{
  "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": 0.012633531354367733,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.0691,
      "num_tokens": 264685.0,
      "reward": 0.045387499034404755,
      "reward_std": 0.09252828359603882,
      "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": 0.008606677874922752,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.0988,
      "num_tokens": 533467.0,
      "reward": 0.09099707007408142,
      "reward_std": 0.17441941797733307,
      "rewards/accuracy_reward_step": 0.03125,
      "rewards/final_brier_reward_step": 0.02965039201080799,
      "rewards/format_reward_step": 0.08984375,
      "step": 2
    },
    {
      "aux_distill/lambda": 0.1,
      "aux_distill/loss": 0.6411322099821908,
      "aux_distill/mean_u": 0.3704558609972279,
      "aux_distill/n_active_tok": 37.142857142857146,
      "calib/answer_extract_rate": 0.06640625,
      "calib/auroc": 0.423076923076923,
      "calib/avg_num_step_conf": 0.25390625,
      "calib/ece": 0.7649374999999999,
      "calib/final_conf_rate": 0.0625,
      "calib/format_rate": 0.05859375,
      "calib/frac_conf_gt_0.9": 0.875,
      "calib/gap": -0.003410256410256385,
      "calib/mean_conf": 0.9324374999999999,
      "calib/mu_c": 0.9296666666666668,
      "calib/mu_w": 0.9330769230769231,
      "calib/nonempty_final_conf_rate": 0.0625,
      "calib/nonempty_reasoning_rate": 0.0703125,
      "calib/nonempty_step_conf_rate": 0.0625,
      "calib/pce": 0.7549374999999999,
      "calib/std_conf": 0.10546502308229966,
      "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": 0.11328125,
      "completions/max_length": 2998.0,
      "completions/max_terminated_length": 2998.0,
      "completions/mean_length": 577.6328125,
      "completions/mean_terminated_length": 651.4273071289062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0032,
      "grad_norm": 0.009327870793640614,
      "learning_rate": 7.5e-07,
      "loss": 0.0704,
      "num_tokens": 786597.0,
      "reward": 0.047834958881139755,
      "reward_std": 0.105415940284729,
      "rewards/accuracy_reward_step": 0.01171875,
      "rewards/final_brier_reward_step": 0.01363866776227951,
      "rewards/format_reward_step": 0.05859375,
      "step": 3
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.6355656509598097,
      "aux_distill/mean_u": 0.32851766073949795,
      "aux_distill/n_active_tok": 54.75,
      "calib/answer_extract_rate": 0.08203125,
      "calib/avg_num_step_conf": 0.3203125,
      "calib/ece": 0.9392307692307692,
      "calib/final_conf_rate": 0.05078125,
      "calib/format_rate": 0.03125,
      "calib/frac_conf_gt_0.9": 0.7692307692307693,
      "calib/mean_conf": 0.9392307692307693,
      "calib/mu_c": NaN,
      "calib/mu_w": 0.9392307692307693,
      "calib/nonempty_final_conf_rate": 0.05078125,
      "calib/nonempty_reasoning_rate": 0.09765625,
      "calib/nonempty_step_conf_rate": 0.0546875,
      "calib/pce": 0.9392307692307692,
      "calib/std_conf": 0.06498065255982423,
      "calib/step_conf_rate": 0.0546875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3031.0,
      "completions/max_terminated_length": 3031.0,
      "completions/mean_length": 705.89453125,
      "completions/mean_terminated_length": 772.2607421875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.004266666666666667,
      "grad_norm": 0.006254538893699646,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0338,
      "num_tokens": 1073474.0,
      "reward": 0.018050584942102432,
      "reward_std": 0.04258500412106514,
      "rewards/accuracy_reward_step": 0.0,
      "rewards/final_brier_reward_step": 0.004851171746850014,
      "rewards/format_reward_step": 0.03125,
      "step": 4
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.6282181943004782,
      "aux_distill/mean_u": 0.2853342183461234,
      "aux_distill/n_active_tok": 51.63636363636363,
      "calib/answer_extract_rate": 0.05859375,
      "calib/auroc": 0.9090909090909091,
      "calib/avg_num_step_conf": 0.27734375,
      "calib/ece": 0.8525000000000003,
      "calib/final_conf_rate": 0.046875,
      "calib/format_rate": 0.0390625,
      "calib/frac_conf_gt_0.9": 0.75,
      "calib/gap": 0.04818181818181799,
      "calib/mean_conf": 0.9358333333333335,
      "calib/mu_c": 0.98,
      "calib/mu_w": 0.931818181818182,
      "calib/nonempty_final_conf_rate": 0.046875,
      "calib/nonempty_reasoning_rate": 0.07421875,
      "calib/nonempty_step_conf_rate": 0.0546875,
      "calib/pce": 0.8525000000000003,
      "calib/std_conf": 0.05780114377953278,
      "calib/step_conf_rate": 0.0546875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3033.0,
      "completions/max_terminated_length": 3033.0,
      "completions/mean_length": 670.96484375,
      "completions/mean_terminated_length": 727.8262939453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.005333333333333333,
      "grad_norm": 0.007472812198102474,
      "learning_rate": 1.25e-06,
      "loss": 0.0364,
      "num_tokens": 1351929.0,
      "reward": 0.02683398313820362,
      "reward_std": 0.07185898721218109,
      "rewards/accuracy_reward_step": 0.00390625,
      "rewards/final_brier_reward_step": 0.006792968139052391,
      "rewards/format_reward_step": 0.0390625,
      "step": 5
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.6285811681300402,
      "aux_distill/mean_u": 0.3202117609486028,
      "aux_distill/n_active_tok": 56.5,
      "calib/answer_extract_rate": 0.09765625,
      "calib/auroc": 0.5357142857142856,
      "calib/avg_num_step_conf": 0.44140625,
      "calib/ece": 0.6455000000000001,
      "calib/final_conf_rate": 0.078125,
      "calib/format_rate": 0.06640625,
      "calib/frac_conf_gt_0.9": 0.8,
      "calib/gap": 0.013095238095237938,
      "calib/mean_conf": 0.8975000000000002,
      "calib/mu_c": 0.9066666666666666,
      "calib/mu_w": 0.8935714285714287,
      "calib/nonempty_final_conf_rate": 0.078125,
      "calib/nonempty_reasoning_rate": 0.12109375,
      "calib/nonempty_step_conf_rate": 0.08984375,
      "calib/pce": 0.6215,
      "calib/std_conf": 0.19531704994700286,
      "calib/step_conf_rate": 0.08984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2517.0,
      "completions/max_terminated_length": 2517.0,
      "completions/mean_length": 567.91015625,
      "completions/mean_terminated_length": 621.303466796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.0064,
      "grad_norm": 0.007908736355602741,
      "learning_rate": 1.5e-06,
      "loss": 0.0879,
      "num_tokens": 1603266.0,
      "reward": 0.0757259726524353,
      "reward_std": 0.16911853849887848,
      "rewards/accuracy_reward_step": 0.02734375,
      "rewards/final_brier_reward_step": 0.030358202755451202,
      "rewards/format_reward_step": 0.06640625,
      "step": 6
    },
    {
      "aux_distill/lambda": 0.1,
      "aux_distill/loss": 0.7187803174768176,
      "aux_distill/mean_u": 0.35720473139848385,
      "aux_distill/n_active_tok": 37.714285714285715,
      "calib/answer_extract_rate": 0.08984375,
      "calib/auroc": 0.7181818181818183,
      "calib/avg_num_step_conf": 0.265625,
      "calib/ece": 0.535,
      "calib/final_conf_rate": 0.0625,
      "calib/format_rate": 0.04296875,
      "calib/frac_conf_gt_0.9": 0.8125,
      "calib/gap": 0.18109090909090886,
      "calib/mean_conf": 0.8474999999999999,
      "calib/mu_c": 0.9719999999999999,
      "calib/mu_w": 0.790909090909091,
      "calib/nonempty_final_conf_rate": 0.0625,
      "calib/nonempty_reasoning_rate": 0.1015625,
      "calib/nonempty_step_conf_rate": 0.06640625,
      "calib/pce": 0.535,
      "calib/std_conf": 0.2674532669458348,
      "calib/step_conf_rate": 0.06640625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3029.0,
      "completions/max_terminated_length": 3029.0,
      "completions/mean_length": 691.8046875,
      "completions/mean_terminated_length": 770.0086669921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.007466666666666667,
      "grad_norm": 0.009156746789813042,
      "learning_rate": 1.75e-06,
      "loss": 0.0396,
      "num_tokens": 1887792.0,
      "reward": 0.05140624940395355,
      "reward_std": 0.13243496417999268,
      "rewards/accuracy_reward_step": 0.01953125,
      "rewards/final_brier_reward_step": 0.020781250670552254,
      "rewards/format_reward_step": 0.04296875,
      "step": 7
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.803349397248692,
      "aux_distill/mean_u": 0.3246804844449429,
      "aux_distill/n_active_tok": 51.833333333333336,
      "calib/answer_extract_rate": 0.109375,
      "calib/auroc": 0.575,
      "calib/avg_num_step_conf": 0.45703125,
      "calib/ece": 0.5033333333333333,
      "calib/final_conf_rate": 0.0703125,
      "calib/format_rate": 0.06640625,
      "calib/frac_conf_gt_0.9": 0.8333333333333334,
      "calib/gap": 0.010749999999999815,
      "calib/mean_conf": 0.9477777777777777,
      "calib/mu_c": 0.95375,
      "calib/mu_w": 0.9430000000000002,
      "calib/nonempty_final_conf_rate": 0.0703125,
      "calib/nonempty_reasoning_rate": 0.13671875,
      "calib/nonempty_step_conf_rate": 0.09765625,
      "calib/pce": 0.5033333333333333,
      "calib/std_conf": 0.03866602809178958,
      "calib/step_conf_rate": 0.09765625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2839.0,
      "completions/max_terminated_length": 2839.0,
      "completions/mean_length": 670.27734375,
      "completions/mean_terminated_length": 717.9539794921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.008533333333333334,
      "grad_norm": 0.01636502891778946,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0735,
      "num_tokens": 2165895.0,
      "reward": 0.0802062526345253,
      "reward_std": 0.13919153809547424,
      "rewards/accuracy_reward_step": 0.03125,
