0625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4761849976990336,
      "calib/avg_num_step_conf": 6.0546875,
      "calib/ece": 0.14114624505928852,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0032581684307409105,
      "calib/mean_conf": 0.0283399209486166,
      "calib/mu_c": 0.025609756097560978,
      "calib/mu_w": 0.028867924528301888,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.003715415019762845,
      "calib/std_conf": 0.05019734835438953,
      "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": 1015.0,
      "completions/max_terminated_length": 1015.0,
      "completions/mean_length": 264.578125,
      "completions/mean_terminated_length": 265.6156921386719,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 92.0,
      "epoch": 0.042666666666666665,
      "grad_norm": 0.009926322847604752,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.0494,
      "num_tokens": 8487654.0,
      "reward": 0.9869291186332703,
      "reward_std": 0.05567903444170952,
      "rewards/accuracy_reward_step": 0.16015625,
      "rewards/final_brier_reward_step": 0.8293269872665405,
      "rewards/format_reward_step": 0.984375,
      "step": 40
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4719663066789508,
      "aux_distill/mean_u": 0.29660021391646385,
      "aux_distill/n_active_tok": 391.3125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5099252013808976,
      "calib/avg_num_step_conf": 6.125,
      "calib/ece": 0.29356862745098045,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00031933256616800687,
      "calib/mean_conf": 0.01623529411764706,
      "calib/mu_c": 0.016455696202531647,
      "calib/mu_w": 0.01613636363636364,
      "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": 0.02471689763775105,
      "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": 868.0,
      "completions/max_terminated_length": 868.0,
      "completions/mean_length": 269.08203125,
      "completions/mean_terminated_length": 270.13726806640625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 84.0,
      "epoch": 0.04373333333333333,
      "grad_norm": 0.008122802712023258,
      "learning_rate": 4.416666666666667e-06,
      "loss": 0.0428,
      "num_tokens": 8663787.0,
      "reward": 1.0007362365722656,
      "reward_std": 0.019822387024760246,
      "rewards/accuracy_reward_step": 0.30859375,
      "rewards/final_brier_reward_step": 0.6967852115631104,
      "rewards/format_reward_step": 0.99609375,
      "step": 41
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4728801045566797,
      "aux_distill/mean_u": 0.306885805886724,
      "aux_distill/n_active_tok": 396.75,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5105218855218855,
      "calib/avg_num_step_conf": 6.22265625,
      "calib/ece": 0.21075396825396825,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0065151515151515155,
      "calib/mean_conf": 0.014563492063492063,
      "calib/mu_c": 0.009444444444444446,
      "calib/mu_w": 0.015959595959595962,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.005515873015873016,
      "calib/std_conf": 0.04327156697463605,
      "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": 660.0,
      "completions/max_terminated_length": 660.0,
      "completions/mean_length": 263.90625,
      "completions/mean_terminated_length": 264.9411926269531,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.0448,
      "grad_norm": 0.009029773995280266,
      "learning_rate": 4.388888888888889e-06,
      "loss": 0.039,
      "num_tokens": 8835715.0,
      "reward": 0.9872943162918091,
      "reward_std": 0.04495030641555786,
      "rewards/accuracy_reward_step": 0.21484375,
      "rewards/final_brier_reward_step": 0.7753698825836182,
      "rewards/format_reward_step": 0.984375,
      "step": 42
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.44131849333643913,
      "aux_distill/mean_u": 0.2632251547550202,
      "aux_distill/n_active_tok": 413.8125,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4897545246515498,
      "calib/avg_num_step_conf": 6.4609375,
      "calib/ece": 0.17211764705882354,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0036966923236946133,
      "calib/mean_conf": 0.00827450980392157,
      "calib/mu_c": 0.011304347826086957,
      "calib/mu_w": 0.0076076555023923435,
      "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": 0.014258232477665623,
      "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": 777.0,
      "completions/max_terminated_length": 777.0,
      "completions/mean_length": 283.2265625,
      "completions/mean_terminated_length": 284.3372802734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.04586666666666667,
      "grad_norm": 0.006555310450494289,
      "learning_rate": 4.361111111111112e-06,
      "loss": 0.0454,
      "num_tokens": 9013445.0,
      "reward": 0.9940841197967529,
      "reward_std": 0.026495356112718582,
      "rewards/accuracy_reward_step": 0.1796875,
      "rewards/final_brier_reward_step": 0.8162933588027954,
      "rewards/format_reward_step": 0.9921875,
      "step": 43
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.46258422918617725,
      "aux_distill/mean_u": 0.31472563813177606,
      "aux_distill/n_active_tok": 493.59375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.437609756097561,
      "calib/avg_num_step_conf": 7.73828125,
      "calib/ece": 0.19019607843137254,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0028390243902439026,
      "calib/mean_conf": 0.005882352941176471,
      "calib/mu_c": 0.0036000000000000003,
      "calib/mu_w": 0.006439024390243903,
      "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": 0.010285430508159473,
      "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": 1352.0,
      "completions/max_terminated_length": 1352.0,
      "completions/mean_length": 324.62109375,
      "completions/mean_terminated_length": 325.8941345214844,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 103.0,
      "epoch": 0.046933333333333334,
      "grad_norm": 0.006308534182608128,
      "learning_rate": 4.333333333333334e-06,
      "loss": 0.0601,
      "num_tokens": 9202868.0,
      "reward": 0.9967269897460938,
      "reward_std": 0.012606674805283546,
      "rewards/accuracy_reward_step": 0.1953125,
      "rewards/final_brier_reward_step": 0.802047610282898,
      "rewards/format_reward_step": 0.99609375,
      "step": 44
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4415023336187005,
      "aux_distill/mean_u": 0.2417768633845245,
      "aux_distill/n_active_tok": 490.8125,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.5660124888492417,
      "calib/avg_num_step_conf": 7.5859375,
      "calib/ece": 0.2319277108433735,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0020838537020517398,
      "calib/mean_conf": 0.0050200803212851405,
      "calib/mu_c": 0.006610169491525424,
      "calib/mu_w": 0.004526315789473684,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.009188979692257087,
      "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": 2734.0,
      "completions/max_terminated_length": 2734.0,
      "completions/mean_length": 352.97265625,
      "completions/mean_terminated_length": 352.97265625,
      "completions/min_length": 91.0,
      "completions/min_terminated_length": 91.0,
      "epoch": 0.048,
      "grad_norm": 0.006134659983217716,
      "learning_rate": 4.305555555555556e-06,
      "loss": 0.1236,
      "num_tokens": 9398277.0,
      "reward": 0.9702200889587402,
      "reward_std": 0.0761524960398674,
      "rewards/accuracy_reward_step": 0.23046875,
      "rewards/final_brier_reward_step": 0.7412214875221252,
      "rewards/format_reward_step": 0.96875,
      "step": 45
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.46812527999281883,
      "aux_distill/mean_u": 0.3197349763021953,
      "aux_distill/n_active_tok": 522.28125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.49359292328042326,
      "calib/avg_num_step_conf": 8.1171875,
      "calib/ece": 0.24723320158102768,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0021585648148148145,
      "calib/mean_conf": 0.0057312252964426885,
      "calib/mu_c": 0.00734375,
      "calib/mu_w": 0.005185185185185185,
      "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": 0.013743805405934633,
      "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": 2256.0,
      "completions/max_terminated_length": 2256.0,
      "completions/mean_length": 351.23046875,
      "completions/mean_terminated_length": 351.23046875,
      "completions/min_length": 130.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.04906666666666667,
      "grad_norm": 0.00599120045080781,
      "learning_rate": 4.277777777777778e-06,
      "loss": 0.1249,
      "num_tokens": 9592960.0,
      "reward": 0.9900075793266296,
      "reward_std": 0.03566751629114151,
      "rewards/accuracy_reward_step": 0.25,
      "rewards/final_brier_reward_step": 0.741733968257904,
      "rewards/format_reward_step": 0.98828125,
      "step": 46
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.45625958032906055,
      "aux_distill/mean_u": 0.2927239472315959,
      "aux_distill/n_active_tok": 530.875,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5433224248091009,
      "calib/avg_num_step_conf": 8.4140625,
      "calib/ece": 0.26509803921568625,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0008181393174380557,
      "calib/mean_conf": 0.005490196078431373,
      "calib/mu_c": 0.006086956521739131,
      "calib/mu_w": 0.0052688172043010755,
      "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": 0.00988437576558473,
      "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": 985.0,
      "completions/max_terminated_length": 985.0,
      "completions/mean_length": 341.44921875,
      "completions/mean_terminated_length": 342.78826904296875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 115.0,
      "epoch": 0.050133333333333335,
      "grad_norm": 0.006031285040080547,
      "learning_rate": 4.25e-06,
      "loss": 0.0569,
      "num_tokens": 9786347.0,
      "reward": 0.9996237754821777,
      "reward_std": 0.008969930931925774,
      "rewards/accuracy_reward_step": 0.2734375,
      "rewards/final_brier_reward_step": 0.729716420173645,
      "rewards/format_reward_step": 0.99609375,
      "step": 47
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4554118439555168,
      "aux_distill/mean_u": 0.29516383089519194,
      "aux_distill/n_active_tok": 497.375,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5036963434022258,
      "calib/avg_num_step_conf": 7.796875,
      "calib/ece": 0.26541501976284587,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 3.1001589825119854e-05,
      "calib/mean_conf": 0.0033596837944664037,
      "calib/mu_c": 0.0033823529411764713,
      "calib/mu_w": 0.0033513513513513515,
      "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": 0.00630084653786219,
      "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": 853.0,
      "completions/max_terminated_length": 853.0,
      "completions/mean_length": 321.41796875,
      "completions/mean_terminated_length": 322.6784362792969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 101.0,
      "epoch": 0.0512,
      "grad_norm": 0.006986531894654036,
      "learning_rate": 4.222222222222223e-06,
      "loss": 0.0418,
      "num_tokens": 9972318.0,
      "reward": 0.9872015118598938,
      "reward_std": 0.04042886197566986,
      "rewards/accuracy_reward_step": 0.26953125,
      "rewards/final_brier_reward_step": 0.7204968333244324,
      "rewards/format_reward_step": 0.984375,
      "step": 48
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.45771760307252407,
      "aux_distill/mean_u": 0.30600149759672407,
      "aux_distill/n_active_tok": 529.34375,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.4506413192853871,
      "calib/avg_num_step_conf": 8.265625,
      "calib/ece": 0.29179282868525896,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0026523133302794317,
      "calib/mean_conf": 0.004302788844621514,
      "calib/mu_c": 0.0024324324324324323,
      "calib/mu_w": 0.005084745762711864,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0006374501992031873,
      "calib/std_conf": 0.01255315202204136,
      "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": 2488.0,
      "completions/max_terminated_length": 2488.0,
      "completions/mean_length": 358.4375,
      "completions/mean_terminated_length": 358.4375,
      "completions/min_length": 44.0,
      "completions/min_terminated_length": 44.0,
      "epoch": 0.05226666666666667,
      "grad_norm": 0.005493524018675089,
      "learning_rate": 4.194444444444445e-06,
      "loss": 0.0846,
      "num_tokens": 10168614.0,
      "reward": 0.9810855388641357,
      "reward_std": 0.05687147378921509,
      "rewards/accuracy_reward_step": 0.2890625,
      "rewards/final_brier_reward_step": 0.6926398277282715,
      "rewards/format_reward_step": 0.98046875,
      "step": 49
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.44665439892560244,
      "aux_distill/mean_u": 0.29054208955618344,
      "aux_distill/n_active_tok": 519.5625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.45397571484528004,
      "calib/avg_num_step_conf": 8.08203125,
      "calib/ece": 0.2678346456692914,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0022585193889541715,
      "calib/mean_conf": 0.0038188976377952757,
      "calib/mu_c": 0.0021739130434782613,
      "calib/mu_w": 0.004432432432432433,
      "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": 0.008693302085341467,
      "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": 1978.0,
      "completions/max_terminated_length": 1978.0,
      "completions/mean_length": 361.5,
      "completions/mean_terminated_length": 361.5,
      "completions/min_length": 130.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.05333333333333334,
      "grad_norm": 0.00480527663603425,
      "learning_rate": 4.166666666666667e-06,
      "loss": 0.0762,
      "num_tokens": 10366518.0,
      "reward": 0.9927287697792053,
      "reward_std": 0.023400694131851196,
      "rewards/accuracy_reward_step": 0.26953125,
      "rewards/final_brier_reward_step": 0.7237386703491211,
      "rewards/format_reward_step": 0.9921875,
      "step": 50
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4320847960188985,
      "aux_distill/mean_u": 0.25169377117910835,
      "aux_distill/n_active_tok": 547.625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.572090909090909,
      "calib/avg_num_step_conf": 8.68359375,
      "calib/ece": 0.21168627450980393,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0025500000000000006,
      "calib/mean_conf": 0.004,
      "calib/mu_c": 0.006000000000000001,
      "calib/mu_w": 0.0034500000000000004,
      "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": 0.00833607797938763,
      "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": 1333.0,
      "completions/max_terminated_length": 1333.0,
      "completions/mean_length": 357.3046875,
      "completions/mean_terminated_length": 358.7059020996094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 5.0,
      "epoch": 0.0544,
      "grad_norm": 0.005411915946751833,
      "learning_rate": 4.138888888888889e-06,
      "loss": 0.0307,
      "num_tokens": 10567284.0,
      "reward": 0.9856216907501221,
      "reward_std": 0.046699557453393936,
      "rewards/accuracy_reward_step": 0.21484375,
      "rewards/final_brier_reward_step": 0.7720246315002441,
      "rewards/format_reward_step": 0.984375,
      "step": 51
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.46485963091254234,
      "aux_distill/mean_u": 0.29594562104785566,
      "aux_distill/n_active_tok": 458.3125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4904846659364732,
      "calib/avg_num_step_conf": 7.25,
      "calib/ece": 0.3437007874015748,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -4.381161007667054e-05,
      "calib/mean_conf": 0.002755905511811024,
      "calib/mu_c": 0.0027272727272727275,
      "calib/mu_w": 0.002771084337349398,
      "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": 0.0057749343459240635,
      "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": 1609.0,
      "completions/max_terminated_length": 1609.0,
      "completions/mean_length": 325.39453125,
      "completions/mean_terminated_length": 326.67059326171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 120.0,
      "epoch": 0.055466666666666664,
      "grad_norm": 0.0054336413741111755,
      "learning_rate": 4.111111111111111e-06,
      "loss": 0.0414,
      "num_tokens": 10758537.0,
      "reward": 0.9891984462738037,
      "reward_std": 0.03528884798288345,
      "rewards/accuracy_reward_step": 0.34375,
      "rewards/final_brier_reward_step": 0.6463655829429626,
      "rewards/format_reward_step": 0.98828125,
      "step": 52
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.44012278597801924,
      "aux_distill/mean_u": 0.2862151821432428,
      "aux_distill/n_active_tok": 500.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5097378277153558,
      "calib/avg_num_step_conf": 8.2109375,
      "calib/ece": 0.2937944664031621,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0005947565543071168,
      "calib/mean_conf": 0.002648221343873518,
      "calib/mu_c": 0.003066666666666667,
      "calib/mu_w": 0.0024719101123595504,
      "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": 0.006136037429351788,
      "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": 2396.0,
      "completions/max_terminated_length": 2396.0,
      "completions/mean_length": 344.328125,
      "completions/mean_terminated_length": 345.678466796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 107.0,
      "epoch": 0.05653333333333333,
      "grad_norm": 0.005466809030622244,
      "learning_rate": 4.083333333333334e-06,
      "loss": 0.0771,
      "num_tokens": 10952509.0,
      "reward": 0.9852514863014221,
      "reward_std": 0.03851320222020149,
      "rewards/accuracy_reward_step": 0.29296875,
      "rewards/final_brier_reward_step": 0.6931593418121338,
      "rewards/format_reward_step": 0.984375,
      "step": 53
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43958625476807356,
      "aux_distill/mean_u": 0.2571615242030315,
      "aux_distill/n_active_tok": 463.28125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4883675464320626,
      "calib/avg_num_step_conf": 7.29296875,
      "calib/ece": 0.38732283464566925,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00035842293906810036,
      "calib/mean_conf": 0.0024409448818897644,
      "calib/mu_c": 0.0022222222222222222,
      "calib/mu_w": 0.0025806451612903226,
      "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": 0.0058480140559455576,
      "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": 939.0,
      "completions/max_terminated_length": 939.0,
      "completions/mean_length": 311.4921875,
      "completions/mean_terminated_length": 312.7137451171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 107.0,
      "epoch": 0.0576,
      "grad_norm": 0.005956499837338924,
      "learning_rate": 4.055555555555556e-06,
      "loss": 0.0511,
      "num_tokens": 11138483.0,
      "reward": 0.9930269718170166,
      "reward_std": 0.023992538452148438,
      "rewards/accuracy_reward_step": 0.38671875,
      "rewards/final_brier_reward_step": 0.6071476340293884,
      "rewards/format_reward_step": 0.9921875,
      "step": 54
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4438884211704135,
      "aux_distill/mean_u": 0.3088732302722088,
      "aux_distill/n_active_tok": 503.1875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5007692307692309,
      "calib/avg_num_step_conf": 7.84765625,
      "calib/ece": 0.23266666666666666,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0006025641025641029,
      "calib/mean_conf": 0.002627450980392157,
      "calib/mu_c": 0.0021666666666666666,
      "calib/mu_w": 0.0027692307692307695,
      "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": 0.006051934658466017,
      "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": 1874.0,
      "completions/max_terminated_length": 1874.0,
      "completions/mean_length": 349.70703125,
      "completions/mean_terminated_length": 349.70703125,
      "completions/min_length": 116.0,
      "completions/min_terminated_length": 116.0,
      "epoch": 0.058666666666666666,
      "grad_norm": 0.004821429029107094,
      "learning_rate": 4.027777777777779e-06,
      "loss": 0.0885,
      "num_tokens": 11335832.0,
      "reward": 0.9965798258781433,
      "reward_std": 0.012168581597507,
      "rewards/accuracy_reward_step": 0.234375,
      "rewards/final_brier_reward_step": 0.7626910209655762,
      "rewards/format_reward_step": 0.99609375,
      "step": 55
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43843644205480814,
      "aux_distill/mean_u": 0.2908842595016219,
      "aux_distill/n_active_tok": 492.84375,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4863632452895122,
      "calib/avg_num_step_conf": 7.72265625,
      "calib/ece": 0.2266015625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.001141701798158824,
      "calib/mean_conf": 0.0038671875,
      "calib/mu_c": 0.004745762711864407,
      "calib/mu_w": 0.0036040609137055835,
      "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": 0.007972483041050872,
      "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": 1079.0,
      "completions/max_terminated_length": 1079.0,
      "completions/mean_length": 333.1328125,
      "completions/mean_terminated_length": 334.4392395019531,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.05973333333333333,
      "grad_norm": 0.0058794510550796986,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0311,
      "num_tokens": 11527954.0,
      "reward": 1.0010545253753662,
      "reward_std": 0.0027934759855270386,
      "rewards/accuracy_reward_step": 0.23046875,
      "rewards/final_brier_reward_step": 0.7716401815414429,
      "rewards/format_reward_step": 1.0,
      "step": 56
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.448484574444592,
      "aux_distill/mean_u": 0.2557524902818454,
      "aux_distill/n_active_tok": 464.9375,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5290054249547921,
      "calib/avg_num_step_conf": 7.2578125,
      "calib/ece": 0.3068503937007874,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0007406871609403256,
      "calib/mean_conf": 0.004173228346456693,
      "calib/mu_c": 0.0046835443037974685,
      "calib/mu_w": 0.003942857142857143,
      "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": 0.007202465116264535,
      "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": 1941.0,
      "completions/max_terminated_length": 1941.0,
      "completions/mean_length": 322.2421875,
      "completions/mean_terminated_length": 322.2421875,
      "completions/min_length": 113.0,
      "completions/min_terminated_length": 113.0,
      "epoch": 0.0608,
      "grad_norm": 0.005439405329525471,
      "learning_rate": 3.972222222222223e-06,
      "loss": 0.0748,
      "num_tokens": 11717240.0,
      "reward": 0.9935984015464783,
      "reward_std": 0.024633333086967468,
      "rewards/accuracy_reward_step": 0.30859375,
      "rewards/final_brier_reward_step": 0.6864155530929565,
      "rewards/format_reward_step": 0.9921875,
      "step": 57
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43026210833340883,
      "aux_distill/mean_u": 0.27150247398639,
      "aux_distill/n_active_tok": 540.8125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4831393867323537,
      "calib/avg_num_step_conf": 8.44921875,
      "calib/ece": 0.23622047244094488,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0006506412978849914,
      "calib/mean_conf": 0.003937007874015748,
      "calib/mu_c": 0.0034426229508196723,
      "calib/mu_w": 0.004093264248704664,
      "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": 0.008050771989682681,
      "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": 2337.0,
      "completions/max_terminated_length": 2337.0,
      "completions/mean_length": 370.8203125,
      "completions/mean_terminated_length": 370.8203125,
      "completions/min_length": 94.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.06186666666666667,
      "grad_norm": 0.004779992159456015,
      "learning_rate": 3.944444444444445e-06,
      "loss": 0.0792,
      "num_tokens": 11918490.0,
      "reward": 0.992967963218689,
      "reward_std": 0.024048078805208206,
      "rewards/accuracy_reward_step": 0.23828125,
      "rewards/final_brier_reward_step": 0.7554671764373779,
      "rewards/format_reward_step": 0.9921875,
      "step": 58
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4273897521197796,
      "aux_distill/mean_u": 0.24913954184704293,
      "aux_distill/n_active_tok": 500.46875,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4096424010217114,
      "calib/avg_num_step_conf": 7.79296875,
      "calib/ece": 0.31388235294117645,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0017177522349936139,
      "calib/mean_conf": 0.003764705882352941,
      "calib/mu_c": 0.002592592592592593,
      "calib/mu_w": 0.004310344827586207,
      "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": 0.006381923722804189,
      "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": 1845.0,
      "completions/max_terminated_length": 1845.0,
      "completions/mean_length": 332.90625,
      "completions/mean_terminated_length": 332.90625,
      "completions/min_length": 114.0,
      "completions/min_terminated_length": 114.0,
      "epoch": 0.06293333333333333,
      "grad_norm": 0.004225445911288261,
      "learning_rate": 3.916666666666667e-06,
      "loss": 0.0848,
      "num_tokens": 12109962.0,
      "reward": 0.9968867301940918,
      "reward_std": 0.013021206483244896,
      "rewards/accuracy_reward_step": 0.31640625,
      "rewards/final_brier_reward_step": 0.6812734603881836,
      "rewards/format_reward_step": 0.99609375,
      "step": 59
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43960097804665565,
      "aux_distill/mean_u": 0.261367273698288,
      "aux_distill/n_active_tok": 483.0,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5204790249433107,
      "calib/avg_num_step_conf": 7.546875,
      "calib/ece": 0.3288888888888889,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0006547619047619054,
      "calib/mean_conf": 0.0044444444444444444,
      "calib/mu_c": 0.004880952380952382,
      "calib/mu_w": 0.004226190476190476,
      "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": 0.007133150370990895,
      "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": 2427.0,
      "completions/max_terminated_length": 2427.0,
      "completions/mean_length": 340.19921875,
      "completions/mean_terminated_length": 340.19921875,
      "completions/min_length": 128.0,
      "completions/min_terminated_length": 128.0,
      "epoch": 0.064,
      "grad_norm": 0.006468915846198797,
      "learning_rate": 3.88888888888889e-06,
      "loss": 0.1286,
      "num_tokens": 12305909.0,
      "reward": 0.9859417676925659,
      "reward_std": 0.04724940285086632,
      "rewards/accuracy_reward_step": 0.328125,
      "rewards/final_brier_reward_step": 0.6593835353851318,
      "rewards/format_reward_step": 0.984375,
      "step": 60
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43225711584091187,
      "aux_distill/mean_u": 0.24751454538576817,
      "aux_distill/n_active_tok": 421.46875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.519921875,
      "calib/avg_num_step_conf": 6.578125,
      "calib/ece": 0.3690234375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00043750000000000126,
      "calib/mean_conf": 0.0059765625,
      "calib/mu_c": 0.006250000000000001,
      "calib/mu_w": 0.0058125,
      "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": 0.00942815335490433,
      "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": 738.0,
      "completions/max_terminated_length": 738.0,
      "completions/mean_length": 286.140625,
      "completions/mean_terminated_length": 287.26275634765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 106.0,
      "epoch": 0.06506666666666666,
      "grad_norm": 0.005420437548309565,
      "learning_rate": 3.861111111111112e-06,
      "loss": 0.0414,
      "num_tokens": 12483225.0,
      "reward": 0.9983751773834229,
      "reward_std": 0.014877854846417904,
      "rewards/accuracy_reward_step": 0.375,
      "rewards/final_brier_reward_step": 0.6256566047668457,
      "rewards/format_reward_step": 0.99609375,
      "step": 61
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4199416805058718,
      "aux_distill/mean_u": 0.25439553840276957,
      "aux_distill/n_active_tok": 476.78125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5237889273356402,
      "calib/avg_num_step_conf": 7.44921875,
      "calib/ece": 0.2592549019607843,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.001323529411764706,
      "calib/mean_conf": 0.007411764705882353,
      "calib/mu_c": 0.008382352941176471,
      "calib/mu_w": 0.007058823529411765,
      "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": 0.011494545118187515,
      "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": 1541.0,
      "completions/max_terminated_length": 1541.0,
      "completions/mean_length": 337.1015625,
      "completions/mean_terminated_length": 337.1015625,
      "completions/min_length": 127.0,
      "completions/min_terminated_length": 127.0,
      "epoch": 0.06613333333333334,
      "grad_norm": 0.005449065938591957,
      "learning_rate": 3.833333333333334e-06,
      "loss": 0.0391,
      "num_tokens": 12676603.0,
      "reward": 0.9982271194458008,
      "reward_std": 0.01538827084004879,
      "rewards/accuracy_reward_step": 0.265625,
      "rewards/final_brier_reward_step": 0.7347354888916016,
      "rewards/format_reward_step": 0.99609375,
      "step": 62
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.436340908985585,
      "aux_distill/mean_u": 0.28511982269740827,
      "aux_distill/n_active_tok": 459.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5146272412708399,
      "calib/avg_num_step_conf": 7.21875,
      "calib/ece": 0.2592549019607843,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00012032085561497399,
      "calib/mean_conf": 0.007411764705882354,
      "calib/mu_c": 0.0075,
      "calib/mu_w": 0.007379679144385026,
      "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": 0.009998385106053397,
      "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": 1058.0,
      "completions/max_terminated_length": 1058.0,
      "completions/mean_length": 320.92578125,
      "completions/mean_terminated_length": 322.184326171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 98.0,
      "epoch": 0.0672,
      "grad_norm": 0.0054280711337924,
      "learning_rate": 3.8055555555555556e-06,
      "loss": 0.0384,
      "num_tokens": 12867400.0,
      "reward": 0.9980087280273438,
      "reward_std": 0.015086276456713676,
      "rewards/accuracy_reward_step": 0.265625,
      "rewards/final_brier_reward_step": 0.734298825263977,
      "rewards/format_reward_step": 0.99609375,
      "step": 63
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4061308619566262,
      "aux_distill/mean_u": 0.2416082616255521,
      "aux_distill/n_active_tok": 445.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.47493517718236816,
      "calib/avg_num_step_conf": 7.0546875,
      "calib/ece": 0.2969921875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0011106309420916153,
      "calib/mean_conf": 0.0076953125,
      "calib/mu_c": 0.006923076923076923,
      "calib/mu_w": 0.008033707865168539,
      "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": 0.008736649273453969,
      "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": 1143.0,
      "completions/max_terminated_length": 1143.0,
      "completions/mean_length": 292.5625,
      "completions/mean_terminated_length": 293.7098083496094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 112.0,
      "epoch": 0.06826666666666667,
      "grad_norm": 0.006516458000987768,
      "learning_rate": 3.777777777777778e-06,
      "loss": 0.0512,
      "num_tokens": 13046072.0,
      "reward": 1.0020415782928467,
      "reward_std": 0.003972196485847235,
      "rewards/accuracy_reward_step": 0.3046875,
      "rewards/final_brier_reward_step": 0.6993956565856934,
      "rewards/format_reward_step": 1.0,
      "step": 64
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4346964657306671,
      "aux_distill/mean_u": 0.25160164239136495,
      "aux_distill/n_active_tok": 407.65625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5213645278789577,
      "calib/avg_num_step_conf": 6.49609375,
      "calib/ece": 0.3174901960784314,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0006430372653404327,
      "calib/mean_conf": 0.008,
      "calib/mu_c": 0.008433734939759036,
      "calib/mu_w": 0.0077906976744186035,
      "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": 0.008466774740702857,
      "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": 789.0,
      "completions/max_terminated_length": 789.0,
      "completions/mean_length": 281.6328125,
      "completions/mean_terminated_length": 282.7372741699219,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 96.0,
      "epoch": 0.06933333333333333,
      "grad_norm": 0.005870714783668518,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.0537,
      "num_tokens": 13223194.0,
      "reward": 0.9948544502258301,
      "reward_std": 0.02620820514857769,
      "rewards/accuracy_reward_step": 0.32421875,
      "rewards/final_brier_reward_step": 0.6733027696609497,
      "rewards/format_reward_step": 0.9921875,
      "step": 65
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43573713721707463,
      "aux_distill/mean_u": 0.27839431637130474,
      "aux_distill/n_active_tok": 472.9375,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5154248892468788,
      "calib/avg_num_step_conf": 7.40625,
      "calib/ece": 0.245,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.000641159887233185,
      "calib/mean_conf": 0.008906250000000001,
      "calib/mu_c": 0.009384615384615384,
      "calib/mu_w": 0.008743455497382199,
      "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": 0.009579794409980833,
      "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": 887.0,
      "completions/max_terminated_length": 887.0,
      "completions/mean_length": 312.73046875,
      "completions/mean_terminated_length": 313.9568786621094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.0704,
      "grad_norm": 0.005807508714497089,
      "learning_rate": 3.7222222222222225e-06,
      "loss": 0.0313,
      "num_tokens": 13409605.0,
      "reward": 1.0022971630096436,
      "reward_std": 0.0038538267835974693,
      "rewards/accuracy_reward_step": 0.25390625,
      "rewards/final_brier_reward_step": 0.7506882548332214,
      "rewards/format_reward_step": 1.0,
      "step": 66
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43293993175029755,
      "aux_distill/mean_u": 0.2693137654714224,
      "aux_distill/n_active_tok": 454.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6072888763569716,
      "calib/avg_num_step_conf": 7.09375,
      "calib/ece": 0.3119607843137255,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.002733681094036374,
      "calib/mean_conf": 0.011803921568627451,
      "calib/mu_c": 0.013658536585365855,
      "calib/mu_w": 0.010924855491329481,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0010980392156862747,
      "calib/std_conf": 0.019620936804645387,
      "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": 2418.0,
      "completions/max_terminated_length": 2418.0,
      "completions/mean_length": 318.67578125,
      "completions/mean_terminated_length": 318.67578125,
      "completions/min_length": 109.0,
      "completions/min_terminated_length": 109.0,
      "epoch": 0.07146666666666666,
      "grad_norm": 0.006566639989614487,
      "learning_rate": 3.694444444444445e-06,
      "loss": 0.0716,
      "num_tokens": 13596194.0,
      "reward": 1.000207543373108,
      "reward_std": 0.01748921349644661,
      "rewards/accuracy_reward_step": 0.3203125,
      "rewards/final_brier_reward_step": 0.6840089559555054,
      "rewards/format_reward_step": 0.99609375,
      "step": 67
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.42459358740597963,
      "aux_distill/mean_u": 0.25818829225068923,
      "aux_distill/n_active_tok": 435.625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5195084485407067,
      "calib/avg_num_step_conf": 6.80078125,
      "calib/ece": 0.2613671875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0002964669738863262,
      "calib/mean_conf": 0.0120703125,
      "calib/mu_c": 0.012285714285714285,
      "calib/mu_w": 0.011989247311827959,
      "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": 0.01231150604728535,
      "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": 299.2109375,
      "completions/mean_terminated_length": 300.38433837890625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 93.0,
      "epoch": 0.07253333333333334,
      "grad_norm": 0.006065750028938055,
      "learning_rate": 3.6666666666666666e-06,
      "loss": 0.0443,
      "num_tokens": 13776880.0,
      "reward": 1.0032107830047607,
      "reward_std": 0.005062174052000046,
      "rewards/accuracy_reward_step": 0.2734375,
      "rewards/final_brier_reward_step": 0.7329839468002319,
      "rewards/format_reward_step": 1.0,
      "step": 68
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43279402423650026,
      "aux_distill/mean_u": 0.27462993316679674,
      "aux_distill/n_active_tok": 432.625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5322975964579381,
      "calib/avg_num_step_conf": 6.91015625,
      "calib/ece": 0.25610236220472443,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0004063883617963348,
      "calib/mean_conf": 0.011614173228346457,
      "calib/mu_c": 0.011911764705882354,
      "calib/mu_w": 0.01150537634408602,
      "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": 0.013107398430191524,
      "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": 1351.0,
      "completions/max_terminated_length": 1351.0,
      "completions/mean_length": 316.20703125,
      "completions/mean_terminated_length": 317.44708251953125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 132.0,
      "epoch": 0.0736,
      "grad_norm": 0.006692867260426283,
      "learning_rate": 3.638888888888889e-06,
      "loss": 0.0452,
      "num_tokens": 13962325.0,
      "reward": 0.99129319190979,
      "reward_std": 0.03864618018269539,
      "rewards/accuracy_reward_step": 0.265625,
      "rewards/final_brier_reward_step": 0.7286800742149353,
      "rewards/format_reward_step": 0.98828125,
      "step": 69
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43750183191150427,
      "aux_distill/mean_u": 0.25953609603320626,
      "aux_distill/n_active_tok": 443.5625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5386486486486486,
      "calib/avg_num_step_conf": 6.94921875,
      "calib/ece": 0.26250980392156864,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0009845559845559869,
      "calib/mean_conf": 0.012000000000000002,
      "calib/mu_c": 0.012714285714285716,
      "calib/mu_w": 0.01172972972972973,
      "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": 0.010267023166715966,
      "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": 721.0,
      "completions/max_terminated_length": 721.0,
      "completions/mean_length": 315.14453125,
      "completions/mean_terminated_length": 316.3804016113281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 80.0,
      "epoch": 0.07466666666666667,
      "grad_norm": 0.006321345455944538,
      "learning_rate": 3.6111111111111115e-06,
      "loss": 0.0588,
      "num_tokens": 14149994.0,
      "reward": 0.9994460344314575,
      "reward_std": 0.015884429216384888,
      "rewards/accuracy_reward_step": 0.2734375,
      "rewards/final_brier_reward_step": 0.7293609380722046,
      "rewards/format_reward_step": 0.99609375,
      "step": 70
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.42223017755895853,
      "aux_distill/mean_u": 0.2762904479705797,
      "aux_distill/n_active_tok": 491.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.46249290981281904,
      "calib/avg_num_step_conf": 7.67578125,
      "calib/ece": 0.3091338582677165,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0015031196823596157,
      "calib/mean_conf": 0.013700787401574804,
      "calib/mu_c": 0.012682926829268294,
      "calib/mu_w": 0.01418604651162791,
      "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": 0.012536758526055635,
      "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": 2448.0,
      "completions/max_terminated_length": 2448.0,
      "completions/mean_length": 364.0234375,
      "completions/mean_terminated_length": 364.0234375,
      "completions/min_length": 111.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.07573333333333333,
      "grad_norm": 0.005532794166356325,
      "learning_rate": 3.5833333333333335e-06,
      "loss": 0.0521,
      "num_tokens": 14347592.0,
      "reward": 0.9960788488388062,
      "reward_std": 0.028619343414902687,
      "rewards/accuracy_reward_step": 0.3203125,
      "rewards/final_brier_reward_step": 0.6796578168869019,
      "rewards/format_reward_step": 0.9921875,
      "step": 71
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4384244615212083,
      "aux_distill/mean_u": 0.2782737296333412,
      "aux_distill/n_active_tok": 492.34375,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.542580170157068,
      "calib/avg_num_step_conf": 7.7109375,
      "calib/ece": 0.23709803921568628,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0014921465968586387,
      "calib/mean_conf": 0.01388235294117647,
      "calib/mu_c": 0.015,
      "calib/mu_w": 0.01350785340314136,
      "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": 0.011956862102007072,
      "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": 897.0,
      "completions/max_terminated_length": 897.0,
      "completions/mean_length": 330.2734375,
      "completions/mean_terminated_length": 331.5686340332031,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 139.0,
      "epoch": 0.0768,
      "grad_norm": 0.006522673182189465,
      "learning_rate": 3.555555555555556e-06,
      "loss": 0.0565,
      "num_tokens": 14536550.0,
      "reward": 1.0016295909881592,
      "reward_std": 0.01137695461511612,
      "rewards/accuracy_reward_step": 0.25390625,
      "rewards/final_brier_reward_step": 0.7532593607902527,
      "rewards/format_reward_step": 0.99609375,
      "step": 72
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4227779684588313,
      "aux_distill/mean_u": 0.21715057211677818,
      "aux_distill/n_active_tok": 431.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4812515915457091,
      "calib/avg_num_step_conf": 6.8203125,
      "calib/ece": 0.38316406249999996,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.001501145912910621,
      "calib/mean_conf": 0.0152734375,
      "calib/mu_c": 0.016176470588235296,
      "calib/mu_w": 0.014675324675324675,
      "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": 0.020289011359196234,
      "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": 979.0,
      "completions/max_terminated_length": 979.0,
      "completions/mean_length": 326.24609375,
      "completions/mean_terminated_length": 327.5255126953125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.07786666666666667,
      "grad_norm": 0.00622026901692152,
      "learning_rate": 3.5277777777777784e-06,
      "loss": 0.0268,
      "num_tokens": 14727101.0,
      "reward": 1.0022165775299072,
      "reward_std": 0.020840127021074295,
      "rewards/accuracy_reward_step": 0.3984375,
      "rewards/final_brier_reward_step": 0.6099019646644592,
      "rewards/format_reward_step": 0.99609375,
      "step": 73
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4327737260609865,
      "aux_distill/mean_u": 0.24665208091788085,
      "aux_distill/n_active_tok": 452.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4544215425531915,
      "calib/avg_num_step_conf": 7.203125,
      "calib/ece": 0.3535433070866142,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0033670212765957407,
      "calib/mean_conf": 0.01850393700787402,
      "calib/mu_c": 0.01638297872340426,
      "calib/mu_w": 0.01975,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.000984251968503937,
      "calib/std_conf": 0.019665659751870696,
      "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": 1317.0,
      "completions/max_terminated_length": 1317.0,
      "completions/mean_length": 333.16796875,
      "completions/mean_terminated_length": 334.4745178222656,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 107.0,
      "epoch": 0.07893333333333333,
      "grad_norm": 0.006490938365459442,
      "learning_rate": 3.5e-06,
      "loss": 0.0469,
      "num_tokens": 14916320.0,
      "reward": 0.9997944235801697,
      "reward_std": 0.02420826628804207,
      "rewards/accuracy_reward_step": 0.37109375,
      "rewards/final_brier_reward_step": 0.6363078355789185,
      "rewards/format_reward_step": 0.9921875,
      "step": 74
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4080852912738919,
      "aux_distill/mean_u": 0.2462691897632438,
      "aux_distill/n_active_tok": 426.53125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4701713299642548,
      "calib/avg_num_step_conf": 6.6640625,
      "calib/ece": 0.5009803921568627,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.001543818562800444,
      "calib/mean_conf": 0.02058823529411765,
      "calib/mu_c": 0.019849624060150377,
      "calib/mu_w": 0.02139344262295082,
      "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": 0.014820728026610377,
      "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": 1676.0,
      "completions/max_terminated_length": 1676.0,
      "completions/mean_length": 334.79296875,
      "completions/mean_terminated_length": 334.79296875,
      "completions/min_length": 131.0,
      "completions/min_terminated_length": 131.0,
      "epoch": 0.08,
      "grad_norm": 0.006999581586569548,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 0.0716,
      "num_tokens": 15106779.0,
      "reward": 1.006085753440857,
      "reward_std": 0.022710826247930527,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/final_brier_reward_step": 0.49654650688171387,
      "rewards/format_reward_step": 0.99609375,
      "step": 75
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4355225879698992,
      "aux_distill/mean_u": 0.2736861558638344,
      "aux_distill/n_active_tok": 451.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5060080106809078,
      "calib/avg_num_step_conf": 7.0546875,
      "calib/ece": 0.40169291338582674,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0007482993197278874,
      "calib/mean_conf": 0.019566929133858272,
      "calib/mu_c": 0.019999999999999997,
      "calib/mu_w": 0.01925170068027211,
      "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": 0.015376170617330107,
      "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": 1253.0,
      "completions/max_terminated_length": 1253.0,
      "completions/mean_length": 341.36328125,
      "completions/mean_terminated_length": 342.7019958496094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.08106666666666666,
      "grad_norm": 0.006661584135144949,
      "learning_rate": 3.444444444444445e-06,
      "loss": 0.0017,
      "num_tokens": 15297224.0,
      "reward": 1.000239610671997,
      "reward_std": 0.03131163865327835,
      "rewards/accuracy_reward_step": 0.41796875,
      "rewards/final_brier_reward_step": 0.5903230905532837,
      "rewards/format_reward_step": 0.9921875,
      "step": 76
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4306439831852913,
      "aux_distill/mean_u": 0.27393377477924047,
      "aux_distill/n_active_tok": 436.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5247457206648474,
      "calib/avg_num_step_conf": 6.8515625,
      "calib/ece": 0.4325882352941176,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0014487720168692604,
      "calib/mean_conf": 0.022313725490196085,
      "calib/mu_c": 0.023103448275862068,
      "calib/mu_w": 0.021654676258992808,
      "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": 0.01670210764682748,
      "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": 962.0,
      "completions/max_terminated_length": 962.0,
      "completions/mean_length": 335.9609375,
      "completions/mean_terminated_length": 337.2784423828125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 117.0,
      "epoch": 0.08213333333333334,
      "grad_norm": 0.007040133234113455,
      "learning_rate": 3.416666666666667e-06,
      "loss": 0.0235,
      "num_tokens": 15487894.0,
      "reward": 1.0061755180358887,
      "reward_std": 0.024019623175263405,
      "rewards/accuracy_reward_step": 0.453125,
      "rewards/final_brier_reward_step": 0.5631324052810669,
      "rewards/format_reward_step": 0.99609375,
      "step": 77
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4400795288383961,
      "aux_distill/mean_u": 0.27258292125392275,
      "aux_distill/n_active_tok": 472.9375,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.49562637631514556,
      "calib/avg_num_step_conf": 7.4375,
      "calib/ece": 0.44999999999999996,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.00390625,
      "calib/gap": 0.006491313922192314,
      "calib/mean_conf": 0.027421875,
      "calib/mu_c": 0.03081967213114754,
      "calib/mu_w": 0.024328358208955226,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0004296875,
      "calib/std_conf": 0.06321152601768426,
      "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": 973.0,
      "completions/max_terminated_length": 973.0,
      "completions/mean_length": 385.5,
      "completions/mean_terminated_length": 387.01177978515625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 144.0,
      "epoch": 0.0832,
      "grad_norm": 0.006053322460502386,
      "learning_rate": 3.3888888888888893e-06,
      "loss": 0.0424,
      "num_tokens": 15694606.0,
      "reward": 1.0123136043548584,
      "reward_std": 0.016614407300949097,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.5480648279190063,
      "rewards/format_reward_step": 1.0,
      "step": 78
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41587979532778263,
      "aux_distill/mean_u": 0.26884553844139497,
      "aux_distill/n_active_tok": 484.9375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.49537037037037035,
      "calib/avg_num_step_conf": 7.57421875,
      "calib/ece": 0.3965490196078431,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 9.826152683295453e-05,
      "calib/mean_conf": 0.026980392156862747,
      "calib/mu_c": 0.027037037037037037,
      "calib/mu_w": 0.026938775510204082,
      "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": 0.02005628374084471,
      "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": 1201.0,
      "completions/max_terminated_length": 1201.0,
      "completions/mean_length": 411.1953125,
      "completions/mean_terminated_length": 412.807861328125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 132.0,
      "epoch": 0.08426666666666667,
      "grad_norm": 0.0055940039455890656,
      "learning_rate": 3.3611111111111117e-06,
      "loss": 0.0385,
      "num_tokens": 15906248.0,
      "reward": 1.006937026977539,
      "reward_std": 0.024199217557907104,
      "rewards/accuracy_reward_step": 0.421875,
      "rewards/final_brier_reward_step": 0.5959054231643677,
      "rewards/format_reward_step": 0.99609375,
      "step": 79
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43790998961776495,
      "aux_distill/mean_u": 0.28386136577758153,
      "aux_distill/n_active_tok": 444.625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5211038961038962,
      "calib/avg_num_step_conf": 6.9375,
      "calib/ece": 0.41011725490196077,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0009577172827172874,
      "calib/mean_conf": 0.03098078431372549,
      "calib/mu_c": 0.031517857142857146,
      "calib/mu_w": 0.03056013986013986,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0009411764705882352,
      "calib/std_conf": 0.019306680327472275,
      "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": 1641.0,
      "completions/max_terminated_length": 1641.0,
      "completions/mean_length": 368.2890625,
      "completions/mean_terminated_length": 368.2890625,
      "completions/min_length": 140.0,
      "completions/min_terminated_length": 140.0,
      "epoch": 0.08533333333333333,
      "grad_norm": 0.0062352754175662994,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.0641,
      "num_tokens": 16102690.0,
      "reward": 1.0092190504074097,
      "reward_std": 0.025584885850548744,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/final_brier_reward_step": 0.5848445296287537,
      "rewards/format_reward_step": 0.99609375,
      "step": 80
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4371841801330447,
      "aux_distill/mean_u": 0.27223928922651247,
      "aux_distill/n_active_tok": 460.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4884761904761904,
      "calib/avg_num_step_conf": 7.1953125,
      "calib/ece": 0.375843137254902,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.001161904761904753,
      "calib/mean_conf": 0.035921568627450974,
      "calib/mu_c": 0.03523809523809524,
      "calib/mu_w": 0.036399999999999995,
      "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": 0.02466707191287799,
      "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": 2562.0,
      "completions/max_terminated_length": 2562.0,
      "completions/mean_length": 391.625,
      "completions/mean_terminated_length": 391.625,
      "completions/min_length": 144.0,
      "completions/min_terminated_length": 144.0,
      "epoch": 0.0864,
      "grad_norm": 0.006331518292427063,
      "learning_rate": 3.3055555555555558e-06,
      "loss": 0.044,
      "num_tokens": 16309194.0,
      "reward": 1.00960111618042,
      "reward_std": 0.0268442090600729,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.6129523515701294,
      "rewards/format_reward_step": 0.99609375,