      "rewards/final_brier_reward_step": 0.0315062515437603,
      "rewards/format_reward_step": 0.06640625,
      "step": 8
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.6690852016210556,
      "aux_distill/mean_u": 0.4719379662396538,
      "aux_distill/n_active_tok": 43.7,
      "calib/answer_extract_rate": 0.0703125,
      "calib/avg_num_step_conf": 0.2109375,
      "calib/ece": 0.7616666666666667,
      "calib/final_conf_rate": 0.0234375,
      "calib/format_rate": 0.01953125,
      "calib/frac_conf_gt_0.9": 0.6666666666666666,
      "calib/mean_conf": 0.7616666666666666,
      "calib/mu_c": NaN,
      "calib/mu_w": 0.7616666666666666,
      "calib/nonempty_final_conf_rate": 0.0234375,
      "calib/nonempty_reasoning_rate": 0.078125,
      "calib/nonempty_step_conf_rate": 0.04296875,
      "calib/pce": 0.7616666666666667,
      "calib/std_conf": 0.35248719440884974,
      "calib/step_conf_rate": 0.04296875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0859375,
      "completions/max_length": 3053.0,
      "completions/max_terminated_length": 3053.0,
      "completions/mean_length": 659.15625,
      "completions/mean_terminated_length": 721.1282348632812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0096,
      "grad_norm": 0.006472648587077856,
      "learning_rate": 2.25e-06,
      "loss": 0.0353,
      "num_tokens": 2442175.0,
      "reward": 0.011277538724243641,
      "reward_std": 0.03189769759774208,
      "rewards/accuracy_reward_step": 0.0,
      "rewards/final_brier_reward_step": 0.0030238283798098564,
      "rewards/format_reward_step": 0.01953125,
      "step": 9
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.6920461853345236,
      "aux_distill/mean_u": 0.2160209553051777,
      "aux_distill/n_active_tok": 48.0,
      "calib/answer_extract_rate": 0.125,
      "calib/auroc": 0.3695652173913044,
      "calib/avg_num_step_conf": 0.421875,
      "calib/ece": 0.7415384615384615,
      "calib/final_conf_rate": 0.1015625,
      "calib/format_rate": 0.0703125,
      "calib/frac_conf_gt_0.9": 0.7307692307692307,
      "calib/gap": -0.053043478260869636,
      "calib/mean_conf": 0.856923076923077,
      "calib/mu_c": 0.8099999999999999,
      "calib/mu_w": 0.8630434782608696,
      "calib/nonempty_final_conf_rate": 0.1015625,
      "calib/nonempty_reasoning_rate": 0.15234375,
      "calib/nonempty_step_conf_rate": 0.10546875,
      "calib/pce": 0.7415384615384615,
      "calib/std_conf": 0.26039569343551616,
      "calib/step_conf_rate": 0.10546875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3024.0,
      "completions/max_terminated_length": 3024.0,
      "completions/mean_length": 626.609375,
      "completions/mean_terminated_length": 679.7118530273438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.010666666666666666,
      "grad_norm": 0.00994604080915451,
      "learning_rate": 2.5e-06,
      "loss": 0.0655,
      "num_tokens": 2709387.0,
      "reward": 0.051632028073072433,
      "reward_std": 0.11215013265609741,
      "rewards/accuracy_reward_step": 0.01171875,
      "rewards/final_brier_reward_step": 0.00951406266540289,
      "rewards/format_reward_step": 0.0703125,
      "step": 10
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.6356026422977448,
      "aux_distill/mean_u": 0.29512467347753385,
      "aux_distill/n_active_tok": 58.72,
      "calib/answer_extract_rate": 0.13671875,
      "calib/auroc": 0.4682539682539683,
      "calib/avg_num_step_conf": 0.69921875,
      "calib/ece": 0.6884,
      "calib/final_conf_rate": 0.09765625,
      "calib/format_rate": 0.0859375,
      "calib/frac_conf_gt_0.9": 0.84,
      "calib/gap": -0.03428571428571414,
      "calib/mean_conf": 0.9404,
      "calib/mu_c": 0.9157142857142857,
      "calib/mu_w": 0.9499999999999998,
      "calib/nonempty_final_conf_rate": 0.09765625,
      "calib/nonempty_reasoning_rate": 0.1953125,
      "calib/nonempty_step_conf_rate": 0.15234375,
      "calib/pce": 0.6744,
      "calib/std_conf": 0.07345638161521433,
      "calib/step_conf_rate": 0.15234375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2943.0,
      "completions/max_terminated_length": 2943.0,
      "completions/mean_length": 661.31640625,
      "completions/mean_terminated_length": 717.3601684570312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.011733333333333333,
      "grad_norm": 0.01098595280200243,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.0929,
      "num_tokens": 2983164.0,
      "reward": 0.09054765850305557,
      "reward_std": 0.18365205824375153,
      "rewards/accuracy_reward_step": 0.03125,
      "rewards/final_brier_reward_step": 0.032657817006111145,
      "rewards/format_reward_step": 0.0859375,
      "step": 11
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.625461015701294,
      "aux_distill/mean_u": 0.3472341701960909,
      "aux_distill/n_active_tok": 50.88,
      "calib/answer_extract_rate": 0.19921875,
      "calib/auroc": 0.316,
      "calib/avg_num_step_conf": 0.625,
      "calib/ece": 0.6594285714285714,
      "calib/final_conf_rate": 0.13671875,
      "calib/format_rate": 0.1015625,
      "calib/frac_conf_gt_0.9": 0.8,
      "calib/gap": -0.03220000000000012,
      "calib/mean_conf": 0.884,
      "calib/mu_c": 0.861,
      "calib/mu_w": 0.8932000000000001,
      "calib/nonempty_final_conf_rate": 0.13671875,
      "calib/nonempty_reasoning_rate": 0.21875,
      "calib/nonempty_step_conf_rate": 0.13671875,
      "calib/pce": 0.6288571428571428,
      "calib/std_conf": 0.21212125912182533,
      "calib/step_conf_rate": 0.13671875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 3064.0,
      "completions/max_terminated_length": 3064.0,
      "completions/mean_length": 608.4296875,
      "completions/mean_terminated_length": 659.9915161132812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0128,
      "grad_norm": 0.01760835014283657,
      "learning_rate": 3e-06,
      "loss": 0.1187,
      "num_tokens": 3243098.0,
      "reward": 0.1129060611128807,
      "reward_std": 0.21075576543807983,
      "rewards/accuracy_reward_step": 0.0390625,
      "rewards/final_brier_reward_step": 0.04612461104989052,
      "rewards/format_reward_step": 0.1015625,
      "step": 12
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.5849946607714114,
      "aux_distill/mean_u": 0.27308247281210013,
      "aux_distill/n_active_tok": 72.8695652173913,
      "calib/answer_extract_rate": 0.15625,
      "calib/auroc": 0.48214285714285715,
      "calib/avg_num_step_conf": 0.8203125,
      "calib/ece": 0.785,
      "calib/final_conf_rate": 0.125,
      "calib/format_rate": 0.09375,
      "calib/frac_conf_gt_0.9": 0.71875,
      "calib/gap": 0.03142857142857136,
      "calib/mean_conf": 0.9099999999999999,
      "calib/mu_c": 0.9375,
      "calib/mu_w": 0.9060714285714286,
      "calib/nonempty_final_conf_rate": 0.125,
      "calib/nonempty_reasoning_rate": 0.23046875,
      "calib/nonempty_step_conf_rate": 0.17578125,
      "calib/pce": 0.785,
      "calib/std_conf": 0.15313392831113554,
      "calib/step_conf_rate": 0.17578125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2971.0,
      "completions/max_terminated_length": 2971.0,
      "completions/mean_length": 593.6015625,
      "completions/mean_terminated_length": 649.4102783203125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.013866666666666666,
      "grad_norm": 0.00963516067713499,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.0712,
      "num_tokens": 3499652.0,
      "reward": 0.07337792962789536,
      "reward_std": 0.14072483777999878,
      "rewards/accuracy_reward_step": 0.015625,
      "rewards/final_brier_reward_step": 0.02175585925579071,
      "rewards/format_reward_step": 0.09375,
      "step": 13
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5950426235795021,
      "aux_distill/mean_u": 0.2380964431578831,
      "aux_distill/n_active_tok": 60.75,
      "calib/answer_extract_rate": 0.20703125,
      "calib/auroc": 0.4095238095238095,
      "calib/avg_num_step_conf": 0.796875,
      "calib/ece": 0.6704878048780488,
      "calib/final_conf_rate": 0.16015625,
      "calib/format_rate": 0.125,
      "calib/frac_conf_gt_0.9": 0.6341463414634146,
      "calib/gap": -0.09133333333333349,
      "calib/mean_conf": 0.8046341463414632,
      "calib/mu_c": 0.7266666666666666,
      "calib/mu_w": 0.8180000000000001,
      "calib/nonempty_final_conf_rate": 0.16015625,
      "calib/nonempty_reasoning_rate": 0.26171875,
      "calib/nonempty_step_conf_rate": 0.1875,
      "calib/pce": 0.6643902439024391,
      "calib/std_conf": 0.2539166135543427,
      "calib/step_conf_rate": 0.1875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3047.0,
      "completions/max_terminated_length": 3047.0,
      "completions/mean_length": 584.20703125,
      "completions/mean_terminated_length": 633.7161254882812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 7.0,
      "epoch": 0.014933333333333333,
      "grad_norm": 0.01152648776769638,
      "learning_rate": 3.5e-06,
      "loss": 0.0984,
      "num_tokens": 3754609.0,
      "reward": 0.10741367191076279,
      "reward_std": 0.19179916381835938,
      "rewards/accuracy_reward_step": 0.0234375,
      "rewards/final_brier_reward_step": 0.042952343821525574,
      "rewards/format_reward_step": 0.125,
      "step": 14
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5943296541427744,
      "aux_distill/mean_u": 0.35168131655490353,
      "aux_distill/n_active_tok": 101.96551724137932,
      "calib/answer_extract_rate": 0.26171875,
      "calib/auroc": 0.5046583850931676,
      "calib/avg_num_step_conf": 1.4296875,
      "calib/ece": 0.7773584905660378,
      "calib/final_conf_rate": 0.20703125,
      "calib/format_rate": 0.17578125,
      "calib/frac_conf_gt_0.9": 0.7547169811320755,
      "calib/gap": -0.04512422360248458,