      "step": 81
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.419726787135005,
      "aux_distill/mean_u": 0.23885014784892913,
      "aux_distill/n_active_tok": 417.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5527279616064663,
      "calib/avg_num_step_conf": 6.51953125,
      "calib/ece": 0.542470588235294,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0009939378630967471,
      "calib/mean_conf": 0.04325490196078432,
      "calib/mu_c": 0.04283783783783784,
      "calib/mu_w": 0.043831775700934585,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.002666666666666667,
      "calib/std_conf": 0.03865373844246665,
      "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": 2239.0,
      "completions/max_terminated_length": 2239.0,
      "completions/mean_length": 370.8671875,
      "completions/mean_terminated_length": 370.8671875,
      "completions/min_length": 137.0,
      "completions/min_terminated_length": 137.0,
      "epoch": 0.08746666666666666,
      "grad_norm": 0.006565708201378584,
      "learning_rate": 3.277777777777778e-06,
      "loss": 0.0597,
      "num_tokens": 16509688.0,
      "reward": 1.0191833972930908,
      "reward_std": 0.033703260123729706,
      "rewards/accuracy_reward_step": 0.578125,
      "rewards/final_brier_reward_step": 0.46414804458618164,
      "rewards/format_reward_step": 0.99609375,
      "step": 82
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40507653122767806,
      "aux_distill/mean_u": 0.2709985755049188,
      "aux_distill/n_active_tok": 485.53125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5766917293233083,
      "calib/avg_num_step_conf": 7.5859375,
      "calib/ece": 0.43225296442687744,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.009247493734335852,
      "calib/mean_conf": 0.04205533596837945,
      "calib/mu_c": 0.046916666666666676,
      "calib/mu_w": 0.037669172932330824,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.028723241563493374,
      "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": 2696.0,
      "completions/max_terminated_length": 2696.0,
      "completions/mean_length": 466.9375,
      "completions/mean_terminated_length": 466.9375,
      "completions/min_length": 131.0,
      "completions/min_terminated_length": 131.0,
      "epoch": 0.08853333333333334,
      "grad_norm": 0.005869552958756685,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.1236,
      "num_tokens": 16736488.0,
      "reward": 1.0050857067108154,
      "reward_std": 0.06311136484146118,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.5570465326309204,
      "rewards/format_reward_step": 0.984375,
      "step": 83
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4285242296755314,
      "aux_distill/mean_u": 0.26300995633903906,
      "aux_distill/n_active_tok": 421.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5220199092741934,
      "calib/avg_num_step_conf": 6.58203125,
      "calib/ece": 0.44035714285714284,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.003162802419354821,
      "calib/mean_conf": 0.062103174603174605,
      "calib/mu_c": 0.06370967741935483,
      "calib/mu_w": 0.060546875000000014,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.005198412698412698,
      "calib/std_conf": 0.07047979890304709,
      "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": 3005.0,
      "completions/max_terminated_length": 3005.0,
      "completions/mean_length": 384.6875,
      "completions/mean_terminated_length": 386.19610595703125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 133.0,
      "epoch": 0.0896,
      "grad_norm": 0.006847344804555178,
      "learning_rate": 3.2222222222222227e-06,
      "loss": 0.0825,
      "num_tokens": 16940888.0,
      "reward": 1.0128443241119385,
      "reward_std": 0.06419685482978821,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/final_brier_reward_step": 0.553032398223877,
      "rewards/format_reward_step": 0.984375,
      "step": 84
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4118081103079021,
      "aux_distill/mean_u": 0.2785398073769991,
      "aux_distill/n_active_tok": 480.65625,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.5516667798221196,
      "calib/avg_num_step_conf": 7.5546875,
      "calib/ece": 0.3671138211382114,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.0040650406504065045,
      "calib/gap": 0.006341910516667791,
      "calib/mean_conf": 0.062235772357723575,
      "calib/mu_c": 0.06592233009708738,
      "calib/mu_w": 0.05958041958041959,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.00532520325203252,
      "calib/std_conf": 0.0744811408771018,
      "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": 2598.0,
      "completions/max_terminated_length": 2598.0,
      "completions/mean_length": 458.8828125,
      "completions/mean_terminated_length": 460.682373046875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 137.0,
      "epoch": 0.09066666666666667,
      "grad_norm": 0.006029889453202486,
      "learning_rate": 3.1944444444444443e-06,
      "loss": 0.0813,
      "num_tokens": 17166186.0,
      "reward": 0.9829345345497131,
      "reward_std": 0.10405939817428589,
      "rewards/accuracy_reward_step": 0.40234375,
      "rewards/final_brier_reward_step": 0.602587878704071,
      "rewards/format_reward_step": 0.9609375,
      "step": 85
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.42284269724041224,
      "aux_distill/mean_u": 0.25688785203974734,
      "aux_distill/n_active_tok": 382.25,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5004802151363811,
      "calib/avg_num_step_conf": 5.98046875,
      "calib/ece": 0.3923107569721116,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.004747726981687794,
      "calib/mean_conf": 0.070398406374502,
      "calib/mu_c": 0.06780701754385966,
      "calib/mu_w": 0.07255474452554746,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.004262948207171314,
      "calib/std_conf": 0.0637087546243611,
      "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": 2434.0,
      "completions/max_terminated_length": 2434.0,
      "completions/mean_length": 442.07421875,
      "completions/mean_terminated_length": 443.807861328125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 95.0,
      "epoch": 0.09173333333333333,
      "grad_norm": 0.00629810057580471,
      "learning_rate": 3.1666666666666667e-06,
      "loss": 0.074,
      "num_tokens": 17384869.0,
      "reward": 1.002341628074646,
      "reward_std": 0.0852438285946846,
      "rewards/accuracy_reward_step": 0.4453125,
      "rewards/final_brier_reward_step": 0.5828081965446472,
      "rewards/format_reward_step": 0.9765625,
      "step": 86
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40320085268467665,
      "aux_distill/mean_u": 0.2343219956904517,
      "aux_distill/n_active_tok": 333.71875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5096732026143791,
      "calib/avg_num_step_conf": 5.1953125,
      "calib/ece": 0.5114682539682538,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": 0.010615686274509803,
      "calib/mean_conf": 0.08376984126984129,
      "calib/mu_c": 0.08806666666666665,
      "calib/mu_w": 0.07745098039215685,
      "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": 0.08096439181050301,
      "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": 2678.0,
      "completions/max_terminated_length": 2678.0,
      "completions/mean_length": 393.45703125,
      "completions/mean_terminated_length": 393.45703125,
      "completions/min_length": 118.0,
      "completions/min_terminated_length": 118.0,
      "epoch": 0.0928,
      "grad_norm": 0.00688812555745244,
      "learning_rate": 3.138888888888889e-06,
      "loss": 0.1493,
      "num_tokens": 17591090.0,
      "reward": 1.0292962789535522,
      "reward_std": 0.09010984003543854,
      "rewards/accuracy_reward_step": 0.5859375,
      "rewards/final_brier_reward_step": 0.4882800579071045,
      "rewards/format_reward_step": 0.984375,
      "step": 87
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43424535216763616,
      "aux_distill/mean_u": 0.2367017103593896,
      "aux_distill/n_active_tok": 406.75,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.48049410726665404,
      "calib/avg_num_step_conf": 6.3359375,
      "calib/ece": 0.43170634920634926,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": -0.013497825675931177,
      "calib/mean_conf": 0.09162698412698411,
      "calib/mu_c": 0.08503875968992249,
      "calib/mu_w": 0.09853658536585366,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.005714285714285713,
      "calib/std_conf": 0.08985510683449129,
      "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": 2639.0,
      "completions/max_terminated_length": 2639.0,
      "completions/mean_length": 462.8828125,
      "completions/mean_terminated_length": 462.8828125,
      "completions/min_length": 144.0,
      "completions/min_terminated_length": 144.0,
      "epoch": 0.09386666666666667,
      "grad_norm": 0.005841149017214775,
      "learning_rate": 3.1111111111111116e-06,
      "loss": 0.123,
      "num_tokens": 17819436.0,
      "reward": 1.017090082168579,
      "reward_std": 0.08964027464389801,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.5498050451278687,
      "rewards/format_reward_step": 0.98046875,
      "step": 88
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4475083462893963,
      "aux_distill/mean_u": 0.2913353557635836,
      "aux_distill/n_active_tok": 400.65625,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5194451646357272,
      "calib/avg_num_step_conf": 6.203125,
      "calib/ece": 0.3772690763052209,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.005897718434016069,
      "calib/mean_conf": 0.09710843373493976,
      "calib/mu_c": 0.10025862068965517,
      "calib/mu_w": 0.0943609022556391,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.9765625,
      "calib/pce": 0.004257028112449799,
      "calib/std_conf": 0.08784154391005794,
      "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": 2504.0,
      "completions/max_terminated_length": 2504.0,
      "completions/mean_length": 466.37109375,
      "completions/mean_terminated_length": 468.2000427246094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 97.0,
      "epoch": 0.09493333333333333,
      "grad_norm": 0.005623057018965483,
      "learning_rate": 3.0833333333333336e-06,
      "loss": 0.1245,
      "num_tokens": 18047715.0,
      "reward": 0.998233437538147,
      "reward_std": 0.1062500923871994,
      "rewards/accuracy_reward_step": 0.453125,
      "rewards/final_brier_reward_step": 0.5824043154716492,
      "rewards/format_reward_step": 0.9609375,
      "step": 89
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43033392354846,
      "aux_distill/mean_u": 0.2852346709679915,
      "aux_distill/n_active_tok": 404.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.572209723784527,
      "calib/avg_num_step_conf": 6.3203125,
      "calib/ece": 0.39778656126482215,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.007905138339920948,
      "calib/gap": 0.013782652168478945,
      "calib/mean_conf": 0.1165217391304348,
      "calib/mu_c": 0.12338582677165355,
      "calib/mu_w": 0.1096031746031746,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.00616600790513834,
      "calib/std_conf": 0.11560357384745286,
      "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": 2565.0,
      "completions/max_terminated_length": 2565.0,
      "completions/mean_length": 415.140625,
      "completions/mean_terminated_length": 415.140625,
      "completions/min_length": 132.0,
      "completions/min_terminated_length": 132.0,
      "epoch": 0.096,
      "grad_norm": 0.006214203778654337,
      "learning_rate": 3.055555555555556e-06,
      "loss": 0.0922,
      "num_tokens": 18257311.0,
      "reward": 1.0322825908660889,
      "reward_std": 0.0971316397190094,
      "rewards/accuracy_reward_step": 0.49609375,
      "rewards/final_brier_reward_step": 0.5840964913368225,
      "rewards/format_reward_step": 0.984375,
      "step": 90
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4307767665013671,
      "aux_distill/mean_u": 0.23621513864912877,
      "aux_distill/n_active_tok": 395.75,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5367861803268502,
      "calib/avg_num_step_conf": 6.18359375,
      "calib/ece": 0.4246850393700787,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.007874015748031496,
      "calib/gap": -0.004238488783943381,
      "calib/mean_conf": 0.12059055118110239,
      "calib/mu_c": 0.11857142857142855,
      "calib/mu_w": 0.12280991735537193,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.010826771653543307,
      "calib/std_conf": 0.13331939355593167,
      "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": 1697.0,
      "completions/max_terminated_length": 1697.0,
      "completions/mean_length": 429.49609375,
      "completions/mean_terminated_length": 431.180419921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 153.0,
      "epoch": 0.09706666666666666,
      "grad_norm": 0.005459485575556755,
      "learning_rate": 3.0277777777777776e-06,
      "loss": 0.0263,
      "num_tokens": 18474974.0,
      "reward": 1.027876853942871,
      "reward_std": 0.09287009388208389,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/final_brier_reward_step": 0.5557538866996765,
      "rewards/format_reward_step": 0.98046875,
      "step": 91
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41957229003310204,
      "aux_distill/mean_u": 0.22008513333165414,
      "aux_distill/n_active_tok": 414.375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4638769230769231,
      "calib/avg_num_step_conf": 6.47265625,
      "calib/ece": 0.3938039215686274,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.00784313725490196,
      "calib/gap": -0.021046153846153853,
      "calib/mean_conf": 0.14447058823529413,
      "calib/mu_c": 0.13415384615384615,
      "calib/mu_w": 0.1552,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.01423529411764706,
      "calib/std_conf": 0.14214468642916672,
      "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": 1868.0,
      "completions/max_terminated_length": 1868.0,
      "completions/mean_length": 435.60546875,
      "completions/mean_terminated_length": 435.60546875,
      "completions/min_length": 144.0,
      "completions/min_terminated_length": 144.0,
      "epoch": 0.09813333333333334,
      "grad_norm": 0.006024438422173262,
      "learning_rate": 3e-06,
      "loss": 0.0915,
      "num_tokens": 18693209.0,
      "reward": 1.0437605381011963,
      "reward_std": 0.09087259322404861,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.5836148262023926,
      "rewards/format_reward_step": 0.99609375,
      "step": 92
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4368822705000639,
      "aux_distill/mean_u": 0.2506925324287287,
      "aux_distill/n_active_tok": 436.09375,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.4765416401780038,
      "calib/avg_num_step_conf": 6.8125,
      "calib/ece": 0.3680876494023905,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.027888446215139442,
      "calib/gap": -0.017501589319771166,
      "calib/mean_conf": 0.16848605577689246,
      "calib/mu_c": 0.15942148760330577,
      "calib/mu_w": 0.17692307692307693,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.027250996015936255,
      "calib/std_conf": 0.18404296117631527,
      "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": 2461.0,
      "completions/max_terminated_length": 2461.0,
      "completions/mean_length": 466.32421875,
      "completions/mean_terminated_length": 466.32421875,
      "completions/min_length": 115.0,
      "completions/min_terminated_length": 115.0,
      "epoch": 0.0992,
      "grad_norm": 0.005379728972911835,
      "learning_rate": 2.9722222222222225e-06,
      "loss": 0.0783,
      "num_tokens": 18918364.0,
      "reward": 1.0252985954284668,
      "reward_std": 0.13064992427825928,
      "rewards/accuracy_reward_step": 0.47265625,
      "rewards/final_brier_reward_step": 0.5974722504615784,
      "rewards/format_reward_step": 0.98046875,
      "step": 93
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3870116211473942,
      "aux_distill/mean_u": 0.1967927027645488,
      "aux_distill/n_active_tok": 396.5,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5232878467550772,
      "calib/avg_num_step_conf": 6.1953125,
      "calib/ece": 0.359,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.044,
      "calib/gap": 0.02757063232750337,
      "calib/mean_conf": 0.19340000000000002,
      "calib/mu_c": 0.20674418604651165,
      "calib/mu_w": 0.17917355371900828,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.018200000000000004,
      "calib/std_conf": 0.21385799026456784,
      "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": 2775.0,
      "completions/max_terminated_length": 2775.0,
      "completions/mean_length": 412.4296875,
      "completions/mean_terminated_length": 414.0470886230469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.10026666666666667,
      "grad_norm": 0.006159038282930851,
      "learning_rate": 2.944444444444445e-06,
      "loss": 0.0775,
      "num_tokens": 19132626.0,
      "reward": 1.038222074508667,
      "reward_std": 0.15837010741233826,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.595975399017334,
      "rewards/format_reward_step": 0.97265625,
      "step": 94
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4291246635839343,
      "aux_distill/mean_u": 0.2764642272026086,
      "aux_distill/n_active_tok": 399.0,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5379575379575381,
      "calib/avg_num_step_conf": 6.2421875,
      "calib/ece": 0.4132228346456693,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.027559055118110236,
      "calib/gap": 0.027195577395577375,
      "calib/mean_conf": 0.1688244094488189,
      "calib/mu_c": 0.1807090909090909,
      "calib/mu_w": 0.15351351351351353,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.009527559055118111,
      "calib/std_conf": 0.17708300167695082,
      "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": 0.0,
      "completions/max_length": 2584.0,
      "completions/max_terminated_length": 2584.0,
      "completions/mean_length": 466.75390625,
      "completions/mean_terminated_length": 466.75390625,
      "completions/min_length": 88.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.10133333333333333,
      "grad_norm": 0.006041058339178562,
      "learning_rate": 2.916666666666667e-06,
      "loss": 0.1034,
      "num_tokens": 19358243.0,
      "reward": 1.054192066192627,
      "reward_std": 0.1343843936920166,
      "rewards/accuracy_reward_step": 0.55859375,
      "rewards/final_brier_reward_step": 0.5693216323852539,
      "rewards/format_reward_step": 0.98046875,
      "step": 95
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41332088224589825,
      "aux_distill/mean_u": 0.24489968288977096,
      "aux_distill/n_active_tok": 421.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5761417859577369,
      "calib/avg_num_step_conf": 6.6875,
      "calib/ece": 0.4851758893280632,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.02771336059986365,
      "calib/mean_conf": 0.16707707509881423,
      "calib/mu_c": 0.17693558282208588,
      "calib/mu_w": 0.14922222222222223,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.00399209486166008,
      "calib/std_conf": 0.14269122454238534,
      "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": 1696.0,
      "completions/max_terminated_length": 1696.0,
      "completions/mean_length": 417.02734375,
      "completions/mean_terminated_length": 420.31103515625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 108.0,
      "epoch": 0.1024,
      "grad_norm": 0.0056716748513281345,
      "learning_rate": 2.888888888888889e-06,
      "loss": 0.0577,
      "num_tokens": 19570818.0,
      "reward": 1.0770845413208008,
      "reward_std": 0.12300469726324081,
      "rewards/accuracy_reward_step": 0.63671875,
      "rewards/final_brier_reward_step": 0.5291690826416016,
      "rewards/format_reward_step": 0.98828125,
      "step": 96
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.425160467159003,
      "aux_distill/mean_u": 0.2345694086333077,
      "aux_distill/n_active_tok": 426.78125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5501902801172873,
      "calib/avg_num_step_conf": 6.66796875,
      "calib/ece": 0.31923976377952756,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.015748031496062992,
      "calib/gap": 0.023136022209744866,
      "calib/mean_conf": 0.18965787401574805,
      "calib/mu_c": 0.20213675213675217,
      "calib/mu_w": 0.1790007299270073,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.024133858267716538,
      "calib/std_conf": 0.18904196800273304,
      "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": 1863.0,
      "completions/max_terminated_length": 1863.0,
      "completions/mean_length": 436.94921875,
      "completions/mean_terminated_length": 436.94921875,
      "completions/min_length": 130.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.10346666666666667,
      "grad_norm": 0.005866437219083309,
      "learning_rate": 2.861111111111111e-06,
      "loss": 0.0759,
      "num_tokens": 19787749.0,
      "reward": 1.0457253456115723,
      "reward_std": 0.13500013947486877,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/final_brier_reward_step": 0.6461380124092102,
      "rewards/format_reward_step": 0.98828125,
      "step": 97
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4224176909774542,
      "aux_distill/mean_u": 0.30284410290391983,
      "aux_distill/n_active_tok": 406.28125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5889187632033801,
      "calib/avg_num_step_conf": 6.34765625,
      "calib/ece": 0.31840120000000005,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.036,
      "calib/gap": 0.05760982651558799,
      "calib/mean_conf": 0.2200788,
      "calib/mu_c": 0.2484228346456693,
      "calib/mu_w": 0.1908130081300813,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.9765625,
      "calib/pce": 0.015240000000000007,
      "calib/std_conf": 0.22231795732814746,
      "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": 2507.0,
      "completions/max_terminated_length": 2507.0,
      "completions/mean_length": 483.60546875,
      "completions/mean_terminated_length": 485.5019836425781,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 123.0,
      "epoch": 0.10453333333333334,
      "grad_norm": 0.005700993817299604,
      "learning_rate": 2.8333333333333335e-06,
      "loss": 0.0761,
      "num_tokens": 20017736.0,
      "reward": 1.0433740615844727,
      "reward_std": 0.18059629201889038,
      "rewards/accuracy_reward_step": 0.49609375,
      "rewards/final_brier_reward_step": 0.6258106231689453,
      "rewards/format_reward_step": 0.96484375,
      "step": 98
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40645847423002124,
      "aux_distill/mean_u": 0.23056967960458863,
      "aux_distill/n_active_tok": 440.0,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5040017785682525,
      "calib/avg_num_step_conf": 6.9453125,
      "calib/ece": 0.20314741035856573,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.03187250996015936,
      "calib/gap": -0.0025900400177856775,
      "calib/mean_conf": 0.2260159362549801,
      "calib/mu_c": 0.2242307692307692,
      "calib/mu_w": 0.22682080924855488,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.05920318725099602,
      "calib/std_conf": 0.21963403041367904,
      "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": 0.0078125,
      "completions/max_length": 2446.0,
      "completions/max_terminated_length": 2446.0,
      "completions/mean_length": 502.109375,
      "completions/mean_terminated_length": 506.06298828125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 124.0,
      "epoch": 0.1056,
      "grad_norm": 0.005459653213620186,
      "learning_rate": 2.805555555555556e-06,
      "loss": 0.0616,
      "num_tokens": 20252076.0,
      "reward": 0.9931430816650391,
      "reward_std": 0.16258558630943298,
      "rewards/accuracy_reward_step": 0.3046875,
      "rewards/final_brier_reward_step": 0.7089425921440125,
      "rewards/format_reward_step": 0.97265625,
      "step": 99
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4357182774692774,
      "aux_distill/mean_u": 0.28317105445502583,
      "aux_distill/n_active_tok": 429.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5066358024691358,
      "calib/avg_num_step_conf": 6.76953125,
      "calib/ece": 0.2976078431372549,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.01568627450980392,
      "calib/gap": 0.011833333333333335,
      "calib/mean_conf": 0.2115686274509804,
      "calib/mu_c": 0.21783333333333335,
      "calib/mu_w": 0.20600000000000002,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.019294117647058823,
      "calib/std_conf": 0.19430895702583822,
      "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": 1653.0,
      "completions/max_terminated_length": 1653.0,
      "completions/mean_length": 476.0078125,
      "completions/mean_terminated_length": 477.8745422363281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 144.0,
      "epoch": 0.10666666666666667,
      "grad_norm": 0.005962701514363289,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 0.0681,
      "num_tokens": 20481342.0,
      "reward": 1.0551525354385376,
      "reward_std": 0.1266723871231079,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.6493675708770752,
      "rewards/format_reward_step": 0.9921875,
      "step": 100
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41964778350666165,
      "aux_distill/mean_u": 0.2572269155690119,
      "aux_distill/n_active_tok": 507.96875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5264010366051183,
      "calib/avg_num_step_conf": 7.9375,
      "calib/ece": 0.25271706349206347,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.03571428571428571,
      "calib/gap": 0.024962040816326525,
      "calib/mean_conf": 0.22148928571428572,
      "calib/mu_c": 0.23605047619047623,
      "calib/mu_w": 0.2110884353741497,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.028769841269841265,
      "calib/std_conf": 0.20629327983683568,
      "calib/step_conf_rate": 0.98046875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2700.0,
      "completions/max_terminated_length": 2700.0,
      "completions/mean_length": 539.640625,
      "completions/mean_terminated_length": 539.640625,
      "completions/min_length": 132.0,
      "completions/min_terminated_length": 132.0,
      "epoch": 0.10773333333333333,
      "grad_norm": 0.00484135327860713,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.1312,
      "num_tokens": 20726482.0,
      "reward": 1.0217359066009521,
      "reward_std": 0.17179188132286072,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.6645654439926147,
      "rewards/format_reward_step": 0.96875,
      "step": 101
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.412255329079926,
      "aux_distill/mean_u": 0.2408152103888565,
      "aux_distill/n_active_tok": 433.21875,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5201820464532328,
      "calib/avg_num_step_conf": 6.72265625,
      "calib/ece": 0.335296442687747,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.043478260869565216,
      "calib/gap": 0.009948524795982377,
      "calib/mean_conf": 0.25395256916996045,
      "calib/mu_c": 0.2585925925925926,
      "calib/mu_w": 0.24864406779661022,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.027826086956521737,
      "calib/std_conf": 0.2319248160071721,
      "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": 2744.0,
      "completions/max_terminated_length": 2744.0,
      "completions/mean_length": 433.71875,
      "completions/mean_terminated_length": 433.71875,
      "completions/min_length": 69.0,
      "completions/min_terminated_length": 69.0,
      "epoch": 0.1088,
      "grad_norm": 0.005836244206875563,
      "learning_rate": 2.7222222222222224e-06,
      "loss": 0.0772,
      "num_tokens": 20944210.0,
      "reward": 1.0662009716033936,
      "reward_std": 0.15195344388484955,
      "rewards/accuracy_reward_step": 0.52734375,
      "rewards/final_brier_reward_step": 0.6167769432067871,
      "rewards/format_reward_step": 0.98828125,
      "step": 102
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.44190127309411764,
      "aux_distill/mean_u": 0.28445447424545817,
      "aux_distill/n_active_tok": 472.09375,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5412144702842377,
      "calib/avg_num_step_conf": 7.375,
      "calib/ece": 0.3301204819277109,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.05622489959839357,
      "calib/gap": 0.0429496124031008,
      "calib/mean_conf": 0.25108433734939756,
      "calib/mu_c": 0.27178294573643413,
      "calib/mu_w": 0.22883333333333333,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.031566265060240954,
      "calib/std_conf": 0.23890317667827385,
      "calib/step_conf_rate": 0.98046875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2630.0,
      "completions/max_terminated_length": 2630.0,
      "completions/mean_length": 561.2734375,
      "completions/mean_terminated_length": 561.2734375,
      "completions/min_length": 130.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.10986666666666667,
      "grad_norm": 0.004805284086614847,
      "learning_rate": 2.6944444444444444e-06,
      "loss": 0.091,
      "num_tokens": 21192448.0,
      "reward": 1.0435025691986084,
      "reward_std": 0.1977090835571289,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.6221612691879272,
      "rewards/format_reward_step": 0.9609375,
      "step": 103
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3976264260709286,
      "aux_distill/mean_u": 0.22159318477906012,
      "aux_distill/n_active_tok": 488.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.48812297734627846,
      "calib/avg_num_step_conf": 7.6328125,
      "calib/ece": 0.3151778656126482,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.05138339920948617,
      "calib/gap": -0.05741618122977346,
      "calib/mean_conf": 0.2656916996047431,
      "calib/mu_c": 0.23165048543689323,
      "calib/mu_w": 0.2890666666666667,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.08687747035573123,
      "calib/std_conf": 0.2508531210690347,
      "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": 2328.0,
      "completions/max_terminated_length": 2328.0,
      "completions/mean_length": 501.20703125,
      "completions/mean_terminated_length": 505.1535339355469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 134.0,
      "epoch": 0.11093333333333333,
      "grad_norm": 0.005391786806285381,
      "learning_rate": 2.666666666666667e-06,
      "loss": 0.0078,
      "num_tokens": 21427437.0,
      "reward": 1.0155071020126343,
      "reward_std": 0.16230294108390808,
      "rewards/accuracy_reward_step": 0.40234375,
      "rewards/final_brier_reward_step": 0.640389084815979,
      "rewards/format_reward_step": 0.98828125,
      "step": 104
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41611582646146417,
      "aux_distill/mean_u": 0.23099319336662405,
      "aux_distill/n_active_tok": 480.46875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5330399285623165,
      "calib/avg_num_step_conf": 7.41796875,
      "calib/ece": 0.30884462151394415,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.02390438247011952,
      "calib/gap": 0.03201811455542802,
      "calib/mean_conf": 0.27059760956175305,
      "calib/mu_c": 0.28552238805970154,
      "calib/mu_w": 0.2535042735042735,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.022788844621513947,
      "calib/std_conf": 0.20847396924042944,
      "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": 2963.0,
      "completions/max_terminated_length": 2963.0,
      "completions/mean_length": 524.15625,
      "completions/mean_terminated_length": 524.15625,
      "completions/min_length": 169.0,
      "completions/min_terminated_length": 169.0,
      "epoch": 0.112,
      "grad_norm": 0.005053746979683638,
      "learning_rate": 2.6388888888888893e-06,
      "loss": 0.0846,
      "num_tokens": 21667381.0,
      "reward": 1.0659711360931396,
      "reward_std": 0.19321054220199585,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/final_brier_reward_step": 0.6358484029769897,
      "rewards/format_reward_step": 0.97265625,
      "step": 105
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4285811148583889,
      "aux_distill/mean_u": 0.2643595084666108,
      "aux_distill/n_active_tok": 457.21875,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.49419411838949895,
      "calib/avg_num_step_conf": 7.140625,
      "calib/ece": 0.2772727272727272,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.02766798418972332,
      "calib/gap": -0.022519878833775098,
      "calib/mean_conf": 0.2731620553359684,
      "calib/mu_c": 0.2607894736842105,
      "calib/mu_w": 0.2833093525179856,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.04992094861660079,
      "calib/std_conf": 0.22256739757956648,
      "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": 2932.0,
      "completions/max_terminated_length": 2932.0,
      "completions/mean_length": 468.84375,
      "completions/mean_terminated_length": 470.682373046875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 159.0,
      "epoch": 0.11306666666666666,
      "grad_norm": 0.00532987155020237,
      "learning_rate": 2.6111111111111113e-06,
      "loss": 0.0613,
      "num_tokens": 21891989.0,
      "reward": 1.0404398441314697,
      "reward_std": 0.16531819105148315,
      "rewards/accuracy_reward_step": 0.4453125,
      "rewards/final_brier_reward_step": 0.6511921882629395,
      "rewards/format_reward_step": 0.984375,
      "step": 106
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4220479386858642,
      "aux_distill/mean_u": 0.2698333611404751,
      "aux_distill/n_active_tok": 492.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4615858163014036,
      "calib/avg_num_step_conf": 7.6484375,
      "calib/ece": 0.29654901960784313,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.023529411764705882,
      "calib/gap": -0.024373307067224903,
      "calib/mean_conf": 0.2836078431372549,
      "calib/mu_c": 0.2717557251908397,
      "calib/mu_w": 0.2961290322580646,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.033215686274509805,
      "calib/std_conf": 0.2114087276688539,
      "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": 1274.0,
      "completions/max_terminated_length": 1274.0,
      "completions/mean_length": 467.12109375,
      "completions/mean_terminated_length": 468.9529724121094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 168.0,
      "epoch": 0.11413333333333334,
      "grad_norm": 0.005111105740070343,
      "learning_rate": 2.5833333333333337e-06,
      "loss": 0.0419,
      "num_tokens": 22116188.0,
      "reward": 1.0728371143341064,
      "reward_std": 0.14746427536010742,
      "rewards/accuracy_reward_step": 0.51171875,
      "rewards/final_brier_reward_step": 0.6378617286682129,
      "rewards/format_reward_step": 0.99609375,
      "step": 107
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40239980118349195,
      "aux_distill/mean_u": 0.257174486721168,
      "aux_distill/n_active_tok": 518.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.45993031358885017,
      "calib/avg_num_step_conf": 8.09765625,
      "calib/ece": 0.4353721568627451,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.027450980392156862,
      "calib/gap": -0.037388528544626154,
      "calib/mean_conf": 0.27364745098039217,
      "calib/mu_c": 0.2603048780487805,
      "calib/mu_w": 0.29769340659340665,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.03294117647058824,
      "calib/std_conf": 0.2238614374995945,
      "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": 0.0,
      "completions/max_length": 2637.0,
      "completions/max_terminated_length": 2637.0,
      "completions/mean_length": 518.359375,
      "completions/mean_terminated_length": 518.359375,
      "completions/min_length": 116.0,
      "completions/min_terminated_length": 116.0,
      "epoch": 0.1152,
      "grad_norm": 0.004731772933155298,
      "learning_rate": 2.5555555555555557e-06,
      "loss": 0.0327,
      "num_tokens": 22352120.0,
      "reward": 1.093104362487793,
      "reward_std": 0.16300244629383087,
      "rewards/accuracy_reward_step": 0.640625,
      "rewards/final_brier_reward_step": 0.5612086057662964,
      "rewards/format_reward_step": 0.984375,
      "step": 108
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4148107245564461,
      "aux_distill/mean_u": 0.24032306112382526,
      "aux_distill/n_active_tok": 531.125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5454946439823566,
      "calib/avg_num_step_conf": 8.375,
      "calib/ece": 0.23936758893280635,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.019762845849802372,
      "calib/gap": 0.020057971014492693,
      "calib/mean_conf": 0.2885375494071146,
      "calib/mu_c": 0.2994782608695652,
      "calib/mu_w": 0.2794202898550725,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.036679841897233195,
      "calib/std_conf": 0.2033898386243882,
      "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": 2597.0,
      "completions/max_terminated_length": 2597.0,
      "completions/mean_length": 519.44140625,
      "completions/mean_terminated_length": 521.4784545898438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 145.0,
      "epoch": 0.11626666666666667,
      "grad_norm": 0.00585621502250433,
      "learning_rate": 2.5277777777777778e-06,
      "loss": 0.0727,
      "num_tokens": 22589697.0,
      "reward": 1.0482683181762695,
      "reward_std": 0.1747893989086151,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.6746617555618286,
      "rewards/format_reward_step": 0.97265625,
      "step": 109
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4172109765931964,
      "aux_distill/mean_u": 0.22872385472207696,
      "aux_distill/n_active_tok": 423.8125,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5273184759166067,
      "calib/avg_num_step_conf": 6.62890625,
      "calib/ece": 0.22584,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.036,
      "calib/gap": 0.006161688778511143,
      "calib/mean_conf": 0.30984,
      "calib/mu_c": 0.31336448598130834,
      "calib/mu_w": 0.3072027972027972,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.05384,
      "calib/std_conf": 0.23109646124508268,
      "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": 2802.0,
      "completions/max_terminated_length": 2802.0,
      "completions/mean_length": 470.04296875,
      "completions/mean_terminated_length": 471.88629150390625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 104.0,
      "epoch": 0.11733333333333333,
      "grad_norm": 0.004899916704744101,
      "learning_rate": 2.5e-06,
      "loss": 0.0564,
      "num_tokens": 22814948.0,
      "reward": 1.0319619178771973,
      "reward_std": 0.18836748600006104,
      "rewards/accuracy_reward_step": 0.41796875,
      "rewards/final_brier_reward_step": 0.6732988357543945,
      "rewards/format_reward_step": 0.97265625,
      "step": 110
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41732750553637743,
      "aux_distill/mean_u": 0.2677599661508874,
      "aux_distill/n_active_tok": 472.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5099623572170301,
      "calib/avg_num_step_conf": 7.3984375,
      "calib/ece": 0.2764541832669323,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.04780876494023904,
      "calib/gap": -0.016223390446521324,
      "calib/mean_conf": 0.30968127490039843,
      "calib/mu_c": 0.30037383177570093,
      "calib/mu_w": 0.31659722222222225,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.07992031872509961,
      "calib/std_conf": 0.22857408766908183,
      "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": 2010.0,
      "completions/max_terminated_length": 2010.0,
      "completions/mean_length": 482.55078125,
      "completions/mean_terminated_length": 484.44317626953125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 128.0,
      "epoch": 0.1184,
      "grad_norm": 0.005739626474678516,
      "learning_rate": 2.4722222222222226e-06,
      "loss": 0.0478,
      "num_tokens": 23045889.0,
      "reward": 1.0353412628173828,
      "reward_std": 0.1884884238243103,
      "rewards/accuracy_reward_step": 0.421875,
      "rewards/final_brier_reward_step": 0.6683386564254761,
      "rewards/format_reward_step": 0.98046875,
      "step": 111
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3907640604302287,
      "aux_distill/mean_u": 0.27975034768577645,
      "aux_distill/n_active_tok": 493.25,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5553652230122819,
      "calib/avg_num_step_conf": 7.75390625,
      "calib/ece": 0.235140562248996,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.020080321285140562,
      "calib/gap": 0.04242921784098258,
      "calib/mean_conf": 0.27473895582329316,
      "calib/mu_c": 0.29689075630252104,
      "calib/mu_w": 0.25446153846153846,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.015983935742971887,
      "calib/std_conf": 0.19097307398810948,
      "calib/step_conf_rate": 0.98046875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2538.0,
      "completions/max_terminated_length": 2538.0,
      "completions/mean_length": 539.71875,
      "completions/mean_terminated_length": 543.968505859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.11946666666666667,
      "grad_norm": 0.005129808560013771,
      "learning_rate": 2.4444444444444447e-06,
      "loss": 0.0865,
      "num_tokens": 23291977.0,
      "reward": 1.0484795570373535,
      "reward_std": 0.16111013293266296,
      "rewards/accuracy_reward_step": 0.46484375,
      "rewards/final_brier_reward_step": 0.6672718524932861,
      "rewards/format_reward_step": 0.96484375,
      "step": 112
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4189460719935596,
      "aux_distill/mean_u": 0.27368923599010614,
      "aux_distill/n_active_tok": 495.3125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5023419203747073,
      "calib/avg_num_step_conf": 7.71484375,
      "calib/ece": 0.2673333333333334,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.01568627450980392,
      "calib/gap": -0.0001429804018241998,
      "calib/mean_conf": 0.3007450980392157,
      "calib/mu_c": 0.3006766917293233,
      "calib/mu_w": 0.3008196721311475,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.023254901960784318,
      "calib/std_conf": 0.18694193741952836,
      "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": 2569.0,
      "completions/max_terminated_length": 2569.0,
      "completions/mean_length": 469.48046875,
      "completions/mean_terminated_length": 469.48046875,
      "completions/min_length": 148.0,
      "completions/min_terminated_length": 148.0,
      "epoch": 0.12053333333333334,
      "grad_norm": 0.00572775537148118,
      "learning_rate": 2.4166666666666667e-06,
      "loss": 0.0455,
      "num_tokens": 23517364.0,
      "reward": 1.089852213859558,
      "reward_std": 0.14501136541366577,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/final_brier_reward_step": 0.6640793085098267,
      "rewards/format_reward_step": 0.99609375,
      "step": 113
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4068376966752112,
      "aux_distill/mean_u": 0.24630880528646237,
      "aux_distill/n_active_tok": 466.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5650943396226416,
      "calib/avg_num_step_conf": 7.3515625,
      "calib/ece": 0.3013671875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.01953125,
      "calib/gap": 0.03843144654088049,
      "calib/mean_conf": 0.31082031250000003,
      "calib/mu_c": 0.3267333333333333,
      "calib/mu_w": 0.28830188679245283,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.013125000000000001,
      "calib/std_conf": 0.19955544082385313,
      "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": 1676.0,
      "completions/max_terminated_length": 1676.0,
      "completions/mean_length": 435.13671875,
      "completions/mean_terminated_length": 436.8431701660156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 137.0,
      "epoch": 0.1216,
      "grad_norm": 0.005413229577243328,
      "learning_rate": 2.388888888888889e-06,
      "loss": 0.0163,
      "num_tokens": 23733783.0,
      "reward": 1.1232295036315918,
      "reward_std": 0.13653072714805603,
      "rewards/accuracy_reward_step": 0.5859375,
      "rewards/final_brier_reward_step": 0.6605215072631836,
      "rewards/format_reward_step": 1.0,
      "step": 114
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3964782766997814,
      "aux_distill/mean_u": 0.20929134778443492,
      "aux_distill/n_active_tok": 471.625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.477160797833046,
      "calib/avg_num_step_conf": 7.4375,
      "calib/ece": 0.27051176470588234,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.047058823529411764,
      "calib/gap": -0.03655121891159807,
      "calib/mean_conf": 0.3403901960784314,
      "calib/mu_c": 0.3216129032258065,
      "calib/mu_w": 0.35816412213740456,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.06231372549019608,
      "calib/std_conf": 0.23257696533802288,
      "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": 1745.0,
      "completions/max_terminated_length": 1745.0,
      "completions/mean_length": 431.984375,
      "completions/mean_terminated_length": 433.678466796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 143.0,
      "epoch": 0.12266666666666666,
      "grad_norm": 0.005717757623642683,
      "learning_rate": 2.361111111111111e-06,
      "loss": -0.0018,
      "num_tokens": 23949635.0,
      "reward": 1.0635097026824951,
      "reward_std": 0.17919078469276428,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.6504567861557007,
      "rewards/format_reward_step": 0.9921875,
      "step": 115
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39956394489854574,
      "aux_distill/mean_u": 0.23843049430854713,
      "aux_distill/n_active_tok": 543.0625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5721250000000001,
      "calib/avg_num_step_conf": 8.51171875,
      "calib/ece": 0.255098814229249,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.02766798418972332,
      "calib/gap": 0.044694375000000064,
      "calib/mean_conf": 0.2843478260869565,
      "calib/mu_c": 0.30696,
      "calib/mu_w": 0.26226562499999995,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.02268774703557313,
      "calib/std_conf": 0.19661580874948692,
      "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": 1718.0,
      "completions/max_terminated_length": 1718.0,
      "completions/mean_length": 515.09765625,
      "completions/mean_terminated_length": 517.11767578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 95.0,
      "epoch": 0.12373333333333333,
      "grad_norm": 0.005036481190472841,
      "learning_rate": 2.3333333333333336e-06,
      "loss": 0.0429,
      "num_tokens": 24186020.0,
      "reward": 1.0783438682556152,
      "reward_std": 0.16028830409049988,
      "rewards/accuracy_reward_step": 0.4921875,
      "rewards/final_brier_reward_step": 0.6801249980926514,
      "rewards/format_reward_step": 0.984375,
      "step": 116
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3888081591576338,
      "aux_distill/mean_u": 0.20686839885684585,
      "aux_distill/n_active_tok": 542.09375,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.4939897698209719,
      "calib/avg_num_step_conf": 8.4375,
      "calib/ece": 0.20772908366533865,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.03187250996015936,
      "calib/gap": -0.0021815856777493425,
      "calib/mean_conf": 0.33752988047808763,
      "calib/mu_c": 0.3363478260869565,
      "calib/mu_w": 0.33852941176470586,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.04354581673306772,
      "calib/std_conf": 0.2156641601742187,
      "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": 2425.0,
      "completions/max_terminated_length": 2425.0,
      "completions/mean_length": 501.34375,
      "completions/mean_terminated_length": 501.34375,
      "completions/min_length": 173.0,
      "completions/min_terminated_length": 173.0,
      "epoch": 0.1248,
      "grad_norm": 0.0053986916318535805,
      "learning_rate": 2.305555555555556e-06,
      "loss": 0.0998,
      "num_tokens": 24420964.0,
      "reward": 1.052910566329956,
      "reward_std": 0.18577823042869568,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.6761335730552673,
      "rewards/format_reward_step": 0.98046875,
      "step": 117
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3707348401658237,
      "aux_distill/mean_u": 0.19880193935700557,
      "aux_distill/n_active_tok": 502.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.57578125,
      "calib/avg_num_step_conf": 7.84765625,
      "calib/ece": 0.21328063241106723,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.02766798418972332,
      "calib/gap": 0.05713500000000005,
      "calib/mean_conf": 0.3311462450592885,
      "calib/mu_c": 0.359375,
      "calib/mu_w": 0.30223999999999995,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.019249011857707513,
      "calib/std_conf": 0.20804726953127312,
      "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": 2293.0,
      "completions/max_terminated_length": 2293.0,
      "completions/mean_length": 466.609375,
      "completions/mean_terminated_length": 468.4392395019531,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 61.0,
      "epoch": 0.12586666666666665,
      "grad_norm": 0.005608038976788521,
      "learning_rate": 2.277777777777778e-06,
      "loss": -0.0022,
      "num_tokens": 24644424.0,
      "reward": 1.0923941135406494,
      "reward_std": 0.15614479780197144,
      "rewards/accuracy_reward_step": 0.5,
      "rewards/final_brier_reward_step": 0.6965069770812988,
      "rewards/format_reward_step": 0.98828125,
      "step": 118
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.42202150681987405,
      "aux_distill/mean_u": 0.285246961711442,
      "aux_distill/n_active_tok": 570.25,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5502906976744186,
      "calib/avg_num_step_conf": 9.1953125,
      "calib/ece": 0.23879518072289155,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.0321285140562249,
      "calib/gap": 0.02000968992248059,
      "calib/mean_conf": 0.3208835341365462,
      "calib/mu_c": 0.33125,
      "calib/mu_w": 0.3112403100775194,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.03887550200803214,
      "calib/std_conf": 0.20838454589009983,
      "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": 0.01171875,
      "completions/max_length": 2785.0,
      "completions/max_terminated_length": 2785.0,
      "completions/mean_length": 521.65234375,
      "completions/mean_terminated_length": 527.8379516601562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.12693333333333334,
      "grad_norm": 0.005744331516325474,
      "learning_rate": 2.25e-06,
      "loss": 0.0611,
      "num_tokens": 24883031.0,
      "reward": 1.0567359924316406,
      "reward_std": 0.2007881999015808,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.6720656156539917,
      "rewards/format_reward_step": 0.97265625,
      "step": 119
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40057709300890565,
      "aux_distill/mean_u": 0.23066852190770373,
      "aux_distill/n_active_tok": 513.03125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5443661077235772,
      "calib/avg_num_step_conf": 8.1015625,
      "calib/ece": 0.21824701195219126,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.0199203187250996,
      "calib/gap": 0.02910251524390245,
      "calib/mean_conf": 0.32581673306772907,
      "calib/mu_c": 0.340078125,
      "calib/mu_w": 0.31097560975609756,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.017051792828685252,
      "calib/std_conf": 0.18867864678303412,
      "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": 2296.0,
      "completions/max_terminated_length": 2296.0,
      "completions/mean_length": 468.4609375,
      "completions/mean_terminated_length": 470.2980651855469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 168.0,
      "epoch": 0.128,
      "grad_norm": 0.006349902134388685,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.0562,
      "num_tokens": 25109645.0,
      "reward": 1.0797877311706543,
      "reward_std": 0.17021773755550385,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.6791066527366638,
      "rewards/format_reward_step": 0.9765625,
      "step": 120
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39616987807676196,
      "aux_distill/mean_u": 0.23927794179594877,
      "aux_distill/n_active_tok": 600.28125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5476857452091424,
      "calib/avg_num_step_conf": 9.37890625,
      "calib/ece": 0.1884126984126984,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.03968253968253968,
      "calib/gap": 0.021945629337238093,
      "calib/mean_conf": 0.33214285714285713,
      "calib/mu_c": 0.3442477876106194,
      "calib/mu_w": 0.3223021582733813,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03607142857142857,
      "calib/std_conf": 0.2082070365478247,
      "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": 2723.0,
      "completions/max_terminated_length": 2723.0,
      "completions/mean_length": 545.3125,
      "completions/mean_terminated_length": 545.3125,
      "completions/min_length": 143.0,
      "completions/min_terminated_length": 143.0,
      "epoch": 0.12906666666666666,
      "grad_norm": 0.0052563464269042015,
      "learning_rate": 2.1944444444444445e-06,
      "loss": 0.0883,
      "num_tokens": 25354301.0,
      "reward": 1.0606942176818848,
      "reward_std": 0.19225087761878967,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.69560706615448,
      "rewards/format_reward_step": 0.984375,
      "step": 121
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4316580416634679,
      "aux_distill/mean_u": 0.27906441364180823,
      "aux_distill/n_active_tok": 497.375,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5692455242966752,