      "calib/mean_conf": 0.8777358490566037,
      "calib/mu_c": 0.8385714285714286,
      "calib/mu_w": 0.8836956521739132,
      "calib/nonempty_final_conf_rate": 0.20703125,
      "calib/nonempty_reasoning_rate": 0.3125,
      "calib/nonempty_step_conf_rate": 0.25,
      "calib/pce": 0.7615094339622641,
      "calib/std_conf": 0.20207080311010325,
      "calib/step_conf_rate": 0.25,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 2978.0,
      "completions/max_terminated_length": 2978.0,
      "completions/mean_length": 571.8984375,
      "completions/mean_terminated_length": 610.0250244140625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.016,
      "grad_norm": 0.012240985408425331,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.145,
      "num_tokens": 4008895.0,
      "reward": 0.13784393668174744,
      "reward_std": 0.20476070046424866,
      "rewards/accuracy_reward_step": 0.02734375,
      "rewards/final_brier_reward_step": 0.045219141989946365,
      "rewards/format_reward_step": 0.17578125,
      "step": 15
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5600810666238109,
      "aux_distill/mean_u": 0.3982821064576951,
      "aux_distill/n_active_tok": 90.41935483870968,
      "calib/answer_extract_rate": 0.296875,
      "calib/auroc": 0.523109243697479,
      "calib/avg_num_step_conf": 1.3359375,
      "calib/ece": 0.5483203389830508,
      "calib/final_conf_rate": 0.23046875,
      "calib/format_rate": 0.203125,
      "calib/frac_conf_gt_0.9": 0.5084745762711864,
      "calib/gap": 0.028301540616246457,
      "calib/mean_conf": 0.7563237288135594,
      "calib/mu_c": 0.7764705882352941,
      "calib/mu_w": 0.7481690476190477,
      "calib/nonempty_final_conf_rate": 0.23046875,
      "calib/nonempty_reasoning_rate": 0.3828125,
      "calib/nonempty_step_conf_rate": 0.30078125,
      "calib/pce": 0.5082542372881356,
      "calib/std_conf": 0.30900877138501365,
      "calib/step_conf_rate": 0.30078125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.04296875,
      "completions/max_length": 2765.0,
      "completions/max_terminated_length": 2765.0,
      "completions/mean_length": 599.84375,
      "completions/mean_terminated_length": 626.7755126953125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.017066666666666667,
      "grad_norm": 0.01254871767014265,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.1227,
      "num_tokens": 4271303.0,
      "reward": 0.21736013889312744,
      "reward_std": 0.37067097425460815,
      "rewards/accuracy_reward_step": 0.06640625,
      "rewards/final_brier_reward_step": 0.09878277778625488,
      "rewards/format_reward_step": 0.203125,
      "step": 16
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.518870398402214,
      "aux_distill/mean_u": 0.30059635668181844,
      "aux_distill/n_active_tok": 112.71875,
      "calib/answer_extract_rate": 0.359375,
      "calib/auroc": 0.546875,
      "calib/avg_num_step_conf": 1.77734375,
      "calib/ece": 0.5814285714285714,
      "calib/final_conf_rate": 0.30078125,
      "calib/format_rate": 0.24609375,
      "calib/frac_conf_gt_0.9": 0.4155844155844156,
      "calib/gap": 0.050576923076923186,
      "calib/mean_conf": 0.711038961038961,
      "calib/mu_c": 0.7530769230769232,
      "calib/mu_w": 0.7025,
      "calib/nonempty_final_conf_rate": 0.30078125,
      "calib/nonempty_reasoning_rate": 0.46875,
      "calib/nonempty_step_conf_rate": 0.39453125,
      "calib/pce": 0.5618181818181818,
      "calib/std_conf": 0.32062270749271565,
      "calib/step_conf_rate": 0.39453125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 2799.0,
      "completions/max_terminated_length": 2799.0,
      "completions/mean_length": 539.28515625,
      "completions/mean_terminated_length": 565.807373046875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.018133333333333335,
      "grad_norm": 0.011570603586733341,
      "learning_rate": 4.25e-06,
      "loss": 0.1764,
      "num_tokens": 4512888.0,
      "reward": 0.24242910742759705,
      "reward_std": 0.3743348717689514,
      "rewards/accuracy_reward_step": 0.0625,
      "rewards/final_brier_reward_step": 0.1137644499540329,
      "rewards/format_reward_step": 0.24609375,
      "step": 17
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5079508516937494,
      "aux_distill/mean_u": 0.3073023040476281,
      "aux_distill/n_active_tok": 135.84375,
      "calib/answer_extract_rate": 0.39453125,
      "calib/auroc": 0.47294117647058825,
      "calib/avg_num_step_conf": 2.09765625,
      "calib/ece": 0.5014130434782609,
      "calib/final_conf_rate": 0.359375,
      "calib/format_rate": 0.3125,
      "calib/frac_conf_gt_0.9": 0.33695652173913043,
      "calib/gap": -0.026847058823529557,
      "calib/mean_conf": 0.6542391304347825,
      "calib/mu_c": 0.6323529411764705,
      "calib/mu_w": 0.6592,
      "calib/nonempty_final_conf_rate": 0.359375,
      "calib/nonempty_reasoning_rate": 0.4921875,
      "calib/nonempty_step_conf_rate": 0.4296875,
      "calib/pce": 0.48543478260869566,
      "calib/std_conf": 0.32658055033959604,
      "calib/step_conf_rate": 0.4296875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03515625,
      "completions/max_length": 2965.0,
      "completions/max_terminated_length": 2965.0,
      "completions/mean_length": 593.0625,
      "completions/mean_terminated_length": 614.6720581054688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0192,
      "grad_norm": 0.011249953880906105,
      "learning_rate": 4.5e-06,
      "loss": 0.1976,
      "num_tokens": 4775432.0,
      "reward": 0.3123294711112976,
      "reward_std": 0.43367162346839905,
      "rewards/accuracy_reward_step": 0.0703125,
      "rewards/final_brier_reward_step": 0.1715339869260788,
      "rewards/format_reward_step": 0.3125,
      "step": 18
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.555460131727159,
      "aux_distill/mean_u": 0.351152560522153,
      "aux_distill/n_active_tok": 189.0625,
      "calib/answer_extract_rate": 0.6015625,
      "calib/auroc": 0.5336134453781513,
      "calib/avg_num_step_conf": 2.86328125,
      "calib/ece": 0.44894987468671677,
      "calib/final_conf_rate": 0.51953125,
      "calib/format_rate": 0.4140625,
      "calib/frac_conf_gt_0.9": 0.21052631578947367,
      "calib/gap": 0.017854341736694757,
      "calib/mean_conf": 0.5275964912280702,
      "calib/mu_c": 0.5435714285714286,
      "calib/mu_w": 0.5257170868347338,
      "calib/nonempty_final_conf_rate": 0.51953125,
      "calib/nonempty_reasoning_rate": 0.80078125,
      "calib/nonempty_step_conf_rate": 0.65625,
      "calib/pce": 0.43564160401002505,
      "calib/std_conf": 0.3389578727804567,
      "calib/step_conf_rate": 0.65625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3045.0,
      "completions/max_terminated_length": 3045.0,
      "completions/mean_length": 441.5,
      "completions/mean_terminated_length": 455.7419128417969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.020266666666666665,
      "grad_norm": 0.012942974455654621,
      "learning_rate": 4.75e-06,
      "loss": 0.2163,
      "num_tokens": 4993216.0,
      "reward": 0.385407954454422,
      "reward_std": 0.459591805934906,
      "rewards/accuracy_reward_step": 0.05859375,
      "rewards/final_brier_reward_step": 0.23956595361232758,
      "rewards/format_reward_step": 0.4140625,
      "step": 19
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5594816422089934,
      "aux_distill/mean_u": 0.3344566962682378,
      "aux_distill/n_active_tok": 277.34375,
      "calib/answer_extract_rate": 0.74609375,
      "calib/auroc": 0.5335101679929265,
      "calib/avg_num_step_conf": 4.0703125,
      "calib/ece": 0.31548260869565214,
      "calib/final_conf_rate": 0.71875,
      "calib/format_rate": 0.64453125,
      "calib/frac_conf_gt_0.9": 0.1358695652173913,
      "calib/gap": 0.03500841732979665,
      "calib/mean_conf": 0.43625543478260864,
      "calib/mu_c": 0.46384358974358975,
      "calib/mu_w": 0.4288351724137931,
      "calib/nonempty_final_conf_rate": 0.71875,
      "calib/nonempty_reasoning_rate": 0.89453125,
      "calib/nonempty_step_conf_rate": 0.8359375,
      "calib/pce": 0.2698907608695652,
      "calib/std_conf": 0.33988152375181335,
      "calib/step_conf_rate": 0.8359375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2946.0,
      "completions/max_terminated_length": 2946.0,
      "completions/mean_length": 394.421875,
      "completions/mean_terminated_length": 395.9686584472656,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 16.0,
      "epoch": 0.021333333333333333,
      "grad_norm": 0.013653420843183994,
      "learning_rate": 5e-06,
      "loss": 0.1654,
      "num_tokens": 5199060.0,
      "reward": 0.7013116478919983,
      "reward_std": 0.555061936378479,
      "rewards/accuracy_reward_step": 0.16015625,
      "rewards/final_brier_reward_step": 0.4377795457839966,
      "rewards/format_reward_step": 0.64453125,
      "step": 20
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5351730575785041,
      "aux_distill/mean_u": 0.34520445152837353,
      "aux_distill/n_active_tok": 267.0,
      "calib/answer_extract_rate": 0.78515625,
      "calib/auroc": 0.4835216572504708,
      "calib/avg_num_step_conf": 4.0859375,
      "calib/ece": 0.34985621890547264,
      "calib/final_conf_rate": 0.78515625,
      "calib/format_rate": 0.72265625,
      "calib/frac_conf_gt_0.9": 0.1691542288557214,
      "calib/gap": -0.02221129943502831,
      "calib/mean_conf": 0.42572587064676615,
      "calib/mu_c": 0.4061666666666666,
      "calib/mu_w": 0.42837796610169493,
      "calib/nonempty_final_conf_rate": 0.78515625,
      "calib/nonempty_reasoning_rate": 0.9375,
      "calib/nonempty_step_conf_rate": 0.90234375,
      "calib/pce": 0.328089552238806,
      "calib/std_conf": 0.35222877223109705,