      "calib/avg_num_step_conf": 7.82421875,
      "calib/ece": 0.20302788844621517,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.027888446215139442,
      "calib/gap": 0.049783248081841425,
      "calib/mean_conf": 0.3712350597609562,
      "calib/mu_c": 0.3940441176470588,
      "calib/mu_w": 0.3442608695652174,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.016215139442231078,
      "calib/std_conf": 0.2060703969325257,
      "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": 2694.0,
      "completions/max_terminated_length": 2694.0,
      "completions/mean_length": 450.72265625,
      "completions/mean_terminated_length": 454.2716369628906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 151.0,
      "epoch": 0.13013333333333332,
      "grad_norm": 0.00618218956515193,
      "learning_rate": 2.166666666666667e-06,
      "loss": 0.0498,
      "num_tokens": 25577030.0,
      "reward": 1.097528338432312,
      "reward_std": 0.1832428276538849,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/final_brier_reward_step": 0.687244176864624,
      "rewards/format_reward_step": 0.9765625,
      "step": 122
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40500340377911925,
      "aux_distill/mean_u": 0.2747171806388661,
      "aux_distill/n_active_tok": 577.6875,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5615291571173924,
      "calib/avg_num_step_conf": 9.09765625,
      "calib/ece": 0.16545816733067728,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0398406374501992,
      "calib/gap": 0.052460529666412015,
      "calib/mean_conf": 0.36585657370517927,
      "calib/mu_c": 0.3934453781512605,
      "calib/mu_w": 0.3409848484848485,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.028605577689243038,
      "calib/std_conf": 0.2113953909180801,
      "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": 0.0078125,
      "completions/max_length": 3053.0,
      "completions/max_terminated_length": 3053.0,
      "completions/mean_length": 542.421875,
      "completions/mean_terminated_length": 546.6929321289062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 155.0,
      "epoch": 0.1312,
      "grad_norm": 0.0051978956907987595,
      "learning_rate": 2.138888888888889e-06,
      "loss": 0.0095,
      "num_tokens": 25821178.0,
      "reward": 1.0758334398269653,
      "reward_std": 0.17953169345855713,
      "rewards/accuracy_reward_step": 0.46484375,
      "rewards/final_brier_reward_step": 0.7063542604446411,
      "rewards/format_reward_step": 0.98046875,
      "step": 123
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39497670345008373,
      "aux_distill/mean_u": 0.20548810591655767,
      "aux_distill/n_active_tok": 525.03125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5154095701540956,
      "calib/avg_num_step_conf": 8.203125,
      "calib/ece": 0.232755905511811,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.027559055118110236,
      "calib/gap": -0.005556179424792529,
      "calib/mean_conf": 0.3860629921259843,
      "calib/mu_c": 0.38350364963503647,
      "calib/mu_w": 0.389059829059829,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0397244094488189,
      "calib/std_conf": 0.20238988443924427,
      "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": 2281.0,
      "completions/max_terminated_length": 2281.0,
      "completions/mean_length": 486.09765625,
      "completions/mean_terminated_length": 486.09765625,
      "completions/min_length": 172.0,
      "completions/min_terminated_length": 172.0,
      "epoch": 0.13226666666666667,
      "grad_norm": 0.0059423367492854595,
      "learning_rate": 2.1111111111111114e-06,
      "loss": 0.0639,
      "num_tokens": 26052435.0,
      "reward": 1.0993671417236328,
      "reward_std": 0.16392850875854492,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/final_brier_reward_step": 0.6752969026565552,
      "rewards/format_reward_step": 0.98828125,
      "step": 124
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36721099307760596,
      "aux_distill/mean_u": 0.194680674087944,
      "aux_distill/n_active_tok": 510.96875,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.572389632936508,
      "calib/avg_num_step_conf": 7.96875,
      "calib/ece": 0.20338937007874017,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.04330708661417323,
      "calib/gap": 0.04803913690476186,
      "calib/mean_conf": 0.39015393700787404,
      "calib/mu_c": 0.41398437499999996,
      "calib/mu_w": 0.3659452380952381,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.044803149606299233,
      "calib/std_conf": 0.22311851579552777,
      "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": 2315.0,
      "completions/max_terminated_length": 2315.0,
      "completions/mean_length": 494.0546875,
      "completions/mean_terminated_length": 494.0546875,
      "completions/min_length": 113.0,
      "completions/min_terminated_length": 113.0,
      "epoch": 0.13333333333333333,
      "grad_norm": 0.005903557408601046,
      "learning_rate": 2.0833333333333334e-06,
      "loss": 0.086,
      "num_tokens": 26283721.0,
      "reward": 1.0950615406036377,
      "reward_std": 0.19785240292549133,
      "rewards/accuracy_reward_step": 0.5,
      "rewards/final_brier_reward_step": 0.7018418312072754,
      "rewards/format_reward_step": 0.98828125,
      "step": 125
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3900165860541165,
      "aux_distill/mean_u": 0.24280415522821283,
      "aux_distill/n_active_tok": 535.25,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5641883116883115,
      "calib/avg_num_step_conf": 8.5078125,
      "calib/ece": 0.1463548,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.06,
      "calib/gap": 0.05018551948051958,
      "calib/mean_conf": 0.41180520000000004,
      "calib/mu_c": 0.43990909090909097,
      "calib/mu_w": 0.3897235714285714,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.059080000000000014,
      "calib/std_conf": 0.2240988442918883,
      "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": 2910.0,
      "completions/max_terminated_length": 2910.0,
      "completions/mean_length": 495.5390625,
      "completions/mean_terminated_length": 499.4409484863281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 127.0,
      "epoch": 0.1344,
      "grad_norm": 0.005371227860450745,
      "learning_rate": 2.0555555555555555e-06,
      "loss": 0.0593,
      "num_tokens": 26516043.0,
      "reward": 1.0582598447799683,
      "reward_std": 0.18028494715690613,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.7102696895599365,
      "rewards/format_reward_step": 0.9765625,
      "step": 126
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3885276550427079,
      "aux_distill/mean_u": 0.234251354106569,
      "aux_distill/n_active_tok": 564.40625,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.5701081612586036,
      "calib/avg_num_step_conf": 8.8046875,
      "calib/ece": 0.11475806451612904,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.05241935483870968,
      "calib/gap": 0.054674532940019505,
      "calib/mean_conf": 0.4484677419354839,
      "calib/mu_c": 0.47823008849557513,
      "calib/mu_w": 0.42355555555555563,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.053790322580645186,
      "calib/std_conf": 0.21490074361104142,
      "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": 2906.0,
      "completions/max_terminated_length": 2906.0,
      "completions/mean_length": 516.74609375,
      "completions/mean_terminated_length": 516.74609375,
      "completions/min_length": 152.0,
      "completions/min_terminated_length": 152.0,
      "epoch": 0.13546666666666668,
      "grad_norm": 0.006517878267914057,
      "learning_rate": 2.027777777777778e-06,
      "loss": 0.1305,
      "num_tokens": 26752002.0,
      "reward": 1.060055136680603,
      "reward_std": 0.21711894869804382,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.7099539041519165,
      "rewards/format_reward_step": 0.96875,
      "step": 127
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4126989454962313,
      "aux_distill/mean_u": 0.2660821506582588,
      "aux_distill/n_active_tok": 519.53125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.518909688843916,
      "calib/avg_num_step_conf": 8.1484375,
      "calib/ece": 0.18773809523809523,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.051587301587301584,
      "calib/gap": 0.010289653427776435,
      "calib/mean_conf": 0.4425793650793651,
      "calib/mu_c": 0.44805084745762713,
      "calib/mu_w": 0.4377611940298507,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.08103174603174601,
      "calib/std_conf": 0.22334420205161598,
      "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": 2783.0,
      "completions/max_terminated_length": 2783.0,
      "completions/mean_length": 516.5234375,
      "completions/mean_terminated_length": 520.590576171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 141.0,
      "epoch": 0.13653333333333334,
      "grad_norm": 0.0053933002054691315,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.052,
      "num_tokens": 26990896.0,
      "reward": 1.0699388980865479,
      "reward_std": 0.21272283792495728,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/final_brier_reward_step": 0.6945652365684509,
      "rewards/format_reward_step": 0.984375,
      "step": 128
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39896715339273214,
      "aux_distill/mean_u": 0.2534112287874238,
      "aux_distill/n_active_tok": 542.78125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5908012820512821,
      "calib/avg_num_step_conf": 8.484375,
      "calib/ece": 0.17408740157480312,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.03937007874015748,
      "calib/gap": 0.05774148717948724,
      "calib/mean_conf": 0.46150314960629923,
      "calib/mu_c": 0.48514533333333343,
      "calib/mu_w": 0.4274038461538462,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.02251968503937008,
      "calib/std_conf": 0.20543797402423522,
      "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": 2706.0,
      "completions/max_terminated_length": 2706.0,
      "completions/mean_length": 464.01171875,
      "completions/mean_terminated_length": 464.01171875,
      "completions/min_length": 137.0,
      "completions/min_terminated_length": 137.0,
      "epoch": 0.1376,
      "grad_norm": 0.0062581077218055725,
      "learning_rate": 1.9722222222222224e-06,
      "loss": 0.0497,
      "num_tokens": 27212067.0,
      "reward": 1.1498541831970215,
      "reward_std": 0.1686449944972992,
      "rewards/accuracy_reward_step": 0.5859375,
      "rewards/final_brier_reward_step": 0.7215834259986877,
      "rewards/format_reward_step": 0.9921875,
      "step": 129
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37909140018746257,
      "aux_distill/mean_u": 0.2216074330409185,
      "aux_distill/n_active_tok": 590.53125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5513638638638639,
      "calib/avg_num_step_conf": 9.234375,
      "calib/ece": 0.17133333333333334,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.050980392156862744,
      "calib/gap": 0.04842342342342343,
      "calib/mean_conf": 0.453921568627451,
      "calib/mu_c": 0.47500000000000003,
      "calib/mu_w": 0.4265765765765766,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.030274509803921566,
      "calib/std_conf": 0.21585469360308243,
      "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": 1955.0,
      "completions/max_terminated_length": 1955.0,
      "completions/mean_length": 490.41015625,
      "completions/mean_terminated_length": 492.3333740234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 119.0,
      "epoch": 0.13866666666666666,
      "grad_norm": 0.005617792718112469,
      "learning_rate": 1.944444444444445e-06,
      "loss": -0.0022,
      "num_tokens": 27442900.0,
      "reward": 1.137455701828003,
      "reward_std": 0.16030901670455933,
      "rewards/accuracy_reward_step": 0.5625,
      "rewards/final_brier_reward_step": 0.7163175344467163,
      "rewards/format_reward_step": 0.99609375,
      "step": 130
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4019861314445734,
      "aux_distill/mean_u": 0.2449434224736697,
      "aux_distill/n_active_tok": 500.78125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5476300197498354,
      "calib/avg_num_step_conf": 8.234375,
      "calib/ece": 0.1700395256916996,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.05928853754940711,
      "calib/gap": 0.04127518104015798,
      "calib/mean_conf": 0.4837944664031621,
      "calib/mu_c": 0.5090816326530612,
      "calib/mu_w": 0.4678064516129032,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.1332411067193676,
      "calib/std_conf": 0.21870222064088155,
      "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": 1762.0,
      "completions/max_terminated_length": 1762.0,
      "completions/mean_length": 432.30078125,
      "completions/mean_terminated_length": 435.7047119140625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 165.0,
      "epoch": 0.13973333333333332,
      "grad_norm": 0.0073953066021203995,
      "learning_rate": 1.916666666666667e-06,
      "loss": -0.0345,
      "num_tokens": 27659777.0,
      "reward": 1.0458250045776367,
      "reward_std": 0.1774011254310608,
      "rewards/accuracy_reward_step": 0.38671875,
      "rewards/final_brier_reward_step": 0.7166500091552734,
      "rewards/format_reward_step": 0.98828125,
      "step": 131
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3832178018055856,
      "aux_distill/mean_u": 0.29191975596897407,
      "aux_distill/n_active_tok": 606.6875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5628139043087497,
      "calib/avg_num_step_conf": 9.4765625,
      "calib/ece": 0.22648221343873517,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.02766798418972332,
      "calib/gap": 0.0352868094105207,
      "calib/mean_conf": 0.4308300395256917,
      "calib/mu_c": 0.4443589743589743,
      "calib/mu_w": 0.4090721649484536,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.020355731225296443,
      "calib/std_conf": 0.19875825676238895,
      "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": 2408.0,
      "completions/max_terminated_length": 2408.0,
      "completions/mean_length": 515.671875,
      "completions/mean_terminated_length": 515.671875,
      "completions/min_length": 125.0,
      "completions/min_terminated_length": 125.0,
      "epoch": 0.1408,
      "grad_norm": 0.005856058094650507,
      "learning_rate": 1.888888888888889e-06,
      "loss": 0.0474,
      "num_tokens": 27897381.0,
      "reward": 1.144197702407837,
      "reward_std": 0.1931305229663849,
      "rewards/accuracy_reward_step": 0.609375,
      "rewards/final_brier_reward_step": 0.6946452856063843,
      "rewards/format_reward_step": 0.984375,
      "step": 132
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4176169987767935,
      "aux_distill/mean_u": 0.2787832522007128,
      "aux_distill/n_active_tok": 694.03125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5484593644447509,
      "calib/avg_num_step_conf": 10.84375,
      "calib/ece": 0.15043650793650792,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.027777777777777776,
      "calib/gap": 0.026141173226718217,
      "calib/mean_conf": 0.4370238095238095,
      "calib/mu_c": 0.45393258426966293,
      "calib/mu_w": 0.4277914110429447,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.11714285714285715,
      "calib/std_conf": 0.20604618700460903,
      "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": 2962.0,
      "completions/max_terminated_length": 2962.0,
      "completions/mean_length": 605.890625,
      "completions/mean_terminated_length": 605.890625,
      "completions/min_length": 144.0,
      "completions/min_terminated_length": 144.0,
      "epoch": 0.14186666666666667,
      "grad_norm": 0.005014988128095865,
      "learning_rate": 1.8611111111111113e-06,
      "loss": 0.0875,
      "num_tokens": 28158833.0,
      "reward": 1.0234687328338623,
      "reward_std": 0.20991873741149902,
      "rewards/accuracy_reward_step": 0.34765625,
      "rewards/final_brier_reward_step": 0.7188124656677246,
      "rewards/format_reward_step": 0.98046875,
      "step": 133
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3869697102345526,
      "aux_distill/mean_u": 0.24196695079998468,
      "aux_distill/n_active_tok": 683.0625,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5955873015873017,
      "calib/avg_num_step_conf": 10.65625,
      "calib/ece": 0.17980000000000002,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.043824701195219126,
      "calib/gap": 0.061628901587301665,
      "calib/mean_conf": 0.42410438247011956,
      "calib/mu_c": 0.45504160000000005,
      "calib/mu_w": 0.3934126984126984,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.05294820717131475,
      "calib/std_conf": 0.22781403355626803,
      "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": 2479.0,
      "completions/max_terminated_length": 2479.0,
      "completions/mean_length": 592.046875,
      "completions/mean_terminated_length": 594.36865234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 172.0,
      "epoch": 0.14293333333333333,
      "grad_norm": 0.005612542387098074,
      "learning_rate": 1.8333333333333333e-06,
      "loss": 0.043,
      "num_tokens": 28419349.0,
      "reward": 1.089038372039795,
      "reward_std": 0.20408809185028076,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/final_brier_reward_step": 0.709326982498169,
      "rewards/format_reward_step": 0.98046875,
      "step": 134
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4061538176611066,
      "aux_distill/mean_u": 0.22896098657794023,
      "aux_distill/n_active_tok": 645.78125,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5387179487179486,
      "calib/avg_num_step_conf": 10.08984375,
      "calib/ece": 0.19088,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.032,
      "calib/gap": 0.036064102564102585,
      "calib/mean_conf": 0.46592,
      "calib/mu_c": 0.48323076923076924,
      "calib/mu_w": 0.44716666666666666,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0684,
      "calib/std_conf": 0.21351195189028646,
      "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": 2642.0,
      "completions/max_terminated_length": 2642.0,
      "completions/mean_length": 578.28125,
      "completions/mean_terminated_length": 578.28125,
      "completions/min_length": 147.0,
      "completions/min_terminated_length": 147.0,
      "epoch": 0.144,
      "grad_norm": 0.0058111692778766155,
      "learning_rate": 1.8055555555555557e-06,
      "loss": 0.1116,
      "num_tokens": 28673269.0,
      "reward": 1.0936968326568604,
      "reward_std": 0.19715605676174164,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.7030187845230103,
      "rewards/format_reward_step": 0.9765625,
      "step": 135
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37998731853440404,
      "aux_distill/mean_u": 0.22645032493399606,
      "aux_distill/n_active_tok": 651.21875,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.630528691166989,
      "calib/avg_num_step_conf": 10.1640625,
      "calib/ece": 0.1035856573705179,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.03187250996015936,
      "calib/gap": 0.08312508059316565,
      "calib/mean_conf": 0.45848605577689244,
      "calib/mu_c": 0.5051818181818181,
      "calib/mu_w": 0.4220567375886524,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.06191235059760956,
      "calib/std_conf": 0.20005004056450834,
      "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": 2530.0,
      "completions/max_terminated_length": 2530.0,
      "completions/mean_length": 541.5,
      "completions/mean_terminated_length": 545.7637939453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 132.0,
      "epoch": 0.14506666666666668,
      "grad_norm": 0.005560385063290596,
      "learning_rate": 1.777777777777778e-06,
      "loss": 0.067,
      "num_tokens": 28920381.0,
      "reward": 1.074867844581604,
      "reward_std": 0.1932879090309143,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.7395796775817871,
      "rewards/format_reward_step": 0.98046875,
      "step": 136
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39566820114851,
      "aux_distill/mean_u": 0.25699592562100493,
      "aux_distill/n_active_tok": 660.40625,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5430187459599225,
      "calib/avg_num_step_conf": 10.34375,
      "calib/ece": 0.15425702811244976,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.028112449799196786,
      "calib/gap": 0.009484809308338615,
      "calib/mean_conf": 0.4516867469879518,
      "calib/mu_c": 0.45663865546218474,
      "calib/mu_w": 0.4471538461538461,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0640160642570281,
      "calib/std_conf": 0.18965506171679003,
      "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": 2625.0,
      "completions/max_terminated_length": 2625.0,
      "completions/mean_length": 533.89453125,
      "completions/mean_terminated_length": 535.98828125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 153.0,
      "epoch": 0.14613333333333334,
      "grad_norm": 0.00525377131998539,
      "learning_rate": 1.75e-06,
      "loss": 0.0355,
      "num_tokens": 29164042.0,
      "reward": 1.0650525093078613,
      "reward_std": 0.20569221675395966,
      "rewards/accuracy_reward_step": 0.46484375,
      "rewards/final_brier_reward_step": 0.6965113282203674,
      "rewards/format_reward_step": 0.96875,
      "step": 137
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3761101681739092,
      "aux_distill/mean_u": 0.22736046414342426,
      "aux_distill/n_active_tok": 639.875,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5541148107121914,
      "calib/avg_num_step_conf": 9.9921875,
      "calib/ece": 0.16638888888888892,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.023809523809523808,
      "calib/gap": 0.04157881019091664,
      "calib/mean_conf": 0.4409920634920635,
      "calib/mu_c": 0.457986577181208,
      "calib/mu_w": 0.41640776699029136,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.008055555555555552,
      "calib/std_conf": 0.19486108181690132,
      "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": 2415.0,
      "completions/max_terminated_length": 2415.0,
      "completions/mean_length": 558.171875,
      "completions/mean_terminated_length": 558.171875,
      "completions/min_length": 136.0,
      "completions/min_terminated_length": 136.0,
      "epoch": 0.1472,
      "grad_norm": 0.005661411210894585,
      "learning_rate": 1.7222222222222224e-06,
      "loss": 0.0969,
      "num_tokens": 29411270.0,
      "reward": 1.1295913457870483,
      "reward_std": 0.20206280052661896,
      "rewards/accuracy_reward_step": 0.58203125,
      "rewards/final_brier_reward_step": 0.7005890607833862,
      "rewards/format_reward_step": 0.9765625,
      "step": 138
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37806188873946667,
      "aux_distill/mean_u": 0.23127254913429085,
      "aux_distill/n_active_tok": 567.0625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5583129741545583,
      "calib/avg_num_step_conf": 8.84375,
      "calib/ece": 0.1582352941176471,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.03137254901960784,
      "calib/gap": 0.044456731387424564,
      "calib/mean_conf": 0.47031372549019607,
      "calib/mu_c": 0.4879220779220779,
      "calib/mu_w": 0.44346534653465336,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.01231372549019608,
      "calib/std_conf": 0.18664819483493628,
      "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": 2649.0,
      "completions/max_terminated_length": 2649.0,
      "completions/mean_length": 472.3671875,
      "completions/mean_terminated_length": 472.3671875,
      "completions/min_length": 97.0,
      "completions/min_terminated_length": 97.0,
      "epoch": 0.14826666666666666,
      "grad_norm": 0.006207121070474386,
      "learning_rate": 1.6944444444444446e-06,
      "loss": 0.0564,
      "num_tokens": 29635292.0,
      "reward": 1.162093162536621,
      "reward_std": 0.17753392457962036,
      "rewards/accuracy_reward_step": 0.6015625,
      "rewards/final_brier_reward_step": 0.7265300750732422,
      "rewards/format_reward_step": 0.99609375,
      "step": 139
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38487443793565035,
      "aux_distill/mean_u": 0.21633979470535,
      "aux_distill/n_active_tok": 583.28125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.503512645523886,
      "calib/avg_num_step_conf": 9.12109375,
      "calib/ece": 0.23739130434782607,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.043478260869565216,
      "calib/gap": -0.0036437842901111406,
      "calib/mean_conf": 0.4660079051383399,
      "calib/mu_c": 0.46465408805031444,
      "calib/mu_w": 0.4682978723404256,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.03747035573122529,
      "calib/std_conf": 0.19472761947733852,
      "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": 2532.0,
      "completions/max_terminated_length": 2532.0,
      "completions/mean_length": 525.58984375,
      "completions/mean_terminated_length": 527.6510009765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 134.0,
      "epoch": 0.14933333333333335,
      "grad_norm": 0.005536575336009264,
      "learning_rate": 1.6666666666666667e-06,
      "loss": 0.0398,
      "num_tokens": 29874859.0,
      "reward": 1.1472833156585693,
      "reward_std": 0.18610472977161407,
      "rewards/accuracy_reward_step": 0.62109375,
      "rewards/final_brier_reward_step": 0.689098060131073,
      "rewards/format_reward_step": 0.984375,
      "step": 140
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36641107965260744,
      "aux_distill/mean_u": 0.23129930774601012,
      "aux_distill/n_active_tok": 599.71875,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.606989247311828,
      "calib/avg_num_step_conf": 9.3671875,
      "calib/ece": 0.1953359683794466,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.02766798418972332,
      "calib/gap": 0.07142137096774198,
      "calib/mean_conf": 0.45968379446640323,
      "calib/mu_c": 0.4859375,
      "calib/mu_w": 0.41451612903225804,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.011304347826086952,
      "calib/std_conf": 0.1996239023403811,
      "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": 2476.0,
      "completions/max_terminated_length": 2476.0,
      "completions/mean_length": 574.19140625,
      "completions/mean_terminated_length": 576.4431762695312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 165.0,
      "epoch": 0.1504,
      "grad_norm": 0.005775410681962967,
      "learning_rate": 1.638888888888889e-06,
      "loss": 0.0661,
      "num_tokens": 30128948.0,
      "reward": 1.167884349822998,
      "reward_std": 0.19001102447509766,
      "rewards/accuracy_reward_step": 0.625,
      "rewards/final_brier_reward_step": 0.7224875092506409,
      "rewards/format_reward_step": 0.98828125,
      "step": 141
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38335499819368124,