      "calib/step_conf_rate": 0.90234375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3019.0,
      "completions/max_terminated_length": 3019.0,
      "completions/mean_length": 370.9453125,
      "completions/mean_terminated_length": 373.86614990234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0224,
      "grad_norm": 0.013144408352673054,
      "learning_rate": 4.9722222222222224e-06,
      "loss": 0.2733,
      "num_tokens": 5396982.0,
      "reward": 0.7004989385604858,
      "reward_std": 0.48331907391548157,
      "rewards/accuracy_reward_step": 0.09375,
      "rewards/final_brier_reward_step": 0.49084165692329407,
      "rewards/format_reward_step": 0.72265625,
      "step": 21
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5181534392759204,
      "aux_distill/mean_u": 0.36928614539473653,
      "aux_distill/n_active_tok": 273.3125,
      "calib/answer_extract_rate": 0.875,
      "calib/auroc": 0.5354075514503321,
      "calib/avg_num_step_conf": 4.1640625,
      "calib/ece": 0.3315731818181818,
      "calib/final_conf_rate": 0.859375,
      "calib/format_rate": 0.79296875,
      "calib/frac_conf_gt_0.9": 0.15454545454545454,
      "calib/gap": 0.03293814616755791,
      "calib/mean_conf": 0.42260863636363644,
      "calib/mu_c": 0.4506060606060606,
      "calib/mu_w": 0.4176679144385027,
      "calib/nonempty_final_conf_rate": 0.859375,
      "calib/nonempty_reasoning_rate": 0.96484375,
      "calib/nonempty_step_conf_rate": 0.9375,
      "calib/pce": 0.3020909090909091,
      "calib/std_conf": 0.340468878390795,
      "calib/step_conf_rate": 0.9375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2813.0,
      "completions/max_terminated_length": 2813.0,
      "completions/mean_length": 298.42578125,
      "completions/mean_terminated_length": 299.5960998535156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 15.0,
      "epoch": 0.023466666666666667,
      "grad_norm": 0.012655031867325306,
      "learning_rate": 4.944444444444445e-06,
      "loss": 0.0866,
      "num_tokens": 5575195.0,
      "reward": 0.8091020584106445,
      "reward_std": 0.4806508421897888,
      "rewards/accuracy_reward_step": 0.13671875,
      "rewards/final_brier_reward_step": 0.5517978072166443,
      "rewards/format_reward_step": 0.79296875,
      "step": 22
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5239614294841886,
      "aux_distill/mean_u": 0.32456098856974674,
      "aux_distill/n_active_tok": 268.9375,
      "calib/answer_extract_rate": 0.85546875,
      "calib/auroc": 0.48690932311621965,
      "calib/avg_num_step_conf": 4.1484375,
      "calib/ece": 0.2829214611872146,
      "calib/final_conf_rate": 0.85546875,
      "calib/format_rate": 0.80859375,
      "calib/frac_conf_gt_0.9": 0.0867579908675799,
      "calib/gap": -0.0038183908045976156,
      "calib/mean_conf": 0.36036712328767123,
      "calib/mu_c": 0.35733333333333345,
      "calib/mu_w": 0.36115172413793106,
      "calib/nonempty_final_conf_rate": 0.85546875,
      "calib/nonempty_reasoning_rate": 0.95703125,
      "calib/nonempty_step_conf_rate": 0.9296875,
      "calib/pce": 0.21890456621004564,
      "calib/std_conf": 0.3172734094783932,
      "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": 0.0,
      "completions/max_length": 2549.0,
      "completions/max_terminated_length": 2549.0,
      "completions/mean_length": 313.2109375,
      "completions/mean_terminated_length": 313.2109375,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.024533333333333334,
      "grad_norm": 0.011424724943935871,
      "learning_rate": 4.9166666666666665e-06,
      "loss": 0.1664,
      "num_tokens": 5759313.0,
      "reward": 0.867918848991394,
      "reward_std": 0.5167853236198425,
      "rewards/accuracy_reward_step": 0.17578125,
      "rewards/final_brier_reward_step": 0.5756815075874329,
      "rewards/format_reward_step": 0.80859375,
      "step": 23
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5008599422872066,
      "aux_distill/mean_u": 0.3815701796434009,
      "aux_distill/n_active_tok": 279.75,
      "calib/answer_extract_rate": 0.8828125,
      "calib/auroc": 0.46545454545454545,
      "calib/avg_num_step_conf": 4.35546875,
      "calib/ece": 0.332652466367713,
      "calib/final_conf_rate": 0.87109375,
      "calib/format_rate": 0.8125,
      "calib/frac_conf_gt_0.9": 0.1210762331838565,
      "calib/gap": -0.03694595959595959,
      "calib/mean_conf": 0.3912040358744394,
      "calib/mu_c": 0.35840000000000005,
      "calib/mu_w": 0.39534595959595964,
      "calib/nonempty_final_conf_rate": 0.87109375,
      "calib/nonempty_reasoning_rate": 0.98046875,
      "calib/nonempty_step_conf_rate": 0.94140625,
      "calib/pce": 0.30587443946188336,
      "calib/std_conf": 0.3324925950580931,
      "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": 0.00390625,
      "completions/max_length": 2154.0,
      "completions/max_terminated_length": 2154.0,
      "completions/mean_length": 304.56640625,
      "completions/mean_terminated_length": 305.76080322265625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 24.0,
      "epoch": 0.0256,
      "grad_norm": 0.012385115027427673,
      "learning_rate": 4.888888888888889e-06,
      "loss": 0.1908,
      "num_tokens": 5941794.0,
      "reward": 0.7888060808181763,
      "reward_std": 0.4247359037399292,
      "rewards/accuracy_reward_step": 0.09765625,
      "rewards/final_brier_reward_step": 0.5697996616363525,
      "rewards/format_reward_step": 0.8125,
      "step": 24
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48966554924845695,
      "aux_distill/mean_u": 0.34171363359404794,
      "aux_distill/n_active_tok": 260.9375,
      "calib/answer_extract_rate": 0.8828125,
      "calib/auroc": 0.4392361111111111,
      "calib/avg_num_step_conf": 4.08203125,
      "calib/ece": 0.25293288288288285,
      "calib/final_conf_rate": 0.8671875,
      "calib/format_rate": 0.8046875,
      "calib/frac_conf_gt_0.9": 0.07657657657657657,
      "calib/gap": -0.07746406250000001,
      "calib/mean_conf": 0.29932927927927927,
      "calib/mu_c": 0.2323333333333333,
      "calib/mu_w": 0.3097973958333333,
      "calib/nonempty_final_conf_rate": 0.8671875,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.95703125,
      "calib/pce": 0.2085635135135135,
      "calib/std_conf": 0.301357494295312,
      "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": 0.0078125,
      "completions/max_length": 2563.0,
      "completions/max_terminated_length": 2563.0,
      "completions/mean_length": 267.8515625,
      "completions/mean_terminated_length": 269.96063232421875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 25.0,
      "epoch": 0.02666666666666667,
      "grad_norm": 0.011866779066622257,
      "learning_rate": 4.861111111111111e-06,
      "loss": 0.1395,
      "num_tokens": 6113588.0,
      "reward": 0.8221957683563232,
      "reward_std": 0.45238590240478516,
      "rewards/accuracy_reward_step": 0.12109375,
      "rewards/final_brier_reward_step": 0.5975164175033569,
      "rewards/format_reward_step": 0.8046875,
      "step": 25
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.508696723729372,
      "aux_distill/mean_u": 0.43063295485899017,
      "aux_distill/n_active_tok": 251.625,
      "calib/answer_extract_rate": 0.95703125,
      "calib/auroc": 0.5144890260631001,
      "calib/avg_num_step_conf": 3.91796875,
      "calib/ece": 0.2462139917695473,
      "calib/final_conf_rate": 0.94921875,
      "calib/format_rate": 0.9140625,
      "calib/frac_conf_gt_0.9": 0.053497942386831275,
      "calib/gap": 0.015416666666666634,
      "calib/mean_conf": 0.29962962962962963,
      "calib/mu_c": 0.31333333333333335,
      "calib/mu_w": 0.2979166666666667,
      "calib/nonempty_final_conf_rate": 0.94921875,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.2173662551440329,
      "calib/std_conf": 0.2912303026184558,
      "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": 0.00390625,
      "completions/max_length": 713.0,
      "completions/max_terminated_length": 713.0,
      "completions/mean_length": 214.05859375,
      "completions/mean_terminated_length": 214.89805603027344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 35.0,
      "epoch": 0.027733333333333332,
      "grad_norm": 0.01172387320548296,
      "learning_rate": 4.833333333333333e-06,
      "loss": 0.0078,
      "num_tokens": 6273627.0,
      "reward": 0.9250505566596985,
      "reward_std": 0.32336753606796265,
      "rewards/accuracy_reward_step": 0.109375,
      "rewards/final_brier_reward_step": 0.7172887325286865,
      "rewards/format_reward_step": 0.9140625,
      "step": 26
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.49942528642714024,
      "aux_distill/mean_u": 0.3542849390346178,
      "aux_distill/n_active_tok": 236.875,
      "calib/answer_extract_rate": 0.953125,
      "calib/auroc": 0.4569633995037221,
      "calib/avg_num_step_conf": 3.609375,
      "calib/ece": 0.21559665271966524,
      "calib/final_conf_rate": 0.93359375,
      "calib/format_rate": 0.89453125,
      "calib/frac_conf_gt_0.9": 0.0041841004184100415,
      "calib/gap": -0.03214708436724567,
      "calib/mean_conf": 0.24733221757322174,
      "calib/mu_c": 0.21935483870967737,
      "calib/mu_w": 0.25150192307692304,
      "calib/nonempty_final_conf_rate": 0.93359375,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.95703125,
      "calib/pce": 0.16661087866108784,
      "calib/std_conf": 0.2538213858358564,
      "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": 987.0,
      "completions/max_terminated_length": 987.0,
      "completions/mean_length": 221.66015625,
      "completions/mean_terminated_length": 222.5294189453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 44.0,
      "epoch": 0.0288,
      "grad_norm": 0.012423808686435223,