      "aux_distill/mean_u": 0.23487243225391793,
      "aux_distill/n_active_tok": 622.34375,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5993613824192336,
      "calib/avg_num_step_conf": 9.7265625,
      "calib/ece": 0.10956521739130433,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.02766798418972332,
      "calib/gap": 0.06858126721763086,
      "calib/mean_conf": 0.4528063241106719,
      "calib/mu_c": 0.48560606060606065,
      "calib/mu_w": 0.4170247933884298,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.020316205533596848,
      "calib/std_conf": 0.19820446493578275,
      "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": 2296.0,
      "completions/max_terminated_length": 2296.0,
      "completions/mean_length": 571.28125,
      "completions/mean_terminated_length": 573.5216064453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 158.0,
      "epoch": 0.15146666666666667,
      "grad_norm": 0.005249022971838713,
      "learning_rate": 1.6111111111111113e-06,
      "loss": 0.0134,
      "num_tokens": 30380356.0,
      "reward": 1.109088659286499,
      "reward_std": 0.1934710443019867,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/final_brier_reward_step": 0.7220836281776428,
      "rewards/format_reward_step": 0.9765625,
      "step": 142
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37024654168635607,
      "aux_distill/mean_u": 0.2144742621975383,
      "aux_distill/n_active_tok": 635.3125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.6517263427109975,
      "calib/avg_num_step_conf": 9.90625,
      "calib/ece": 0.11812749003984062,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.03187250996015936,
      "calib/gap": 0.09728069053708438,
      "calib/mean_conf": 0.470796812749004,
      "calib/mu_c": 0.5153676470588235,
      "calib/mu_w": 0.41808695652173916,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.02354581673306773,
      "calib/std_conf": 0.19683391018067173,
      "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": 2570.0,
      "completions/max_terminated_length": 2570.0,
      "completions/mean_length": 594.14453125,
      "completions/mean_terminated_length": 594.14453125,
      "completions/min_length": 200.0,
      "completions/min_terminated_length": 200.0,
      "epoch": 0.15253333333333333,
      "grad_norm": 0.006905294954776764,
      "learning_rate": 1.5833333333333333e-06,
      "loss": 0.1369,
      "num_tokens": 30639793.0,
      "reward": 1.1266040802001953,
      "reward_std": 0.19484950602054596,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/final_brier_reward_step": 0.7414894104003906,
      "rewards/format_reward_step": 0.98046875,
      "step": 143
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3681063880212605,
      "aux_distill/mean_u": 0.18628796554254137,
      "aux_distill/n_active_tok": 611.34375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5160638644509612,
      "calib/avg_num_step_conf": 9.55078125,
      "calib/ece": 0.17539370078740157,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.04330708661417323,
      "calib/gap": 0.018401433691756397,
      "calib/mean_conf": 0.49011811023622043,
      "calib/mu_c": 0.49729032258064526,
      "calib/mu_w": 0.47888888888888886,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.02763779527559055,
      "calib/std_conf": 0.2006396901477136,
      "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": 2564.0,
      "completions/max_terminated_length": 2564.0,
      "completions/mean_length": 563.98828125,
      "completions/mean_terminated_length": 563.98828125,
      "completions/min_length": 101.0,
      "completions/min_terminated_length": 101.0,
      "epoch": 0.1536,
      "grad_norm": 0.005503041669726372,
      "learning_rate": 1.5555555555555558e-06,
      "loss": 0.0812,
      "num_tokens": 30888302.0,
      "reward": 1.154140830039978,
      "reward_std": 0.18355412781238556,
      "rewards/accuracy_reward_step": 0.60546875,
      "rewards/final_brier_reward_step": 0.710625410079956,
      "rewards/format_reward_step": 0.9921875,
      "step": 144
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3513887752778828,
      "aux_distill/mean_u": 0.21696741285072296,
      "aux_distill/n_active_tok": 651.15625,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5948333333333333,
      "calib/avg_num_step_conf": 10.171875,
      "calib/ece": 0.14659999999999995,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.028,
      "calib/gap": 0.06279999999999986,
      "calib/mean_conf": 0.48148,
      "calib/mu_c": 0.5065999999999998,
      "calib/mu_w": 0.4438,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.014039999999999987,
      "calib/std_conf": 0.18749669223748988,
      "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": 2476.0,
      "completions/max_terminated_length": 2476.0,
      "completions/mean_length": 552.65234375,
      "completions/mean_terminated_length": 554.8196411132812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 38.0,
      "epoch": 0.15466666666666667,
      "grad_norm": 0.005455151200294495,
      "learning_rate": 1.527777777777778e-06,
      "loss": 0.0276,
      "num_tokens": 31132485.0,
      "reward": 1.143038034439087,
      "reward_std": 0.20722763240337372,
      "rewards/accuracy_reward_step": 0.5859375,
      "rewards/final_brier_reward_step": 0.7235761880874634,
      "rewards/format_reward_step": 0.9765625,
      "step": 145
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4044023398309946,
      "aux_distill/mean_u": 0.2234341881191083,
      "aux_distill/n_active_tok": 641.28125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5734043903057988,
      "calib/avg_num_step_conf": 10.00390625,
      "calib/ece": 0.11881422924901189,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.04743083003952569,
      "calib/gap": 0.04737152645603343,
      "calib/mean_conf": 0.5039525691699605,
      "calib/mu_c": 0.5305405405405406,
      "calib/mu_w": 0.48316901408450713,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.09201581027667986,
      "calib/std_conf": 0.2019455512148131,
      "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": 2770.0,
      "completions/max_terminated_length": 2770.0,
      "completions/mean_length": 571.01953125,
      "completions/mean_terminated_length": 571.01953125,
      "completions/min_length": 170.0,
      "completions/min_terminated_length": 170.0,
      "epoch": 0.15573333333333333,
      "grad_norm": 0.005942014046013355,
      "learning_rate": 1.5e-06,
      "loss": 0.0738,
      "num_tokens": 31385882.0,
      "reward": 1.0697505474090576,
      "reward_std": 0.21587388217449188,
      "rewards/accuracy_reward_step": 0.43359375,
      "rewards/final_brier_reward_step": 0.7215324640274048,
      "rewards/format_reward_step": 0.984375,
      "step": 146
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37331899954006076,
      "aux_distill/mean_u": 0.2227285418246145,
      "aux_distill/n_active_tok": 642.90625,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6430228810828231,
      "calib/avg_num_step_conf": 10.21875,
      "calib/ece": 0.13431230158730162,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.05952380952380952,
      "calib/gap": 0.10045484369964552,
      "calib/mean_conf": 0.5290210317460317,
      "calib/mu_c": 0.5868224299065421,
      "calib/mu_w": 0.4863675862068966,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.1193650793650794,
      "calib/std_conf": 0.21020622318098686,
      "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": 0.0078125,
      "completions/max_length": 3001.0,
      "completions/max_terminated_length": 3001.0,
      "completions/mean_length": 612.515625,
      "completions/mean_terminated_length": 617.3385620117188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 186.0,
      "epoch": 0.1568,
      "grad_norm": 0.004775840323418379,
      "learning_rate": 1.4722222222222225e-06,
      "loss": 0.0178,
      "num_tokens": 31646366.0,
      "reward": 1.0663350820541382,
      "reward_std": 0.2053927630186081,
      "rewards/accuracy_reward_step": 0.41796875,
      "rewards/final_brier_reward_step": 0.7342325448989868,
      "rewards/format_reward_step": 0.98046875,
      "step": 147
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3720237333327532,
      "aux_distill/mean_u": 0.23919748396660928,
      "aux_distill/n_active_tok": 592.03125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5598808341608739,
      "calib/avg_num_step_conf": 9.203125,
      "calib/ece": 0.170748031496063,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.051181102362204724,
      "calib/gap": 0.03987884806355502,
      "calib/mean_conf": 0.5388582677165354,
      "calib/mu_c": 0.5537735849056603,
      "calib/mu_w": 0.5138947368421053,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.041811023622047236,
      "calib/std_conf": 0.20432305069623957,
      "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": 2083.0,
      "completions/max_terminated_length": 2083.0,
      "completions/mean_length": 541.83984375,
      "completions/mean_terminated_length": 541.83984375,
      "completions/min_length": 120.0,
      "completions/min_terminated_length": 120.0,
      "epoch": 0.15786666666666666,
      "grad_norm": 0.0056559499353170395,
      "learning_rate": 1.4444444444444445e-06,
      "loss": 0.0557,
      "num_tokens": 31890189.0,
      "reward": 1.1733251810073853,
      "reward_std": 0.2118576318025589,
      "rewards/accuracy_reward_step": 0.625,
      "rewards/final_brier_reward_step": 0.7294628620147705,
      "rewards/format_reward_step": 0.9921875,
      "step": 148
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37415668834000826,
      "aux_distill/mean_u": 0.18526994675126796,
      "aux_distill/n_active_tok": 631.75,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6356880849987352,
      "calib/avg_num_step_conf": 9.87109375,
      "calib/ece": 0.0841269841269841,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.051587301587301584,
      "calib/gap": 0.08855805717176818,
      "calib/mean_conf": 0.5216666666666666,
      "calib/mu_c": 0.5631343283582089,
      "calib/mu_w": 0.4745762711864407,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.037023809523809514,
      "calib/std_conf": 0.19665355081365565,
      "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": 0.0078125,
      "completions/max_length": 3012.0,
      "completions/max_terminated_length": 3012.0,
      "completions/mean_length": 592.55078125,
      "completions/mean_terminated_length": 597.216552734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 96.0,
      "epoch": 0.15893333333333334,
      "grad_norm": 0.005320643540471792,
      "learning_rate": 1.4166666666666667e-06,
      "loss": 0.0261,
      "num_tokens": 32146338.0,
      "reward": 1.126164436340332,
      "reward_std": 0.19407425820827484,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/final_brier_reward_step": 0.7445164322853088,
      "rewards/format_reward_step": 0.984375,
      "step": 149
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4086106764152646,
      "aux_distill/mean_u": 0.24081592924632741,
      "aux_distill/n_active_tok": 558.3125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5217777777777778,
      "calib/avg_num_step_conf": 8.80078125,
      "calib/ece": 0.16298039215686275,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.058823529411764705,
      "calib/gap": 0.01598095238095243,
      "calib/mean_conf": 0.5743529411764706,
      "calib/mu_c": 0.5809333333333334,
      "calib/mu_w": 0.564952380952381,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.07454901960784314,
      "calib/std_conf": 0.19403323179385817,
      "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": 2204.0,
      "completions/max_terminated_length": 2204.0,
      "completions/mean_length": 487.75,
      "completions/mean_terminated_length": 489.66278076171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 142.0,
      "epoch": 0.16,
      "grad_norm": 0.005738874897360802,
      "learning_rate": 1.3888888888888892e-06,
      "loss": 0.044,
      "num_tokens": 32376162.0,
      "reward": 1.1534371376037598,
      "reward_std": 0.18325306475162506,
      "rewards/accuracy_reward_step": 0.5859375,
      "rewards/final_brier_reward_step": 0.7248429656028748,
      "rewards/format_reward_step": 0.99609375,
      "step": 150
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38912961538881063,
      "aux_distill/mean_u": 0.23468656230850374,
      "aux_distill/n_active_tok": 642.5625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.564843847986956,
      "calib/avg_num_step_conf": 10.04296875,
      "calib/ece": 0.17541501976284588,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.07509881422924901,
      "calib/gap": 0.05584409883356345,
      "calib/mean_conf": 0.5452964426877469,
      "calib/mu_c": 0.574873949579832,
      "calib/mu_w": 0.5190298507462685,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.12517786561264824,
      "calib/std_conf": 0.23259069179202796,
      "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": 2558.0,
      "completions/max_terminated_length": 2558.0,
      "completions/mean_length": 588.3203125,
      "completions/mean_terminated_length": 588.3203125,
      "completions/min_length": 160.0,
      "completions/min_terminated_length": 160.0,
      "epoch": 0.16106666666666666,
      "grad_norm": 0.0050440216436982155,
      "learning_rate": 1.3611111111111112e-06,
      "loss": 0.033,
      "num_tokens": 32633796.0,
      "reward": 1.0779383182525635,
      "reward_std": 0.2225310355424881,
      "rewards/accuracy_reward_step": 0.46484375,
      "rewards/final_brier_reward_step": 0.7066578269004822,
      "rewards/format_reward_step": 0.984375,
      "step": 151
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3880243906751275,
      "aux_distill/mean_u": 0.19693858332008352,
      "aux_distill/n_active_tok": 622.71875,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.572785614436396,
      "calib/avg_num_step_conf": 9.7265625,
      "calib/ece": 0.1451442231075697,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.08366533864541832,
      "calib/gap": 0.04836035074342371,
      "calib/mean_conf": 0.6034709163346613,
      "calib/mu_c": 0.6269767441860465,
      "calib/mu_w": 0.5786163934426228,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.11733545816733064,
      "calib/std_conf": 0.2129893996734671,
      "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": 2883.0,
      "completions/max_terminated_length": 2883.0,
      "completions/mean_length": 587.45703125,
      "completions/mean_terminated_length": 587.45703125,
      "completions/min_length": 175.0,
      "completions/min_terminated_length": 175.0,
      "epoch": 0.16213333333333332,
      "grad_norm": 0.005218581296503544,
      "learning_rate": 1.3333333333333334e-06,
      "loss": 0.072,
      "num_tokens": 32889577.0,
      "reward": 1.0956348180770874,
      "reward_std": 0.2619733214378357,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.7068947553634644,
      "rewards/format_reward_step": 0.98046875,
      "step": 152
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4142996594309807,
      "aux_distill/mean_u": 0.29708617139507415,
      "aux_distill/n_active_tok": 627.0,
      "calib/answer_extract_rate": 0.953125,
      "calib/auroc": 0.5845899607519108,
      "calib/avg_num_step_conf": 9.96875,
      "calib/ece": 0.13467213114754095,
      "calib/final_conf_rate": 0.953125,
      "calib/format_rate": 0.953125,
      "calib/frac_conf_gt_0.9": 0.11885245901639344,
      "calib/gap": 0.06643117813124011,
      "calib/mean_conf": 0.6192622950819673,
      "calib/mu_c": 0.6473049645390071,
      "calib/mu_w": 0.580873786407767,
      "calib/nonempty_final_conf_rate": 0.953125,
      "calib/nonempty_reasoning_rate": 0.96875,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.08803278688524588,
      "calib/std_conf": 0.2170373089963381,
      "calib/step_conf_rate": 0.96875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2735.0,
      "completions/max_terminated_length": 2735.0,
      "completions/mean_length": 632.4375,
      "completions/mean_terminated_length": 637.4172973632812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 167.0,
      "epoch": 0.1632,
      "grad_norm": 0.005001327954232693,
      "learning_rate": 1.3055555555555556e-06,
      "loss": 0.0707,
      "num_tokens": 33158801.0,
      "reward": 1.104444980621338,
      "reward_std": 0.21987763047218323,
      "rewards/accuracy_reward_step": 0.55078125,
      "rewards/final_brier_reward_step": 0.7049835920333862,
      "rewards/format_reward_step": 0.953125,
      "step": 153
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3739353087730706,
      "aux_distill/mean_u": 0.22378105635695517,
      "aux_distill/n_active_tok": 589.40625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.6106679960119641,
      "calib/avg_num_step_conf": 9.3203125,
      "calib/ece": 0.19649606299212602,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.1062992125984252,
      "calib/gap": 0.07342597208374879,
      "calib/mean_conf": 0.6371259842519686,
      "calib/mu_c": 0.6764406779661017,
      "calib/mu_w": 0.603014705882353,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.18452755905511814,
      "calib/std_conf": 0.20850236472633155,
      "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": 1530.0,
      "completions/max_terminated_length": 1530.0,
      "completions/mean_length": 540.6484375,
      "completions/mean_terminated_length": 542.7686767578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 160.0,
      "epoch": 0.16426666666666667,
      "grad_norm": 0.005266482941806316,
      "learning_rate": 1.2777777777777779e-06,
      "loss": 0.0241,
      "num_tokens": 33401647.0,
      "reward": 1.081038475036621,
      "reward_std": 0.24052487313747406,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/final_brier_reward_step": 0.7089519500732422,
      "rewards/format_reward_step": 0.9921875,
      "step": 154
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3764974372461438,
      "aux_distill/mean_u": 0.19954199441576917,
      "aux_distill/n_active_tok": 555.21875,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.568870523415978,
      "calib/avg_num_step_conf": 8.7265625,
      "calib/ece": 0.19988379446640314,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.16600790513833993,
      "calib/gap": 0.055773691460055175,
      "calib/mean_conf": 0.6720577075098814,
      "calib/mu_c": 0.7011570247933884,
      "calib/mu_w": 0.6453833333333332,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.1968403162055336,
      "calib/std_conf": 0.20734349417086043,
      "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": 2756.0,
      "completions/max_terminated_length": 2756.0,
      "completions/mean_length": 504.30859375,
      "completions/mean_terminated_length": 506.28631591796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 183.0,
      "epoch": 0.16533333333333333,
      "grad_norm": 0.006387793924659491,
      "learning_rate": 1.25e-06,
      "loss": 0.0366,
      "num_tokens": 33637966.0,
      "reward": 1.0752594470977783,
      "reward_std": 0.22296142578125,
      "rewards/accuracy_reward_step": 0.47265625,
      "rewards/final_brier_reward_step": 0.6895813345909119,
      "rewards/format_reward_step": 0.98828125,
      "step": 155
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3770381882786751,
      "aux_distill/mean_u": 0.1989017791072949,
      "aux_distill/n_active_tok": 598.09375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6612547288776797,
      "calib/avg_num_step_conf": 9.36328125,
      "calib/ece": 0.1534845238095238,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.09523809523809523,
      "calib/gap": 0.12217446406052968,
      "calib/mean_conf": 0.6170710317460317,
      "calib/mu_c": 0.6800975409836066,
      "calib/mu_w": 0.557923076923077,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.97265625,
      "calib/pce": 0.1432142857142857,
      "calib/std_conf": 0.22878321806284102,
      "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": 2122.0,
      "completions/max_terminated_length": 2122.0,
      "completions/mean_length": 579.515625,
      "completions/mean_terminated_length": 581.7882690429688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 162.0,
      "epoch": 0.1664,
      "grad_norm": 0.005374233238399029,
      "learning_rate": 1.2222222222222223e-06,
      "loss": 0.0421,
      "num_tokens": 33891082.0,
      "reward": 1.0824484825134277,
      "reward_std": 0.2440761774778366,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.719584584236145,
      "rewards/format_reward_step": 0.96875,
      "step": 156
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3959145718254149,
      "aux_distill/mean_u": 0.2793759608083668,
      "aux_distill/n_active_tok": 662.40625,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.596949221949222,
      "calib/avg_num_step_conf": 10.5,
      "calib/ece": 0.14449392712550607,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.1214574898785425,
      "calib/gap": 0.07183660933660929,
      "calib/mean_conf": 0.6466801619433198,
      "calib/mu_c": 0.6754729729729729,
      "calib/mu_w": 0.6036363636363636,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.09599190283400812,
      "calib/std_conf": 0.21002870184418684,
      "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": 2984.0,
      "completions/max_terminated_length": 2984.0,
      "completions/mean_length": 587.05078125,
      "completions/mean_terminated_length": 589.3529663085938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 137.0,
      "epoch": 0.16746666666666668,
      "grad_norm": 0.005070784594863653,
      "learning_rate": 1.1944444444444446e-06,
      "loss": 0.1324,
      "num_tokens": 34145095.0,
      "reward": 1.1265051364898682,
      "reward_std": 0.24920831620693207,
      "rewards/accuracy_reward_step": 0.578125,
      "rewards/final_brier_reward_step": 0.7178539037704468,
      "rewards/format_reward_step": 0.95703125,
      "step": 157
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38080530194565654,
      "aux_distill/mean_u": 0.2144786379452595,
      "aux_distill/n_active_tok": 595.46875,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5020163831127914,
      "calib/avg_num_step_conf": 9.35546875,
      "calib/ece": 0.18814229249011852,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.17786561264822134,
      "calib/gap": 0.017797101449275377,
      "calib/mean_conf": 0.6803162055335968,
      "calib/mu_c": 0.6884057971014492,
      "calib/mu_w": 0.6706086956521738,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.16150197628458496,
      "calib/std_conf": 0.22398453065900517,
      "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": 1801.0,
      "completions/max_terminated_length": 1801.0,
      "completions/mean_length": 542.1484375,
      "completions/mean_terminated_length": 544.2745361328125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 142.0,
      "epoch": 0.16853333333333334,
      "grad_norm": 0.006110852584242821,
      "learning_rate": 1.1666666666666668e-06,
      "loss": 0.0714,
      "num_tokens": 34389125.0,
      "reward": 1.1058812141418457,
      "reward_std": 0.2672828137874603,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/final_brier_reward_step": 0.6844187378883362,
      "rewards/format_reward_step": 0.98828125,
      "step": 158
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36747160041704774,
      "aux_distill/mean_u": 0.18348273788386177,
      "aux_distill/n_active_tok": 580.09375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5950610221205188,
      "calib/avg_num_step_conf": 9.0625,
      "calib/ece": 0.174318253968254,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.14682539682539683,
      "calib/gap": 0.06862860411899308,
      "calib/mean_conf": 0.6717928571428572,
      "calib/mu_c": 0.7028391304347826,
      "calib/mu_w": 0.6342105263157896,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.1492460317460318,
      "calib/std_conf": 0.2146011178681263,
      "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": 2890.0,
      "completions/max_terminated_length": 2890.0,
      "completions/mean_length": 535.01953125,
      "completions/mean_terminated_length": 535.01953125,
      "completions/min_length": 170.0,
      "completions/min_terminated_length": 170.0,
      "epoch": 0.1696,
      "grad_norm": 0.005810961127281189,
      "learning_rate": 1.138888888888889e-06,
      "loss": 0.0909,
      "num_tokens": 34630874.0,
      "reward": 1.118455171585083,
      "reward_std": 0.22580690681934357,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/final_brier_reward_step": 0.713472843170166,
      "rewards/format_reward_step": 0.984375,
      "step": 159
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3862662399187684,
      "aux_distill/mean_u": 0.22606274958942982,
      "aux_distill/n_active_tok": 640.09375,
      "calib/answer_extract_rate": 0.95703125,
      "calib/auroc": 0.5912081339712919,
      "calib/avg_num_step_conf": 10.0,
      "calib/ece": 0.23280487804878053,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 0.16260162601626016,
      "calib/gap": 0.07370414673046244,
      "calib/mean_conf": 0.6848373983739838,
      "calib/mu_c": 0.7243859649122807,
      "calib/mu_w": 0.6506818181818183,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.22711382113821144,
      "calib/std_conf": 0.21200498053605563,
      "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": 2723.0,
      "completions/max_terminated_length": 2723.0,
      "completions/mean_length": 591.74609375,
      "completions/mean_terminated_length": 594.0667114257812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 155.0,
      "epoch": 0.17066666666666666,
      "grad_norm": 0.004924526903778315,
      "learning_rate": 1.111111111111111e-06,
      "loss": 0.0339,
      "num_tokens": 34887201.0,
      "reward": 1.0270812511444092,
      "reward_std": 0.30026695132255554,
      "rewards/accuracy_reward_step": 0.4453125,
      "rewards/final_brier_reward_step": 0.6596312522888184,
      "rewards/format_reward_step": 0.94921875,
      "step": 160
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3662711353972554,
      "aux_distill/mean_u": 0.20840195328305358,
      "aux_distill/n_active_tok": 550.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.575003621613791,
      "calib/avg_num_step_conf": 8.703125,
      "calib/ece": 0.1333333333333333,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.13333333333333333,
      "calib/gap": 0.048624511082138056,
      "calib/mean_conf": 0.6710588235294118,
      "calib/mu_c": 0.6859322033898305,
      "calib/mu_w": 0.6373076923076925,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.05513725490196077,
      "calib/std_conf": 0.20934270727401036,
      "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": 1806.0,
      "completions/max_terminated_length": 1806.0,
      "completions/mean_length": 519.50390625,
      "completions/mean_terminated_length": 521.5411987304688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 102.0,
      "epoch": 0.17173333333333332,
      "grad_norm": 0.005698539316654205,