      "learning_rate": 4.805555555555556e-06,
      "loss": 0.087,
      "num_tokens": 6435588.0,
      "reward": 0.9256427884101868,
      "reward_std": 0.36195361614227295,
      "rewards/accuracy_reward_step": 0.12109375,
      "rewards/final_brier_reward_step": 0.7145668268203735,
      "rewards/format_reward_step": 0.89453125,
      "step": 27
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47512149065732956,
      "aux_distill/mean_u": 0.34499709554861546,
      "aux_distill/n_active_tok": 214.125,
      "calib/answer_extract_rate": 0.95703125,
      "calib/auroc": 0.5100290697674419,
      "calib/avg_num_step_conf": 3.36328125,
      "calib/ece": 0.18277894736842104,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 0.02834008097165992,
      "calib/gap": 0.0035413953488372396,
      "calib/mean_conf": 0.2619174089068826,
      "calib/mu_c": 0.265,
      "calib/mu_w": 0.2614586046511628,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.15757085020242914,
      "calib/std_conf": 0.24821793492654226,
      "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": 1773.0,
      "completions/max_terminated_length": 1773.0,
      "completions/mean_length": 207.30078125,
      "completions/mean_terminated_length": 207.30078125,
      "completions/min_length": 21.0,
      "completions/min_terminated_length": 21.0,
      "epoch": 0.029866666666666666,
      "grad_norm": 0.013800237327814102,
      "learning_rate": 4.777777777777778e-06,
      "loss": 0.082,
      "num_tokens": 6595601.0,
      "reward": 0.9762635231018066,
      "reward_std": 0.2939887046813965,
      "rewards/accuracy_reward_step": 0.12890625,
      "rewards/final_brier_reward_step": 0.7572145462036133,
      "rewards/format_reward_step": 0.9375,
      "step": 28
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5141920652240515,
      "aux_distill/mean_u": 0.3670624937354536,
      "aux_distill/n_active_tok": 226.0,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.5522705314009662,
      "calib/avg_num_step_conf": 3.4765625,
      "calib/ece": 0.22223467741935485,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.024193548387096774,
      "calib/gap": 0.06157283091787441,
      "calib/mean_conf": 0.26212016129032256,
      "calib/mu_c": 0.3179826086956522,
      "calib/mu_w": 0.2564097777777778,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.19580645161290325,
      "calib/std_conf": 0.25058815185795774,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 691.0,
      "completions/max_terminated_length": 691.0,
      "completions/mean_length": 203.31640625,
      "completions/mean_terminated_length": 204.11373901367188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 43.0,
      "epoch": 0.030933333333333334,
      "grad_norm": 0.011709311045706272,
      "learning_rate": 4.75e-06,
      "loss": 0.0294,
      "num_tokens": 6754778.0,
      "reward": 0.9514590501785278,
      "reward_std": 0.2994237542152405,
      "rewards/accuracy_reward_step": 0.08984375,
      "rewards/final_brier_reward_step": 0.7818243503570557,
      "rewards/format_reward_step": 0.94140625,
      "step": 29
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5036658011376858,
      "aux_distill/mean_u": 0.3104102361376659,
      "aux_distill/n_active_tok": 199.09375,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.6265198237885462,
      "calib/avg_num_step_conf": 3.078125,
      "calib/ece": 0.1319047619047619,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.007936507936507936,
      "calib/gap": 0.0725955947136564,
      "calib/mean_conf": 0.22420634920634924,
      "calib/mu_c": 0.2896,
      "calib/mu_w": 0.21700440528634363,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.12845238095238096,
      "calib/std_conf": 0.2019914864409097,
      "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": 775.0,
      "completions/max_terminated_length": 775.0,
      "completions/mean_length": 170.47265625,
      "completions/mean_terminated_length": 171.1411895751953,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 48.0,
      "epoch": 0.032,
      "grad_norm": 0.012775780633091927,
      "learning_rate": 4.722222222222222e-06,
      "loss": 0.0336,
      "num_tokens": 6905403.0,
      "reward": 1.009130835533142,
      "reward_std": 0.21095620095729828,
      "rewards/accuracy_reward_step": 0.09765625,
      "rewards/final_brier_reward_step": 0.846386730670929,
      "rewards/format_reward_step": 0.9765625,
      "step": 30
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.49115212354809046,
      "aux_distill/mean_u": 0.32121877010109623,
      "aux_distill/n_active_tok": 214.40625,
      "calib/answer_extract_rate": 0.96484375,
      "calib/auroc": 0.5238933601609658,
      "calib/avg_num_step_conf": 3.328125,
      "calib/ece": 0.16140622406639005,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.92578125,
      "calib/frac_conf_gt_0.9": 0.029045643153526972,
      "calib/gap": 0.007251710261569483,
      "calib/mean_conf": 0.22216224066390042,
      "calib/mu_c": 0.22857142857142862,
      "calib/mu_w": 0.22131971830985914,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.1336929460580913,
      "calib/std_conf": 0.22215404253743912,
      "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": 600.0,
      "completions/max_terminated_length": 600.0,
      "completions/mean_length": 176.61328125,
      "completions/mean_terminated_length": 177.30589294433594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 5.0,
      "epoch": 0.03306666666666667,
      "grad_norm": 0.013856728561222553,
      "learning_rate": 4.694444444444445e-06,
      "loss": 0.0632,
      "num_tokens": 7056528.0,
      "reward": 0.9648450613021851,
      "reward_std": 0.2989380657672882,
      "rewards/accuracy_reward_step": 0.11328125,
      "rewards/final_brier_reward_step": 0.7773463129997253,
      "rewards/format_reward_step": 0.92578125,
      "step": 31
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5210744366049767,
      "aux_distill/mean_u": 0.33302298813774467,
      "aux_distill/n_active_tok": 206.09375,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5025111607142857,
      "calib/avg_num_step_conf": 3.140625,
      "calib/ece": 0.14899677419354837,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.004032258064516129,
      "calib/gap": -0.005524404761904772,
      "calib/mean_conf": 0.18540645161290323,
      "calib/mu_c": 0.18041666666666667,
      "calib/mu_w": 0.18594107142857144,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.11881451612903227,
      "calib/std_conf": 0.19145325031231744,
      "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": 1639.0,
      "completions/max_terminated_length": 1639.0,
      "completions/mean_length": 175.671875,
      "completions/mean_terminated_length": 175.671875,
      "completions/min_length": 46.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.034133333333333335,
      "grad_norm": 0.013500292785465717,
      "learning_rate": 4.666666666666667e-06,
      "loss": 0.0807,
      "num_tokens": 7208204.0,
      "reward": 0.9942401647567749,
      "reward_std": 0.23118172585964203,
      "rewards/accuracy_reward_step": 0.09375,
      "rewards/final_brier_reward_step": 0.8361366391181946,
      "rewards/format_reward_step": 0.96484375,
      "step": 32
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5026440424844623,
      "aux_distill/mean_u": 0.33876221840326814,
      "aux_distill/n_active_tok": 188.5,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5013993224333481,
      "calib/avg_num_step_conf": 2.9453125,
      "calib/ece": 0.12772,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.008,
      "calib/gap": 0.007213138901163707,
      "calib/mean_conf": 0.18852000000000002,
      "calib/mu_c": 0.19483870967741942,
      "calib/mu_w": 0.18762557077625572,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.09612000000000001,
      "calib/std_conf": 0.1957054153568572,
      "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": 1502.0,
      "completions/max_terminated_length": 1502.0,
      "completions/mean_length": 166.40234375,
      "completions/mean_terminated_length": 166.40234375,
      "completions/min_length": 46.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.0352,
      "grad_norm": 0.013900885358452797,
      "learning_rate": 4.638888888888889e-06,
      "loss": 0.0665,
      "num_tokens": 7357675.0,
      "reward": 1.0091570615768433,
      "reward_std": 0.23295243084430695,
      "rewards/accuracy_reward_step": 0.12109375,
      "rewards/final_brier_reward_step": 0.8151890635490417,
      "rewards/format_reward_step": 0.9609375,
      "step": 33
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.49193944688886404,
      "aux_distill/mean_u": 0.3246782472719872,
      "aux_distill/n_active_tok": 186.25,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.4617512142471667,
      "calib/avg_num_step_conf": 2.92578125,
      "calib/ece": 0.15749999999999997,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.03515650296815975,
      "calib/mean_conf": 0.1901190476190476,
      "calib/mu_c": 0.15970588235294117,
      "calib/mu_w": 0.19486238532110092,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.10634920634920635,
      "calib/std_conf": 0.17530128745454196,
      "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": 497.0,
      "completions/max_terminated_length": 497.0,
      "completions/mean_length": 141.4296875,
      "completions/mean_terminated_length": 141.9843292236328,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 24.0,
      "epoch": 0.03626666666666667,
      "grad_norm": 0.015190926380455494,
      "learning_rate": 4.611111111111112e-06,
      "loss": 0.0128,
      "num_tokens": 7498993.0,
      "reward": 1.024768352508545,