      "learning_rate": 1.0833333333333335e-06,
      "loss": 0.0608,
      "num_tokens": 35124114.0,
      "reward": 1.2242445945739746,
      "reward_std": 0.2015002965927124,
      "rewards/accuracy_reward_step": 0.69140625,
      "rewards/final_brier_reward_step": 0.7609890699386597,
      "rewards/format_reward_step": 0.99609375,
      "step": 161
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3748155296780169,
      "aux_distill/mean_u": 0.1932356987573526,
      "aux_distill/n_active_tok": 531.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5960806428873537,
      "calib/avg_num_step_conf": 8.30859375,
      "calib/ece": 0.11890588235294117,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.15294117647058825,
      "calib/gap": 0.06314239390948828,
      "calib/mean_conf": 0.7045450980392156,
      "calib/mu_c": 0.7248497109826589,
      "calib/mu_w": 0.6617073170731707,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.07250980392156862,
      "calib/std_conf": 0.1852319503140515,
      "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": 2561.0,
      "completions/max_terminated_length": 2561.0,
      "completions/mean_length": 497.8046875,
      "completions/mean_terminated_length": 497.8046875,
      "completions/min_length": 151.0,
      "completions/min_terminated_length": 151.0,
      "epoch": 0.1728,
      "grad_norm": 0.005967023782432079,
      "learning_rate": 1.0555555555555557e-06,
      "loss": 0.0557,
      "num_tokens": 35355696.0,
      "reward": 1.2153667211532593,
      "reward_std": 0.2070452868938446,
      "rewards/accuracy_reward_step": 0.67578125,
      "rewards/final_brier_reward_step": 0.7666709423065186,
      "rewards/format_reward_step": 0.98828125,
      "step": 162
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3876973898150027,
      "aux_distill/mean_u": 0.2564975633289124,
      "aux_distill/n_active_tok": 627.0625,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.6148873783819806,
      "calib/avg_num_step_conf": 10.14453125,
      "calib/ece": 0.2239591836734694,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.15918367346938775,
      "calib/gap": 0.07605957616953207,
      "calib/mean_conf": 0.7071428571428572,
      "calib/mu_c": 0.7453278688524589,
      "calib/mu_w": 0.6692682926829269,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.21657142857142858,
      "calib/std_conf": 0.19477459499930388,
      "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": 2760.0,
      "completions/max_terminated_length": 2760.0,
      "completions/mean_length": 573.625,
      "completions/mean_terminated_length": 578.1417236328125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 147.0,
      "epoch": 0.17386666666666667,
      "grad_norm": 0.00539444712921977,
      "learning_rate": 1.0277777777777777e-06,
      "loss": 0.05,
      "num_tokens": 35607376.0,
      "reward": 1.0547908544540405,
      "reward_std": 0.2592874765396118,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.675987958908081,
      "rewards/format_reward_step": 0.95703125,
      "step": 163
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38780549727380276,
      "aux_distill/mean_u": 0.20125896250529246,
      "aux_distill/n_active_tok": 601.59375,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.6642457719585378,
      "calib/avg_num_step_conf": 9.66796875,
      "calib/ece": 0.09895800000000002,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.164,
      "calib/gap": 0.13103668848881622,
      "calib/mean_conf": 0.685602,
      "calib/mu_c": 0.7348717948717949,
      "calib/mu_w": 0.6038351063829787,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.08028000000000002,
      "calib/std_conf": 0.220712434167176,
      "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": 0.0078125,
      "completions/max_length": 2589.0,
      "completions/max_terminated_length": 2589.0,
      "completions/mean_length": 571.27734375,
      "completions/mean_terminated_length": 575.7755737304688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 147.0,
      "epoch": 0.17493333333333333,
      "grad_norm": 0.005605064332485199,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0258,
      "num_tokens": 35859759.0,
      "reward": 1.1609622240066528,
      "reward_std": 0.20984989404678345,
      "rewards/accuracy_reward_step": 0.609375,
      "rewards/final_brier_reward_step": 0.7477058172225952,
      "rewards/format_reward_step": 0.96484375,
      "step": 164
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41156348772346973,
      "aux_distill/mean_u": 0.2492245281580815,
      "aux_distill/n_active_tok": 574.5625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.568014027316353,
      "calib/avg_num_step_conf": 8.984375,
      "calib/ece": 0.1940694117647059,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.20784313725490197,
      "calib/gap": 0.06060826873385006,
      "calib/mean_conf": 0.693538431372549,
      "calib/mu_c": 0.723486046511628,
      "calib/mu_w": 0.6628777777777779,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.19086274509803924,
      "calib/std_conf": 0.21402834816245078,
      "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": 1605.0,
      "completions/max_terminated_length": 1605.0,
      "completions/mean_length": 554.39453125,
      "completions/mean_terminated_length": 556.5686645507812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 144.0,
      "epoch": 0.176,
      "grad_norm": 0.005580270662903786,
      "learning_rate": 9.722222222222224e-07,
      "loss": 0.0407,
      "num_tokens": 36107260.0,
      "reward": 1.100243091583252,
      "reward_std": 0.2129260003566742,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.6965799331665039,
      "rewards/format_reward_step": 0.99609375,
      "step": 165
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3936434802599251,
      "aux_distill/mean_u": 0.25069887011889913,
      "aux_distill/n_active_tok": 631.53125,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.6650324675324676,
      "calib/avg_num_step_conf": 9.91015625,
      "calib/ece": 0.1648344,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.24,
      "calib/gap": 0.13665610389610383,
      "calib/mean_conf": 0.7110856,
      "calib/mu_c": 0.7712142857142857,
      "calib/mu_w": 0.6345581818181819,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.15796000000000004,
      "calib/std_conf": 0.22916760092264352,
      "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": 2842.0,
      "completions/max_terminated_length": 2842.0,
      "completions/mean_length": 622.55859375,
      "completions/mean_terminated_length": 622.55859375,
      "completions/min_length": 187.0,
      "completions/min_terminated_length": 187.0,
      "epoch": 0.17706666666666668,
      "grad_norm": 0.004876760300248861,
      "learning_rate": 9.444444444444445e-07,
      "loss": 0.0771,
      "num_tokens": 36372819.0,
      "reward": 1.1162844896316528,
      "reward_std": 0.23633524775505066,
      "rewards/accuracy_reward_step": 0.546875,
      "rewards/final_brier_reward_step": 0.7208502292633057,
      "rewards/format_reward_step": 0.96484375,
      "step": 166
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3998649688437581,
      "aux_distill/mean_u": 0.266047704954752,
      "aux_distill/n_active_tok": 591.78125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5493181204428841,
      "calib/avg_num_step_conf": 9.24609375,
      "calib/ece": 0.16944031620553357,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.2924901185770751,
      "calib/gap": 0.04313105590062116,
      "calib/mean_conf": 0.7466861660079053,
      "calib/mu_c": 0.7623701863354038,
      "calib/mu_w": 0.7192391304347826,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.1398814229249012,
      "calib/std_conf": 0.21671540363402186,
      "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": 2852.0,
      "completions/max_terminated_length": 2852.0,
      "completions/mean_length": 530.41015625,
      "completions/mean_terminated_length": 530.41015625,
      "completions/min_length": 110.0,
      "completions/min_terminated_length": 110.0,
      "epoch": 0.17813333333333334,
      "grad_norm": 0.005584842525422573,
      "learning_rate": 9.166666666666666e-07,
      "loss": 0.0538,
      "num_tokens": 36614212.0,
      "reward": 1.1690298318862915,
      "reward_std": 0.2310115396976471,
      "rewards/accuracy_reward_step": 0.62890625,
      "rewards/final_brier_reward_step": 0.7208719849586487,
      "rewards/format_reward_step": 0.98828125,
      "step": 167
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3931912127882242,
      "aux_distill/mean_u": 0.24273921144412486,
      "aux_distill/n_active_tok": 600.0,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6109868421052631,
      "calib/avg_num_step_conf": 9.703125,
      "calib/ece": 0.14272896825396825,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.24206349206349206,
      "calib/gap": 0.08523226315789478,
      "calib/mean_conf": 0.7120329365079365,
      "calib/mu_c": 0.7458552631578947,
      "calib/mu_w": 0.660623,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.1257936507936508,
      "calib/std_conf": 0.22484954243370556,
      "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": 0.0078125,
      "completions/max_length": 2173.0,
      "completions/max_terminated_length": 2173.0,
      "completions/mean_length": 607.05859375,
      "completions/mean_terminated_length": 611.8385620117188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 131.0,
      "epoch": 0.1792,
      "grad_norm": 0.005158184561878443,
      "learning_rate": 8.88888888888889e-07,
      "loss": 0.0627,
      "num_tokens": 36874291.0,
      "reward": 1.1490778923034668,
      "reward_std": 0.2613195478916168,
      "rewards/accuracy_reward_step": 0.59375,
      "rewards/final_brier_reward_step": 0.723936915397644,
      "rewards/format_reward_step": 0.98046875,
      "step": 168
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3681354848667979,
      "aux_distill/mean_u": 0.20839590022768617,
      "aux_distill/n_active_tok": 556.25,
      "calib/answer_extract_rate": 0.96484375,
      "calib/auroc": 0.5548625083836353,
      "calib/avg_num_step_conf": 8.69921875,
      "calib/ece": 0.24605263157894747,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.30364372469635625,
      "calib/gap": 0.035121730382293936,
      "calib/mean_conf": 0.7590485829959515,
      "calib/mu_c": 0.7739788732394366,
      "calib/mu_w": 0.7388571428571427,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.2151012145748989,
      "calib/std_conf": 0.22427603705609578,
      "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": 0.01171875,
      "completions/max_length": 2834.0,
      "completions/max_terminated_length": 2834.0,
      "completions/mean_length": 566.71875,
      "completions/mean_terminated_length": 573.4387817382812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 134.0,
      "epoch": 0.18026666666666666,
      "grad_norm": 0.0056145042181015015,
      "learning_rate": 8.611111111111112e-07,
      "loss": 0.0222,
      "num_tokens": 37123555.0,
      "reward": 1.0919448137283325,
      "reward_std": 0.26493191719055176,
      "rewards/accuracy_reward_step": 0.5546875,
      "rewards/final_brier_reward_step": 0.664358377456665,
      "rewards/format_reward_step": 0.96484375,
      "step": 169
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3833466558717191,
      "aux_distill/mean_u": 0.22643499535792744,
      "aux_distill/n_active_tok": 613.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5134576948700866,
      "calib/avg_num_step_conf": 9.578125,
      "calib/ece": 0.2185335968379447,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.32806324110671936,
      "calib/gap": 0.018214856762158615,
      "calib/mean_conf": 0.7722173913043479,
      "calib/mu_c": 0.7790569620253165,
      "calib/mu_w": 0.7608421052631579,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.18312252964426884,
      "calib/std_conf": 0.2049528726652741,
      "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": 2661.0,
      "completions/max_terminated_length": 2661.0,
      "completions/mean_length": 587.5625,
      "completions/mean_terminated_length": 587.5625,
      "completions/min_length": 166.0,
      "completions/min_terminated_length": 166.0,
      "epoch": 0.18133333333333335,
      "grad_norm": 0.005056075286120176,
      "learning_rate": 8.333333333333333e-07,
      "loss": 0.0434,
      "num_tokens": 37378123.0,
      "reward": 1.1505274772644043,
      "reward_std": 0.24166499078273773,
      "rewards/accuracy_reward_step": 0.6171875,
      "rewards/final_brier_reward_step": 0.6994924545288086,
      "rewards/format_reward_step": 0.984375,
      "step": 170
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35855999030172825,
      "aux_distill/mean_u": 0.21844478544140802,
      "aux_distill/n_active_tok": 546.78125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5903972474194558,
      "calib/avg_num_step_conf": 8.63671875,
      "calib/ece": 0.30744291338582685,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.29133858267716534,
      "calib/gap": 0.07547461370034403,
      "calib/mean_conf": 0.7531082677165355,
      "calib/mu_c": 0.794411304347826,
      "calib/mu_w": 0.718936690647482,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.3038976377952757,
      "calib/std_conf": 0.21654670017537028,
      "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": 2927.0,
      "completions/max_terminated_length": 2927.0,
      "completions/mean_length": 546.68359375,
      "completions/mean_terminated_length": 548.8274536132812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 159.0,
      "epoch": 0.1824,
      "grad_norm": 0.005854424089193344,
      "learning_rate": 8.055555555555557e-07,
      "loss": 0.0422,
      "num_tokens": 37624970.0,
      "reward": 1.0410816669464111,
      "reward_std": 0.2533574104309082,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.6446633338928223,
      "rewards/format_reward_step": 0.98828125,
      "step": 171
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3715131198987365,
      "aux_distill/mean_u": 0.1825784954079074,
      "aux_distill/n_active_tok": 528.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6015790870488323,
      "calib/avg_num_step_conf": 8.26171875,
      "calib/ece": 0.2094071146245059,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.4189723320158103,
      "calib/gap": 0.08736332271762215,
      "calib/mean_conf": 0.8106719367588934,
      "calib/mu_c": 0.8438216560509555,
      "calib/mu_w": 0.7564583333333333,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.19976284584980233,
      "calib/std_conf": 0.19861363142530922,
      "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": 2886.0,
      "completions/max_terminated_length": 2886.0,
      "completions/mean_length": 519.671875,
      "completions/mean_terminated_length": 519.671875,
      "completions/min_length": 175.0,
      "completions/min_terminated_length": 175.0,
      "epoch": 0.18346666666666667,
      "grad_norm": 0.005732356104999781,
      "learning_rate": 7.777777777777779e-07,
      "loss": 0.0639,
      "num_tokens": 37861358.0,
      "reward": 1.161544919013977,
      "reward_std": 0.2524621784687042,
      "rewards/accuracy_reward_step": 0.61328125,
      "rewards/final_brier_reward_step": 0.7215273380279541,
      "rewards/format_reward_step": 0.98828125,
      "step": 172
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35221074521541595,
      "aux_distill/mean_u": 0.1902758039051572,
      "aux_distill/n_active_tok": 588.75,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5617761917586857,
      "calib/avg_num_step_conf": 9.19140625,
      "calib/ece": 0.21011269841269842,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.30158730158730157,
      "calib/gap": 0.047811096148666965,
      "calib/mean_conf": 0.7680619047619047,
      "calib/mu_c": 0.7858962025316456,
      "calib/mu_w": 0.7380851063829786,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.17559523809523814,
      "calib/std_conf": 0.2136894720231082,
      "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": 0.0078125,
      "completions/max_length": 2442.0,
      "completions/max_terminated_length": 2442.0,
      "completions/mean_length": 553.91796875,
      "completions/mean_terminated_length": 558.279541015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 124.0,
      "epoch": 0.18453333333333333,
      "grad_norm": 0.006491128820925951,
      "learning_rate": 7.5e-07,
      "loss": 0.013,
      "num_tokens": 38106321.0,
      "reward": 1.1526883840560913,
      "reward_std": 0.27811115980148315,
      "rewards/accuracy_reward_step": 0.6171875,
      "rewards/final_brier_reward_step": 0.7077205181121826,
      "rewards/format_reward_step": 0.98046875,
      "step": 173
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3911949987523258,
      "aux_distill/mean_u": 0.2356908682254667,
      "aux_distill/n_active_tok": 561.78125,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.46328694068678455,
      "calib/avg_num_step_conf": 8.96875,
      "calib/ece": 0.3510068548387096,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.3588709677419355,
      "calib/gap": -0.0404008064516127,
      "calib/mean_conf": 0.7588318548387097,
      "calib/mu_c": 0.7386314516129033,
      "calib/mu_w": 0.779032258064516,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.3049193548387096,
      "calib/std_conf": 0.23482360760128473,
      "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": 0.0078125,
      "completions/max_length": 2890.0,
      "completions/max_terminated_length": 2890.0,
      "completions/mean_length": 593.02734375,
      "completions/mean_terminated_length": 597.6968383789062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 218.0,
      "epoch": 0.1856,
      "grad_norm": 0.005600844509899616,
      "learning_rate": 7.222222222222222e-07,
      "loss": 0.0809,
      "num_tokens": 38362368.0,
      "reward": 1.0135700702667236,
      "reward_std": 0.30040162801742554,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.5818275809288025,
      "rewards/format_reward_step": 0.9609375,
      "step": 174
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38108411943539977,
      "aux_distill/mean_u": 0.22307086978765725,
      "aux_distill/n_active_tok": 589.625,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.6455332342871908,
      "calib/avg_num_step_conf": 9.2109375,
      "calib/ece": 0.3225364,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.34,
      "calib/gap": 0.10556283831793811,
      "calib/mean_conf": 0.7571436,
      "calib/mu_c": 0.8149920353982301,
      "calib/mu_w": 0.709429197080292,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.31384,
      "calib/std_conf": 0.2222298836318824,
      "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": 2844.0,
      "completions/max_terminated_length": 2844.0,
      "completions/mean_length": 586.3671875,
      "completions/mean_terminated_length": 588.6666870117188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 147.0,
      "epoch": 0.18666666666666668,
      "grad_norm": 0.005803854204714298,
      "learning_rate": 6.944444444444446e-07,
      "loss": 0.0866,
      "num_tokens": 38618302.0,
      "reward": 1.0197913646697998,
      "reward_std": 0.29834282398223877,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.6372389197349548,
      "rewards/format_reward_step": 0.9609375,
      "step": 175
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41542663518339396,
      "aux_distill/mean_u": 0.27843300225684847,
      "aux_distill/n_active_tok": 519.46875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5742283950617284,
      "calib/avg_num_step_conf": 8.08984375,
      "calib/ece": 0.3078478431372549,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.4,
      "calib/gap": 0.0673043518518519,
      "calib/mean_conf": 0.7817992156862745,
      "calib/mu_c": 0.8134718518518519,
      "calib/mu_w": 0.7461675,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.2801176470588235,
      "calib/std_conf": 0.2326337950247559,
      "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": 2417.0,
      "completions/max_terminated_length": 2417.0,
      "completions/mean_length": 552.08984375,
      "completions/mean_terminated_length": 552.08984375,
      "completions/min_length": 180.0,
      "completions/min_terminated_length": 180.0,
      "epoch": 0.18773333333333334,
      "grad_norm": 0.005393526051193476,
      "learning_rate": 6.666666666666667e-07,
      "loss": 0.0335,
      "num_tokens": 38863701.0,
      "reward": 1.09013032913208,
      "reward_std": 0.2410275936126709,
      "rewards/accuracy_reward_step": 0.52734375,
      "rewards/final_brier_reward_step": 0.6607294082641602,
      "rewards/format_reward_step": 0.9921875,
      "step": 176
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38257669610902667,
      "aux_distill/mean_u": 0.20653966698340326,
      "aux_distill/n_active_tok": 523.875,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6535557563242126,
      "calib/avg_num_step_conf": 8.1796875,
      "calib/ece": 0.2193596837944664,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.41106719367588934,
      "calib/gap": 0.12427387712958171,
      "calib/mean_conf": 0.7853043478260869,
      "calib/mu_c": 0.8363892617449663,
      "calib/mu_w": 0.7121153846153846,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.20786561264822134,
      "calib/std_conf": 0.22668207216011574,
      "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": 2861.0,
      "completions/max_terminated_length": 2861.0,
      "completions/mean_length": 523.03125,
      "completions/mean_terminated_length": 527.1495971679688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 168.0,
      "epoch": 0.1888,
      "grad_norm": 0.006122148595750332,
      "learning_rate": 6.388888888888889e-07,
      "loss": 0.0196,
      "num_tokens": 39101429.0,
      "reward": 1.1449565887451172,
      "reward_std": 0.26474517583847046,
      "rewards/accuracy_reward_step": 0.58203125,
      "rewards/final_brier_reward_step": 0.7196007966995239,
      "rewards/format_reward_step": 0.98828125,
      "step": 177
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39593329932540655,
      "aux_distill/mean_u": 0.24124242091843112,
      "aux_distill/n_active_tok": 575.0625,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6205788868600798,
      "calib/avg_num_step_conf": 8.98828125,
      "calib/ece": 0.2195551587301587,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.4365079365079365,
      "calib/gap": 0.07042708460133884,
      "calib/mean_conf": 0.8046511904761906,
      "calib/mu_c": 0.8306421383647797,
      "calib/mu_w": 0.7602150537634409,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.19662698412698412,
      "calib/std_conf": 0.21381766774115116,
      "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": 2646.0,
      "completions/max_terminated_length": 2646.0,
      "completions/mean_length": 542.3046875,
      "completions/mean_terminated_length": 542.3046875,
      "completions/min_length": 105.0,
      "completions/min_terminated_length": 105.0,
      "epoch": 0.18986666666666666,
      "grad_norm": 0.006403418257832527,
      "learning_rate": 6.111111111111112e-07,
      "loss": 0.0719,
      "num_tokens": 39346331.0,
      "reward": 1.15155029296875,
      "reward_std": 0.28153494000434875,
      "rewards/accuracy_reward_step": 0.62109375,
      "rewards/final_brier_reward_step": 0.7054443359375,
      "rewards/format_reward_step": 0.9765625,
      "step": 178
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40228676423430443,
      "aux_distill/mean_u": 0.2207415109289261,
      "aux_distill/n_active_tok": 498.84375,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.5345121867959309,
      "calib/avg_num_step_conf": 7.80078125,
      "calib/ece": 0.25700524193548396,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.40725806451612906,
      "calib/gap": 0.021218126578821717,
      "calib/mean_conf": 0.8117850806451613,
      "calib/mu_c": 0.8200841059602649,
      "calib/mu_w": 0.7988659793814432,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.9765625,
      "calib/nonempty_step_conf_rate": 0.9765625,
      "calib/pce": 0.2299596774193549,
      "calib/std_conf": 0.20379758532788697,
      "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": 0.0078125,
      "completions/max_length": 1941.0,
      "completions/max_terminated_length": 1941.0,
      "completions/mean_length": 529.1640625,
      "completions/mean_terminated_length": 533.3306884765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 175.0,
      "epoch": 0.19093333333333334,
      "grad_norm": 0.006466962397098541,
      "learning_rate": 5.833333333333334e-07,
      "loss": 0.0255,
      "num_tokens": 39588061.0,
      "reward": 1.1131529808044434,
      "reward_std": 0.30939981341362,
      "rewards/accuracy_reward_step": 0.58984375,
      "rewards/final_brier_reward_step": 0.6677122712135315,
      "rewards/format_reward_step": 0.96875,
      "step": 179
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3644904945977032,
      "aux_distill/mean_u": 0.2252707458553707,
      "aux_distill/n_active_tok": 622.5,
      "calib/answer_extract_rate": 0.96484375,
      "calib/auroc": 0.5752621412803532,
      "calib/avg_num_step_conf": 9.95703125,
      "calib/ece": 0.21748987854251006,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.3481781376518219,
      "calib/gap": 0.07612858719646798,
      "calib/mean_conf": 0.7786234817813765,
      "calib/mu_c": 0.8082119205298013,
      "calib/mu_w": 0.7320833333333333,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.19238866396761128,
      "calib/std_conf": 0.21379293004794264,
      "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": 0.01171875,
      "completions/max_length": 2986.0,
      "completions/max_terminated_length": 2986.0,
      "completions/mean_length": 643.69921875,
      "completions/mean_terminated_length": 651.33203125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 182.0,
      "epoch": 0.192,
      "grad_norm": 0.005174067337065935,
      "learning_rate": 5.555555555555555e-07,
      "loss": 0.0508,
      "num_tokens": 39856704.0,
      "reward": 1.1200428009033203,
      "reward_std": 0.2508864402770996,
      "rewards/accuracy_reward_step": 0.58984375,
      "rewards/final_brier_reward_step": 0.6932105422019958,
      "rewards/format_reward_step": 0.95703125,
      "step": 180
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3895543343387544,
      "aux_distill/mean_u": 0.19215650873399412,
      "aux_distill/n_active_tok": 504.0625,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5619111280875988,
      "calib/avg_num_step_conf": 8.0,
      "calib/ece": 0.34328047808764944,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.38645418326693226,
      "calib/gap": 0.037288349885408634,
      "calib/mean_conf": 0.800810358565737,
      "calib/mu_c": 0.8204201680672268,
      "calib/mu_w": 0.7831318181818182,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.33499362549800804,
      "calib/std_conf": 0.20024760303729527,