      "reward_std": 0.22394582629203796,
      "rewards/accuracy_reward_step": 0.1328125,
      "rewards/final_brier_reward_step": 0.8151617050170898,
      "rewards/format_reward_step": 0.96875,
      "step": 34
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4921773290261626,
      "aux_distill/mean_u": 0.33049255483384,
      "aux_distill/n_active_tok": 191.78125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4441244239631337,
      "calib/avg_num_step_conf": 3.0,
      "calib/ece": 0.14278431372549016,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.037167338709677394,
      "calib/mean_conf": 0.17329411764705882,
      "calib/mu_c": 0.1406451612903226,
      "calib/mu_w": 0.17781249999999998,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.09725490196078432,
      "calib/std_conf": 0.1807460892769947,
      "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": 427.0,
      "completions/max_terminated_length": 427.0,
      "completions/mean_length": 144.484375,
      "completions/mean_terminated_length": 145.05099487304688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 45.0,
      "epoch": 0.037333333333333336,
      "grad_norm": 0.01703733019530773,
      "learning_rate": 4.583333333333333e-06,
      "loss": 0.0132,
      "num_tokens": 7645237.0,
      "reward": 1.0347568988800049,
      "reward_std": 0.16965264081954956,
      "rewards/accuracy_reward_step": 0.12109375,
      "rewards/final_brier_reward_step": 0.8390449285507202,
      "rewards/format_reward_step": 0.98828125,
      "step": 35
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.489220367744565,
      "aux_distill/mean_u": 0.3329217827101059,
      "aux_distill/n_active_tok": 191.78125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.47716346153846156,
      "calib/avg_num_step_conf": 3.0,
      "calib/ece": 0.13984251968503938,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": -0.026258361204013342,
      "calib/mean_conf": 0.15976377952755905,
      "calib/mu_c": 0.1382608695652174,
      "calib/mu_w": 0.16451923076923075,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.05925196850393701,
      "calib/std_conf": 0.1777725901431028,
      "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": 1750.0,
      "completions/max_terminated_length": 1750.0,
      "completions/mean_length": 153.39453125,
      "completions/mean_terminated_length": 153.39453125,
      "completions/min_length": 48.0,
      "completions/min_terminated_length": 48.0,
      "epoch": 0.0384,
      "grad_norm": 0.012923480942845345,
      "learning_rate": 4.555555555555556e-06,
      "loss": 0.0928,
      "num_tokens": 7787218.0,
      "reward": 1.0785343647003174,
      "reward_std": 0.21512489020824432,
      "rewards/accuracy_reward_step": 0.1796875,
      "rewards/final_brier_reward_step": 0.8055062294006348,
      "rewards/format_reward_step": 0.9921875,
      "step": 36
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4732179343700409,
      "aux_distill/mean_u": 0.2754847840077296,
      "aux_distill/n_active_tok": 194.5625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5553020767778477,
      "calib/avg_num_step_conf": 3.015625,
      "calib/ece": 0.08060705882352943,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.020079358086847063,
      "calib/mean_conf": 0.11319686274509805,
      "calib/mu_c": 0.13107142857142856,
      "calib/mu_w": 0.1109920704845815,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.04200000000000001,
      "calib/std_conf": 0.1273543914035377,
      "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": 591.0,
      "completions/max_terminated_length": 591.0,
      "completions/mean_length": 151.015625,
      "completions/mean_terminated_length": 151.60784912109375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 45.0,
      "epoch": 0.039466666666666664,
      "grad_norm": 0.01448907982558012,
      "learning_rate": 4.527777777777778e-06,
      "loss": 0.0758,
      "num_tokens": 7932974.0,
      "reward": 1.0467637777328491,
      "reward_std": 0.1426790952682495,
      "rewards/accuracy_reward_step": 0.109375,
      "rewards/final_brier_reward_step": 0.8825901746749878,
      "rewards/format_reward_step": 0.9921875,
      "step": 37
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.524141519330442,
      "aux_distill/mean_u": 0.37779598700656625,
      "aux_distill/n_active_tok": 219.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5299713584288053,
      "calib/avg_num_step_conf": 3.453125,
      "calib/ece": 0.11498039215686277,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.010617839607201293,
      "calib/mean_conf": 0.09176470588235294,
      "calib/mu_c": 0.10042553191489362,
      "calib/mu_w": 0.08980769230769232,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.011215686274509803,
      "calib/std_conf": 0.10985384019371217,
      "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": 509.0,
      "completions/max_terminated_length": 509.0,
      "completions/mean_length": 159.3046875,
      "completions/mean_terminated_length": 159.92942810058594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 45.0,
      "epoch": 0.04053333333333333,
      "grad_norm": 0.013505329377949238,
      "learning_rate": 4.5e-06,
      "loss": 0.0494,
      "num_tokens": 8080644.0,
      "reward": 1.09230375289917,
      "reward_std": 0.17799054086208344,
      "rewards/accuracy_reward_step": 0.18359375,
      "rewards/final_brier_reward_step": 0.8252324461936951,
      "rewards/format_reward_step": 0.9921875,
      "step": 38
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4891305109485984,
      "aux_distill/mean_u": 0.30491045657396,
      "aux_distill/n_active_tok": 212.78125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.48369210697977827,
      "calib/avg_num_step_conf": 3.33203125,
      "calib/ece": 0.10676574803149609,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.007805283757338563,
      "calib/mean_conf": 0.06669881889763779,
      "calib/mu_c": 0.07342857142857144,
      "calib/mu_w": 0.06562328767123288,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.017834645669291337,
      "calib/std_conf": 0.09622918446857423,
      "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": 703.0,
      "completions/max_terminated_length": 703.0,
      "completions/mean_length": 161.66796875,
      "completions/mean_terminated_length": 162.30197143554688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 47.0,
      "epoch": 0.0416,
      "grad_norm": 0.014169767498970032,
      "learning_rate": 4.472222222222223e-06,
      "loss": 0.0422,
      "num_tokens": 8228119.0,
      "reward": 1.0637850761413574,
      "reward_std": 0.1445343792438507,
      "rewards/accuracy_reward_step": 0.140625,
      "rewards/final_brier_reward_step": 0.8580389022827148,
      "rewards/format_reward_step": 0.98828125,
      "step": 39
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47824777383357286,
      "aux_distill/mean_u": 0.2888125463333283,
      "aux_distill/n_active_tok": 234.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.506860465116279,
      "calib/avg_num_step_conf": 3.66796875,
      "calib/ece": 0.1305270588235294,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00796813953488372,
      "calib/mean_conf": 0.05253176470588236,
      "calib/mu_c": 0.059250000000000004,
      "calib/mu_w": 0.051281860465116284,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.013098039215686275,
      "calib/std_conf": 0.09134577386375445,
      "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": 470.0,
      "completions/max_terminated_length": 470.0,
      "completions/mean_length": 166.625,
      "completions/mean_terminated_length": 167.2784423828125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 5.0,
      "epoch": 0.042666666666666665,
      "grad_norm": 0.012960164807736874,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.0551,
      "num_tokens": 8377535.0,
      "reward": 1.077946424484253,
      "reward_std": 0.1492486298084259,
      "rewards/accuracy_reward_step": 0.15625,
      "rewards/final_brier_reward_step": 0.8472990989685059,
      "rewards/format_reward_step": 0.99609375,
      "step": 40
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4979221150279045,
      "aux_distill/mean_u": 0.30807657441087855,
      "aux_distill/n_active_tok": 227.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.521875,
      "calib/avg_num_step_conf": 3.56640625,
      "calib/ece": 0.28539062499999995,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.005045454545454547,
      "calib/mean_conf": 0.03484375000000001,
      "calib/mu_c": 0.031375,
      "calib/mu_w": 0.03642045454545455,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0038671875,
      "calib/std_conf": 0.05553692992898959,
      "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": 457.0,
      "completions/max_terminated_length": 457.0,
      "completions/mean_length": 169.59765625,
      "completions/mean_terminated_length": 170.26275634765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 49.0,
      "epoch": 0.04373333333333333,
      "grad_norm": 0.014377506449818611,
      "learning_rate": 4.416666666666667e-06,
      "loss": 0.0596,
      "num_tokens": 8528200.0,
      "reward": 1.1560933589935303,
      "reward_std": 0.20653408765792847,
      "rewards/accuracy_reward_step": 0.3125,
      "rewards/final_brier_reward_step": 0.6949992179870605,
      "rewards/format_reward_step": 0.9921875,
      "step": 41
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4956301599740982,
      "aux_distill/mean_u": 0.32941970174270324,
      "aux_distill/n_active_tok": 229.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.48813964610234334,
      "calib/avg_num_step_conf": 3.6015625,