      "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": 2805.0,
      "completions/max_terminated_length": 2805.0,
      "completions/mean_length": 530.27734375,
      "completions/mean_terminated_length": 532.3568725585938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 213.0,
      "epoch": 0.19306666666666666,
      "grad_norm": 0.006468856241554022,
      "learning_rate": 5.277777777777779e-07,
      "loss": 0.0567,
      "num_tokens": 40098719.0,
      "reward": 1.0248982906341553,
      "reward_std": 0.2814429998397827,
      "rewards/accuracy_reward_step": 0.46484375,
      "rewards/final_brier_reward_step": 0.6083903312683105,
      "rewards/format_reward_step": 0.9765625,
      "step": 181
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36977091105654836,
      "aux_distill/mean_u": 0.20276270446751285,
      "aux_distill/n_active_tok": 520.3125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6068609022556389,
      "calib/avg_num_step_conf": 8.25390625,
      "calib/ece": 0.27892204724409453,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.4409448818897638,
      "calib/gap": 0.07826878446115282,
      "calib/mean_conf": 0.8070622047244094,
      "calib/mu_c": 0.8421907142857142,
      "calib/mu_w": 0.7639219298245614,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.26740157480314963,
      "calib/std_conf": 0.21123016812436582,
      "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": 1445.0,
      "completions/max_terminated_length": 1445.0,
      "completions/mean_length": 515.19921875,
      "completions/mean_terminated_length": 517.2196655273438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 173.0,
      "epoch": 0.19413333333333332,
      "grad_norm": 0.0057126544415950775,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.0388,
      "num_tokens": 40336770.0,
      "reward": 1.104158639907837,
      "reward_std": 0.26528361439704895,
      "rewards/accuracy_reward_step": 0.546875,
      "rewards/final_brier_reward_step": 0.6731610298156738,
      "rewards/format_reward_step": 0.98828125,
      "step": 182
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3720056042075157,
      "aux_distill/mean_u": 0.22487873949963205,
      "aux_distill/n_active_tok": 632.125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5904737903225807,
      "calib/avg_num_step_conf": 9.875,
      "calib/ece": 0.2851309523809524,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.376984126984127,
      "calib/gap": 0.09010005040322577,
      "calib/mean_conf": 0.7693134920634921,
      "calib/mu_c": 0.8136484374999999,
      "calib/mu_w": 0.7235483870967742,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.27325396825396825,
      "calib/std_conf": 0.2307191159242876,
      "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": 2362.0,
      "completions/max_terminated_length": 2362.0,
      "completions/mean_length": 610.61328125,
      "completions/mean_terminated_length": 613.0078735351562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 161.0,
      "epoch": 0.1952,
      "grad_norm": 0.005356924142688513,
      "learning_rate": 4.7222222222222226e-07,
      "loss": 0.0619,
      "num_tokens": 40599767.0,
      "reward": 1.0603950023651123,
      "reward_std": 0.3010280430316925,
      "rewards/accuracy_reward_step": 0.5,
      "rewards/final_brier_reward_step": 0.6520398855209351,
      "rewards/format_reward_step": 0.96875,
      "step": 183
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3988386648707092,
      "aux_distill/mean_u": 0.2309760148236864,
      "aux_distill/n_active_tok": 556.28125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5913656069364162,
      "calib/avg_num_step_conf": 8.640625,
      "calib/ece": 0.16279525691699614,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.36363636363636365,
      "calib/gap": 0.05032471098265889,
      "calib/mean_conf": 0.7895367588932807,
      "calib/mu_c": 0.8054497109826588,
      "calib/mu_w": 0.7551249999999999,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.13426877470355741,
      "calib/std_conf": 0.19784592767953718,
      "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": 2464.0,
      "completions/max_terminated_length": 2464.0,
      "completions/mean_length": 550.47265625,
      "completions/mean_terminated_length": 550.47265625,
      "completions/min_length": 153.0,
      "completions/min_terminated_length": 153.0,
      "epoch": 0.19626666666666667,
      "grad_norm": 0.006214865017682314,
      "learning_rate": 4.444444444444445e-07,
      "loss": 0.1089,
      "num_tokens": 40845968.0,
      "reward": 1.2052152156829834,
      "reward_std": 0.2249186486005783,
      "rewards/accuracy_reward_step": 0.67578125,
      "rewards/final_brier_reward_step": 0.7463681697845459,
      "rewards/format_reward_step": 0.98828125,
      "step": 184
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37272774055600166,
      "aux_distill/mean_u": 0.20732307433212496,
      "aux_distill/n_active_tok": 603.03125,
      "calib/answer_extract_rate": 0.96484375,
      "calib/auroc": 0.5939581667998934,
      "calib/avg_num_step_conf": 9.98828125,
      "calib/ece": 0.28278987854251014,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.4048582995951417,
      "calib/gap": 0.06050807354116705,
      "calib/mean_conf": 0.7759955465587044,
      "calib/mu_c": 0.8024525179856115,
      "calib/mu_w": 0.7419444444444444,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.24801619433198382,
      "calib/std_conf": 0.24001522037686593,
      "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": 0.01953125,
      "completions/max_length": 2910.0,
      "completions/max_terminated_length": 2910.0,
      "completions/mean_length": 570.33203125,
      "completions/mean_terminated_length": 581.6932373046875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 190.0,
      "epoch": 0.19733333333333333,
      "grad_norm": 0.005448437761515379,
      "learning_rate": 4.1666666666666667e-07,
      "loss": 0.0043,
      "num_tokens": 41098893.0,
      "reward": 1.0783536434173584,
      "reward_std": 0.28535914421081543,
      "rewards/accuracy_reward_step": 0.54296875,
      "rewards/final_brier_reward_step": 0.6528009176254272,
      "rewards/format_reward_step": 0.9609375,
      "step": 185
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40977200120687485,
      "aux_distill/mean_u": 0.2495710649067675,
      "aux_distill/n_active_tok": 578.375,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5549559036461762,
      "calib/avg_num_step_conf": 9.26171875,
      "calib/ece": 0.24531285140562253,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.3654618473895582,
      "calib/gap": 0.044854528103198565,
      "calib/mean_conf": 0.7789441767068271,
      "calib/mu_c": 0.7982190140845071,
      "calib/mu_w": 0.7533644859813086,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.22698795180722897,
      "calib/std_conf": 0.2157856955449833,
      "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": 0.015625,
      "completions/max_length": 2680.0,
      "completions/max_terminated_length": 2680.0,
      "completions/mean_length": 565.27734375,
      "completions/mean_terminated_length": 574.2500610351562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.1984,
      "grad_norm": 0.005148930009454489,
      "learning_rate": 3.8888888888888895e-07,
      "loss": 0.0008,
      "num_tokens": 41348644.0,
      "reward": 1.090340495109558,
      "reward_std": 0.2756432890892029,
      "rewards/accuracy_reward_step": 0.5546875,
      "rewards/final_brier_reward_step": 0.661149799823761,
      "rewards/format_reward_step": 0.96484375,
      "step": 186
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3799622510559857,
      "aux_distill/mean_u": 0.21517597478940167,
      "aux_distill/n_active_tok": 598.03125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.6202479338842976,
      "calib/avg_num_step_conf": 9.4296875,
      "calib/ece": 0.30274701195219134,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.3745019920318725,
      "calib/gap": 0.07251358550540377,
      "calib/mean_conf": 0.7975717131474103,
      "calib/mu_c": 0.8325284615384616,
      "calib/mu_w": 0.7600148760330578,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.2911952191235061,
      "calib/std_conf": 0.20689881747839342,
      "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": 2860.0,
      "completions/max_terminated_length": 2860.0,
      "completions/mean_length": 578.2890625,
      "completions/mean_terminated_length": 580.556884765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 160.0,
      "epoch": 0.19946666666666665,
      "grad_norm": 0.005258694291114807,
      "learning_rate": 3.611111111111111e-07,
      "loss": 0.0736,
      "num_tokens": 41598230.0,
      "reward": 1.0704033374786377,
      "reward_std": 0.2771027088165283,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.6525255441665649,
      "rewards/format_reward_step": 0.98046875,
      "step": 187
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38288334896788,
      "aux_distill/mean_u": 0.2397077996049779,
      "aux_distill/n_active_tok": 601.59375,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.6624349635796046,
      "calib/avg_num_step_conf": 9.40234375,
      "calib/ece": 0.19652540322580647,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.3709677419354839,
      "calib/gap": 0.11033354838709686,
      "calib/mean_conf": 0.7934745967741935,
      "calib/mu_c": 0.8348496774193549,
      "calib/mu_w": 0.724516129032258,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.18250000000000002,
      "calib/std_conf": 0.20560450633201344,
      "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": 0.015625,
      "completions/max_length": 2599.0,
      "completions/max_terminated_length": 2599.0,
      "completions/mean_length": 624.4296875,
      "completions/mean_terminated_length": 634.34130859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 166.0,
      "epoch": 0.20053333333333334,
      "grad_norm": 0.004758658353239298,
      "learning_rate": 3.3333333333333335e-07,
      "loss": 0.0602,
      "num_tokens": 41862156.0,
      "reward": 1.1487858295440674,
      "reward_std": 0.24109750986099243,
      "rewards/accuracy_reward_step": 0.60546875,
      "rewards/final_brier_reward_step": 0.7233529686927795,
      "rewards/format_reward_step": 0.96875,
      "step": 188
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38821121491491795,
      "aux_distill/mean_u": 0.22937579912574796,
      "aux_distill/n_active_tok": 518.25,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6888692115964843,
      "calib/avg_num_step_conf": 8.09765625,
      "calib/ece": 0.2161154150197629,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.4426877470355731,
      "calib/gap": 0.13102106782106793,
      "calib/mean_conf": 0.8089438735177866,
      "calib/mu_c": 0.8602129870129871,
      "calib/mu_w": 0.7291919191919192,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.2081818181818182,
      "calib/std_conf": 0.20390016974333175,
      "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": 2657.0,
      "completions/max_terminated_length": 2657.0,
      "completions/mean_length": 546.7890625,
      "completions/mean_terminated_length": 546.7890625,
      "completions/min_length": 160.0,
      "completions/min_terminated_length": 160.0,
      "epoch": 0.2016,
      "grad_norm": 0.005810108035802841,
      "learning_rate": 3.055555555555556e-07,
      "loss": 0.0947,
      "num_tokens": 42109902.0,
      "reward": 1.1581063270568848,
      "reward_std": 0.22317633032798767,
      "rewards/accuracy_reward_step": 0.6015625,
      "rewards/final_brier_reward_step": 0.7302750945091248,
      "rewards/format_reward_step": 0.984375,
      "step": 189
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.370077112223953,
      "aux_distill/mean_u": 0.2316970306220341,
      "aux_distill/n_active_tok": 632.40625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6490683229813664,
      "calib/avg_num_step_conf": 9.875,
      "calib/ece": 0.15501976284584967,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.33992094861660077,
      "calib/gap": 0.11170807453416143,
      "calib/mean_conf": 0.7834782608695652,
      "calib/mu_c": 0.8240993788819876,
      "calib/mu_w": 0.7123913043478262,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.15106719367588922,
      "calib/std_conf": 0.19828851828568378,
      "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": 2046.0,
      "completions/max_terminated_length": 2046.0,
      "completions/mean_length": 614.49609375,
      "completions/mean_terminated_length": 616.9059448242188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 184.0,
      "epoch": 0.20266666666666666,
      "grad_norm": 0.0051289028488099575,
      "learning_rate": 2.7777777777777776e-07,
      "loss": 0.0136,
      "num_tokens": 42372821.0,
      "reward": 1.1838111877441406,
      "reward_std": 0.21080797910690308,
      "rewards/accuracy_reward_step": 0.62890625,
      "rewards/final_brier_reward_step": 0.7504351735115051,
      "rewards/format_reward_step": 0.98828125,
      "step": 190
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3612460242584348,
      "aux_distill/mean_u": 0.18409684871874585,
      "aux_distill/n_active_tok": 587.25,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5992965018716924,
      "calib/avg_num_step_conf": 9.17578125,
      "calib/ece": 0.27774297188755037,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.3815261044176707,
      "calib/gap": 0.09912056279850268,
      "calib/mean_conf": 0.781293172690763,
      "calib/mu_c": 0.8298582677165355,
      "calib/mu_w": 0.7307377049180328,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.2744979919678716,
      "calib/std_conf": 0.21773484849829758,
      "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": 2912.0,
      "completions/max_terminated_length": 2912.0,
      "completions/mean_length": 573.2890625,
      "completions/mean_terminated_length": 575.5372924804688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 168.0,
      "epoch": 0.20373333333333332,
      "grad_norm": 0.005880712065845728,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.0726,
      "num_tokens": 42623751.0,
      "reward": 1.0624706745147705,
      "reward_std": 0.24356943368911743,
      "rewards/accuracy_reward_step": 0.49609375,
      "rewards/final_brier_reward_step": 0.6600974798202515,
      "rewards/format_reward_step": 0.96875,
      "step": 191
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3918271157890558,
      "aux_distill/mean_u": 0.2257468611779381,
      "aux_distill/n_active_tok": 514.0625,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.6300331125827815,
      "calib/avg_num_step_conf": 8.03515625,
      "calib/ece": 0.20690677290836654,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.3784860557768924,
      "calib/gap": 0.106654728476821,
      "calib/mean_conf": 0.7816988047808765,
      "calib/mu_c": 0.8241907284768212,
      "calib/mu_w": 0.7175360000000002,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.19350597609561754,
      "calib/std_conf": 0.2273512524500433,
      "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": 0.0078125,
      "completions/max_length": 2626.0,
      "completions/max_terminated_length": 2626.0,
      "completions/mean_length": 557.2890625,
      "completions/mean_terminated_length": 561.6771850585938,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 141.0,
      "epoch": 0.2048,
      "grad_norm": 0.006296477280557156,
      "learning_rate": 2.2222222222222224e-07,
      "loss": 0.0647,
      "num_tokens": 42871393.0,
      "reward": 1.1417138576507568,
      "reward_std": 0.2670000195503235,
      "rewards/accuracy_reward_step": 0.58984375,
      "rewards/final_brier_reward_step": 0.7131150364875793,
      "rewards/format_reward_step": 0.98046875,
      "step": 192
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40630976390093565,
      "aux_distill/mean_u": 0.22802555206204686,
      "aux_distill/n_active_tok": 538.90625,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.6023408972352636,
      "calib/avg_num_step_conf": 8.66015625,
      "calib/ece": 0.2670236,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.46,
      "calib/gap": 0.048302217005738224,
      "calib/mean_conf": 0.8181764,
      "calib/mu_c": 0.8390429577464789,
      "calib/mu_w": 0.7907407407407406,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.25860000000000005,
      "calib/std_conf": 0.19366300452858826,
      "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": 0.0078125,
      "completions/max_length": 2099.0,
      "completions/max_terminated_length": 2099.0,
      "completions/mean_length": 521.734375,
      "completions/mean_terminated_length": 525.842529296875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 211.0,
      "epoch": 0.20586666666666667,
      "grad_norm": 0.006823267322033644,
      "learning_rate": 1.9444444444444447e-07,
      "loss": 0.0735,
      "num_tokens": 43110669.0,
      "reward": 1.0947258472442627,
      "reward_std": 0.38915589451789856,
      "rewards/accuracy_reward_step": 0.5546875,
      "rewards/final_brier_reward_step": 0.6621078252792358,
      "rewards/format_reward_step": 0.97265625,
      "step": 193
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37754127802327275,
      "aux_distill/mean_u": 0.2385847335013782,
      "aux_distill/n_active_tok": 527.5,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5573550121623352,
      "calib/avg_num_step_conf": 8.421875,
      "calib/ece": 0.2580699604743084,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.40711462450592883,
      "calib/gap": 0.04432505441044665,
      "calib/mean_conf": 0.8026415019762846,
      "calib/mu_c": 0.8213876712328766,
      "calib/mu_w": 0.7770626168224299,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.24181818181818193,
      "calib/std_conf": 0.20781324405356888,
      "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": 2675.0,
      "completions/max_terminated_length": 2675.0,
      "completions/mean_length": 520.0703125,
      "completions/mean_terminated_length": 522.10986328125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 160.0,
      "epoch": 0.20693333333333333,
      "grad_norm": 0.006866298615932465,
      "learning_rate": 1.6666666666666668e-07,
      "loss": 0.0503,
      "num_tokens": 43349751.0,
      "reward": 1.1137315034866333,
      "reward_std": 0.2670598328113556,
      "rewards/accuracy_reward_step": 0.5703125,
      "rewards/final_brier_reward_step": 0.672775387763977,
      "rewards/format_reward_step": 0.984375,
      "step": 194
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39284872729331255,
      "aux_distill/mean_u": 0.25530160009173675,
      "aux_distill/n_active_tok": 561.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.4998744507219084,
      "calib/avg_num_step_conf": 8.765625,
      "calib/ece": 0.2733596837944663,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.36363636363636365,
      "calib/gap": 0.01785122410546125,
      "calib/mean_conf": 0.7901185770750988,
      "calib/mu_c": 0.7984444444444444,
      "calib/mu_w": 0.7805932203389832,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.26494071146245046,
      "calib/std_conf": 0.19763319361916326,
      "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": 2951.0,
      "completions/max_terminated_length": 2951.0,
      "completions/mean_length": 582.671875,
      "completions/mean_terminated_length": 582.671875,
      "completions/min_length": 162.0,
      "completions/min_terminated_length": 162.0,
      "epoch": 0.208,
      "grad_norm": 0.005913162138313055,
      "learning_rate": 1.3888888888888888e-07,
      "loss": 0.0545,
      "num_tokens": 43604899.0,
      "reward": 1.0780566930770874,
      "reward_std": 0.25560739636421204,
      "rewards/accuracy_reward_step": 0.52734375,
      "rewards/final_brier_reward_step": 0.6443945169448853,
      "rewards/format_reward_step": 0.984375,
      "step": 195
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3867280255071819,
      "aux_distill/mean_u": 0.20433668531479132,
      "aux_distill/n_active_tok": 469.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5802331349206351,
      "calib/avg_num_step_conf": 7.3828125,
      "calib/ece": 0.2953945312499999,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.4921875,
      "calib/gap": 0.06552678571428583,
      "calib/mean_conf": 0.85133203125,
      "calib/mu_c": 0.8800000000000001,
      "calib/mu_w": 0.8144732142857143,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.2921132812499999,
      "calib/std_conf": 0.16652546231432305,
      "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": 1235.0,
      "completions/max_terminated_length": 1235.0,
      "completions/mean_length": 448.0,
      "completions/mean_terminated_length": 449.75689697265625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 155.0,
      "epoch": 0.20906666666666668,
      "grad_norm": 0.00626879557967186,
      "learning_rate": 1.1111111111111112e-07,
      "loss": 0.0276,
      "num_tokens": 43822131.0,
      "reward": 1.1187515258789062,
      "reward_std": 0.21417656540870667,
      "rewards/accuracy_reward_step": 0.5625,
      "rewards/final_brier_reward_step": 0.6750030517578125,
      "rewards/format_reward_step": 1.0,
      "step": 196
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3991149263456464,
      "aux_distill/mean_u": 0.24741979634781106,
      "aux_distill/n_active_tok": 582.53125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5721665721665722,
      "calib/avg_num_step_conf": 9.1015625,
      "calib/ece": 0.2724870078740157,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.44881889763779526,
      "calib/gap": 0.07266039816039827,
      "calib/mean_conf": 0.8078279527559056,
      "calib/mu_c": 0.839581118881119,
      "calib/mu_w": 0.7669207207207207,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.25866141732283465,
      "calib/std_conf": 0.21488679432697524,
      "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": 3025.0,
      "completions/max_terminated_length": 3025.0,
      "completions/mean_length": 559.3671875,
      "completions/mean_terminated_length": 561.560791015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 145.0,
      "epoch": 0.21013333333333334,
      "grad_norm": 0.00560788344591856,
      "learning_rate": 8.333333333333334e-08,
      "loss": 0.0351,
      "num_tokens": 44070385.0,
      "reward": 1.11452054977417,
      "reward_std": 0.2765326201915741,
      "rewards/accuracy_reward_step": 0.55859375,
      "rewards/final_brier_reward_step": 0.6782600283622742,
      "rewards/format_reward_step": 0.9921875,
      "step": 197
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3851682706736028,
      "aux_distill/mean_u": 0.1962447667898319,
      "aux_distill/n_active_tok": 536.71875,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5855562784645412,
      "calib/avg_num_step_conf": 8.51953125,
      "calib/ece": 0.24864541832669326,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.38247011952191234,
      "calib/gap": 0.0758646714378659,
      "calib/mean_conf": 0.800996015936255,
      "calib/mu_c": 0.8330344827586207,
      "calib/mu_w": 0.7571698113207548,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.2359760956175299,
      "calib/std_conf": 0.19730304109153482,
      "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": 3025.0,
      "completions/max_terminated_length": 3025.0,
      "completions/mean_length": 526.3125,
      "completions/mean_terminated_length": 528.3765258789062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 156.0,
      "epoch": 0.2112,
      "grad_norm": 0.005777986254543066,
      "learning_rate": 5.555555555555556e-08,
      "loss": 0.0533,
      "num_tokens": 44310505.0,
      "reward": 1.1186888217926025,
      "reward_std": 0.29297205805778503,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/final_brier_reward_step": 0.6905027627944946,
      "rewards/format_reward_step": 0.98046875,
      "step": 198
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3839912237599492,
      "aux_distill/mean_u": 0.25348926264527655,
      "aux_distill/n_active_tok": 562.03125,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.6176393699758186,
      "calib/avg_num_step_conf": 8.8203125,
      "calib/ece": 0.25571079999999996,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.428,
      "calib/gap": 0.09906385857133526,
      "calib/mean_conf": 0.7651692,
      "calib/mu_c": 0.8075685314685316,
      "calib/mu_w": 0.7085046728971963,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.22443999999999997,
      "calib/std_conf": 0.2612843289050455,
      "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": 0.01171875,
      "completions/max_length": 3061.0,
      "completions/max_terminated_length": 3061.0,
      "completions/mean_length": 599.20703125,
      "completions/mean_terminated_length": 606.312255859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 148.0,
      "epoch": 0.21226666666666666,
      "grad_norm": 0.0047719781287014484,
      "learning_rate": 2.777777777777778e-08,
      "loss": -0.0175,
      "num_tokens": 44568102.0,
      "reward": 1.1084496974945068,
      "reward_std": 0.30771604180336,
      "rewards/accuracy_reward_step": 0.55859375,
      "rewards/final_brier_reward_step": 0.6817431449890137,
      "rewards/format_reward_step": 0.9765625,
      "step": 199
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3788956059142947,
      "aux_distill/mean_u": 0.21847873765485837,
      "aux_distill/n_active_tok": 494.5,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5257306626354245,
      "calib/avg_num_step_conf": 7.71875,
      "calib/ece": 0.24171921568627447,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.39215686274509803,
      "calib/gap": 0.04255623582766466,
      "calib/mean_conf": 0.7831435294117648,
      "calib/mu_c": 0.8011673469387756,
      "calib/mu_w": 0.7586111111111109,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.22419607843137251,
      "calib/std_conf": 0.22645131673339491,
      "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": 2081.0,
      "completions/max_terminated_length": 2081.0,
      "completions/mean_length": 545.921875,
      "completions/mean_terminated_length": 548.0628051757812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 106.0,
      "epoch": 0.21333333333333335,
      "grad_norm": 0.005753261968493462,
      "learning_rate": 0.0,
      "loss": 0.0435,
      "num_tokens": 44815906.0,
      "reward": 1.1214730739593506,
      "reward_std": 0.23278102278709412,
      "rewards/accuracy_reward_step": 0.57421875,
      "rewards/final_brier_reward_step": 0.6765400171279907,
      "rewards/format_reward_step": 0.9921875,
      "step": 200
    },
    {
      "epoch": 0.21333333333333335,
      "step": 200,
      "total_flos": 0.0,
      "train_loss": 0.06408213273854926,
      "train_runtime": 18422.6439,
      "train_samples_per_second": 2.779,
      "train_steps_per_second": 0.011
    }
  ],
  "logging_steps": 1,
  "max_steps": 200,
  "num_input_tokens_seen": 44815906,
  "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
}