      "calib/ece": 0.18374999999999997,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00886848397895744,
      "calib/mean_conf": 0.029062500000000005,
      "calib/mu_c": 0.02196078431372549,
      "calib/mu_w": 0.03082926829268293,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.006796875,
      "calib/std_conf": 0.05621439845582269,
      "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": 426.0,
      "completions/max_terminated_length": 426.0,
      "completions/mean_length": 168.703125,
      "completions/mean_terminated_length": 169.36471557617188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 52.0,
      "epoch": 0.0448,
      "grad_norm": 0.013974346220493317,
      "learning_rate": 4.388888888888889e-06,
      "loss": 0.0691,
      "num_tokens": 8675756.0,
      "reward": 1.1019821166992188,
      "reward_std": 0.13847261667251587,
      "rewards/accuracy_reward_step": 0.19921875,
      "rewards/final_brier_reward_step": 0.8055265545845032,
      "rewards/format_reward_step": 1.0,
      "step": 42
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47923671174794436,
      "aux_distill/mean_u": 0.28819438981801476,
      "aux_distill/n_active_tok": 263.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6338541666666666,
      "calib/avg_num_step_conf": 4.171875,
      "calib/ece": 0.142558203125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.011254166666666666,
      "calib/mean_conf": 0.021504296875000002,
      "calib/mu_c": 0.031,
      "calib/mu_w": 0.019745833333333334,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.00390625,
      "calib/std_conf": 0.04623921771585637,
      "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": 773.0,
      "completions/max_terminated_length": 773.0,
      "completions/mean_length": 197.10546875,
      "completions/mean_terminated_length": 197.87844848632812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 57.0,
      "epoch": 0.04586666666666667,
      "grad_norm": 0.010071014054119587,
      "learning_rate": 4.361111111111112e-06,
      "loss": 0.0641,
      "num_tokens": 8831439.0,
      "reward": 1.0777781009674072,
      "reward_std": 0.115573450922966,
      "rewards/accuracy_reward_step": 0.15625,
      "rewards/final_brier_reward_step": 0.8469623923301697,
      "rewards/format_reward_step": 0.99609375,
      "step": 43
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47694727033376694,
      "aux_distill/mean_u": 0.32135310640373566,
      "aux_distill/n_active_tok": 328.0,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.47371822490305904,
      "calib/avg_num_step_conf": 5.125,
      "calib/ece": 0.16690196078431377,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.006278543731150367,
      "calib/mean_conf": 0.019058823529411767,
      "calib/mu_c": 0.013863636363636363,
      "calib/mu_w": 0.02014218009478673,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.006705882352941176,
      "calib/std_conf": 0.054864328839493465,
      "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": 2174.0,
      "completions/max_terminated_length": 2174.0,
      "completions/mean_length": 227.578125,
      "completions/mean_terminated_length": 227.578125,
      "completions/min_length": 79.0,
      "completions/min_terminated_length": 79.0,
      "epoch": 0.046933333333333334,
      "grad_norm": 0.01117937732487917,
      "learning_rate": 4.333333333333334e-06,
      "loss": 0.0676,
      "num_tokens": 8996019.0,
      "reward": 1.0788277387619019,
      "reward_std": 0.14245754480361938,
      "rewards/accuracy_reward_step": 0.171875,
      "rewards/final_brier_reward_step": 0.8217179775238037,
      "rewards/format_reward_step": 0.9921875,
      "step": 44
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4636860592290759,
      "aux_distill/mean_u": 0.25446549858857836,
      "aux_distill/n_active_tok": 304.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4781467779402908,
      "calib/avg_num_step_conf": 4.75,
      "calib/ece": 0.2226171875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0015082164673492198,
      "calib/mean_conf": 0.0098046875,
      "calib/mu_c": 0.00864406779661017,
      "calib/mu_w": 0.01015228426395939,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0009765625,
      "calib/std_conf": 0.016804277075415763,
      "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": 627.0,
      "completions/max_terminated_length": 627.0,
      "completions/mean_length": 219.796875,
      "completions/mean_terminated_length": 220.65884399414062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 60.0,
      "epoch": 0.048,
      "grad_norm": 0.009709659963846207,
      "learning_rate": 4.305555555555556e-06,
      "loss": 0.042,
      "num_tokens": 9157335.0,
      "reward": 1.1131341457366943,
      "reward_std": 0.14094233512878418,
      "rewards/accuracy_reward_step": 0.23046875,
      "rewards/final_brier_reward_step": 0.7692371010780334,
      "rewards/format_reward_step": 0.99609375,
      "step": 45
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48097407817840576,
      "aux_distill/mean_u": 0.3231606234081447,
      "aux_distill/n_active_tok": 330.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.48965356893745643,
      "calib/avg_num_step_conf": 5.18359375,
      "calib/ece": 0.265234375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0017329303262807105,
      "calib/mean_conf": 0.00546875,
      "calib/mu_c": 0.004202898550724638,
      "calib/mu_w": 0.005935828877005348,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0005859375,
      "calib/std_conf": 0.012708226604743088,
      "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": 758.0,
      "completions/max_terminated_length": 758.0,
      "completions/mean_length": 240.29296875,
      "completions/mean_terminated_length": 241.2353057861328,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 50.0,
      "epoch": 0.04906666666666667,
      "grad_norm": 0.008790974505245686,
      "learning_rate": 4.277777777777778e-06,
      "loss": 0.0275,
      "num_tokens": 9323618.0,
      "reward": 1.1358027458190918,
      "reward_std": 0.12116257846355438,
      "rewards/accuracy_reward_step": 0.26953125,
      "rewards/final_brier_reward_step": 0.7325429916381836,
      "rewards/format_reward_step": 1.0,
      "step": 46
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4657612703740597,
      "aux_distill/mean_u": 0.2934879503928806,
      "aux_distill/n_active_tok": 380.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.3742272727272727,
      "calib/avg_num_step_conf": 5.96484375,
      "calib/ece": 0.21090196078431375,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.004786363636363637,
      "calib/mean_conf": 0.00611764705882353,
      "calib/mu_c": 0.0023636363636363638,
      "calib/mu_w": 0.007150000000000001,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0006666666666666668,
      "calib/std_conf": 0.0137849691641231,
      "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": 705.0,
      "completions/max_terminated_length": 705.0,
      "completions/mean_length": 259.48828125,
      "completions/mean_terminated_length": 260.5058898925781,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 37.0,
      "epoch": 0.050133333333333335,
      "grad_norm": 0.009199056774377823,
      "learning_rate": 4.25e-06,
      "loss": 0.0573,
      "num_tokens": 9496023.0,
      "reward": 1.0960979461669922,
      "reward_std": 0.16379356384277344,
      "rewards/accuracy_reward_step": 0.21484375,
      "rewards/final_brier_reward_step": 0.7742273211479187,
      "rewards/format_reward_step": 0.98828125,
      "step": 47
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.46220773458480835,
      "aux_distill/mean_u": 0.27937701966110146,
      "aux_distill/n_active_tok": 376.75,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.516267648864334,
      "calib/avg_num_step_conf": 5.890625,
      "calib/ece": 0.2817391304347826,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0006606813996316765,
      "calib/mean_conf": 0.004110671936758893,
      "calib/mu_c": 0.004583333333333333,
      "calib/mu_w": 0.003922651933701657,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0006324110671936759,
      "calib/std_conf": 0.013025499646864886,
      "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": 2414.0,
      "completions/max_terminated_length": 2414.0,
      "completions/mean_length": 266.04296875,
      "completions/mean_terminated_length": 266.04296875,
      "completions/min_length": 58.0,
      "completions/min_terminated_length": 58.0,
      "epoch": 0.0512,
      "grad_norm": 0.00915578380227089,
      "learning_rate": 4.222222222222223e-06,
      "loss": 0.1119,
      "num_tokens": 9667818.0,
      "reward": 1.1261968612670898,
      "reward_std": 0.1674359142780304,
      "rewards/accuracy_reward_step": 0.28125,
      "rewards/final_brier_reward_step": 0.7055187225341797,
      "rewards/format_reward_step": 0.984375,
      "step": 48
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4780575539916754,
      "aux_distill/mean_u": 0.3032898873526519,
      "aux_distill/n_active_tok": 370.03125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.460891197775664,
      "calib/avg_num_step_conf": 5.7734375,
      "calib/ece": 0.3107539682539682,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0006694958659544891,
      "calib/mean_conf": 0.0027380952380952383,
      "calib/mu_c": 0.002278481012658228,
      "calib/mu_w": 0.002947976878612717,
      "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": 0.007403000862400835,
      "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": 2160.0,
      "completions/max_terminated_length": 2160.0,
      "completions/mean_length": 269.3359375,
      "completions/mean_terminated_length": 269.3359375,
      "completions/min_length": 81.0,
      "completions/min_terminated_length": 81.0,
      "epoch": 0.05226666666666667,
      "grad_norm": 0.009693208150565624,
      "learning_rate": 4.194444444444445e-06,
      "loss": 0.122,
      "num_tokens": 9841304.0,
      "reward": 1.1432504653930664,
      "reward_std": 0.17814280092716217,
      "rewards/accuracy_reward_step": 0.3125,
      "rewards/final_brier_reward_step": 0.6771261692047119,
      "rewards/format_reward_step": 0.984375,
      "step": 49
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4636885020881891,
      "aux_distill/mean_u": 0.277049195947932,
      "aux_distill/n_active_tok": 394.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.49481781376518214,
      "calib/avg_num_step_conf": 6.1640625,
      "calib/ece": 0.2532941176470588,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0002995951417004048,
      "calib/mean_conf": 0.0016078431372549022,
      "calib/mu_c": 0.0013846153846153847,
      "calib/mu_w": 0.0016842105263157896,
      "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": 0.004266046667756052,
      "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": 2235.0,
      "completions/max_terminated_length": 2235.0,
      "completions/mean_length": 305.46875,
      "completions/mean_terminated_length": 305.46875,
      "completions/min_length": 85.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.05333333333333334,
      "grad_norm": 0.007432255428284407,
      "learning_rate": 4.166666666666667e-06,
      "loss": 0.077,
      "num_tokens": 10024864.0,
      "reward": 1.1233880519866943,
      "reward_std": 0.13776543736457825,
      "rewards/accuracy_reward_step": 0.25390625,
      "rewards/final_brier_reward_step": 0.7428699135780334,
      "rewards/format_reward_step": 0.99609375,
      "step": 50
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.44833634700626135,
      "aux_distill/mean_u": 0.2464335719181652,
      "aux_distill/n_active_tok": 393.0625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4710574113559188,
      "calib/avg_num_step_conf": 6.17578125,
      "calib/ece": 0.2601953125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0008505093579720445,
      "calib/mean_conf": 0.0015234375,
      "calib/mu_c": 0.0008955223880597015,
      "calib/mu_w": 0.001746031746031746,
      "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": 0.004721071190269615,
      "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": 1022.0,
      "completions/max_terminated_length": 1022.0,
      "completions/mean_length": 270.13671875,
      "completions/mean_terminated_length": 271.19610595703125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 79.0,
      "epoch": 0.0544,
      "grad_norm": 0.008282770402729511,
      "learning_rate": 4.138888888888889e-06,
      "loss": 0.0455,
      "num_tokens": 10203315.0,
      "reward": 1.131081461906433,
      "reward_std": 0.13202789425849915,
      "rewards/accuracy_reward_step": 0.26171875,
      "rewards/final_brier_reward_step": 0.7387253642082214,
      "rewards/format_reward_step": 1.0,
      "step": 51
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.46477444563061,
      "aux_distill/mean_u": 0.2906417075283816,
      "aux_distill/n_active_tok": 357.53125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4778928571428571,
      "calib/avg_num_step_conf": 5.65625,
      "calib/ece": 0.3128627450980392,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0007107142857142858,
      "calib/mean_conf": 0.0008627450980392157,
      "calib/mu_c": 0.000375,
      "calib/mu_w": 0.0010857142857142858,
      "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": 0.004606132348243458,
      "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": 923.0,
      "completions/max_terminated_length": 923.0,
      "completions/mean_length": 264.53515625,
      "completions/mean_terminated_length": 265.57257080078125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 74.0,
      "epoch": 0.055466666666666664,
      "grad_norm": 0.010578352957963943,
      "learning_rate": 4.111111111111111e-06,
      "loss": 0.0625,
      "num_tokens": 10378988.0,
      "reward": 1.1524499654769897,
      "reward_std": 0.18429045379161835,
      "rewards/accuracy_reward_step": 0.3125,
      "rewards/final_brier_reward_step": 0.6838062405586243,
      "rewards/format_reward_step": 0.99609375,
      "step": 52
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4559667222201824,
      "aux_distill/mean_u": 0.2902862957850312,
      "aux_distill/n_active_tok": 388.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.481833910034602,
      "calib/avg_num_step_conf": 6.0703125,
      "calib/ece": 0.3993725490196079,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0005555555555555557,
      "calib/mean_conf": 0.0006274509803921569,
      "calib/mu_c": 0.0002941176470588235,
      "calib/mu_w": 0.0008496732026143792,
      "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": 0.003003010152012023,
      "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": 2700.0,
      "completions/max_terminated_length": 2700.0,
      "completions/mean_length": 298.67578125,
      "completions/mean_terminated_length": 298.67578125,
      "completions/min_length": 89.0,
      "completions/min_terminated_length": 89.0,
      "epoch": 0.05653333333333333,
      "grad_norm": 0.008898251689970493,
      "learning_rate": 4.083333333333334e-06,
      "loss": 0.0719,
      "num_tokens": 10561273.0,
      "reward": 1.195425033569336,
      "reward_std": 0.18935687839984894,
      "rewards/accuracy_reward_step": 0.3984375,
      "rewards/final_brier_reward_step": 0.5978812575340271,
      "rewards/format_reward_step": 0.99609375,
      "step": 53
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4361731903627515,
      "aux_distill/mean_u": 0.22795934520457617,
      "aux_distill/n_active_tok": 370.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5071147725145104,
      "calib/avg_num_step_conf": 5.89453125,
      "calib/ece": 0.4253125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0001422954502902078,
      "calib/mean_conf": 0.00046875000000000004,
      "calib/mu_c": 0.0005504587155963302,
      "calib/mu_w": 0.0004081632653061224,
      "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": 0.0021137108216357317,
      "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": 958.0,
      "completions/max_terminated_length": 958.0,
      "completions/mean_length": 269.8671875,
      "completions/mean_terminated_length": 270.9255065917969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 86.0,
      "epoch": 0.0576,
      "grad_norm": 0.007963367737829685,
      "learning_rate": 4.055555555555556e-06,
      "loss": 0.0373,
      "num_tokens": 10736591.0,
      "reward": 1.2111694812774658,
      "reward_std": 0.13964581489562988,
      "rewards/accuracy_reward_step": 0.42578125,
      "rewards/final_brier_reward_step": 0.5746828317642212,
      "rewards/format_reward_step": 0.99609375,
      "step": 54
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4443762181326747,
      "aux_distill/mean_u": 0.27825471713899175,
      "aux_distill/n_active_tok": 381.5,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5139276877643867,
      "calib/avg_num_step_conf": 5.98828125,
      "calib/ece": 0.2632283464566929,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.000670444568600846,
      "calib/mean_conf": 0.0005511811023622047,
      "calib/mu_c": 0.001044776119402985,
      "calib/mu_w": 0.000374331550802139,
      "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": 0.0032741138403609114,
      "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": 2341.0,
      "completions/max_terminated_length": 2341.0,
      "completions/mean_length": 282.6328125,
      "completions/mean_terminated_length": 283.7411804199219,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 100.0,
      "epoch": 0.058666666666666666,
      "grad_norm": 0.009071978740394115,
      "learning_rate": 4.027777777777779e-06,
      "loss": 0.08,
      "num_tokens": 10916769.0,
      "reward": 1.1233148574829102,
      "reward_std": 0.1690075397491455,
      "rewards/accuracy_reward_step": 0.26171875,
      "rewards/final_brier_reward_step": 0.7310047149658203,
      "rewards/format_reward_step": 0.9921875,
      "step": 55
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4458486130461097,
      "aux_distill/mean_u": 0.28278279798028116,
      "aux_distill/n_active_tok": 379.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5131971051511282,
      "calib/avg_num_step_conf": 5.9296875,
      "calib/ece": 0.3173333333333333,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0002639421030225628,
      "calib/mean_conf": 0.00031372549019607844,
      "calib/mu_c": 0.0004938271604938272,
      "calib/mu_w": 0.00022988505747126436,
      "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": 0.0017432243741876759,
      "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": 934.0,
      "completions/max_terminated_length": 934.0,
      "completions/mean_length": 288.24609375,
      "completions/mean_terminated_length": 289.3764953613281,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 89.0,
      "epoch": 0.05973333333333333,
      "grad_norm": 0.007802541833370924,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0256,
      "num_tokens": 11097400.0,
      "reward": 1.154451608657837,
      "reward_std": 0.17325270175933838,
      "rewards/accuracy_reward_step": 0.31640625,
      "rewards/final_brier_reward_step": 0.6799968481063843,
      "rewards/format_reward_step": 0.99609375,
      "step": 56
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43617555499076843,
      "aux_distill/mean_u": 0.21965836460819538,
      "aux_distill/n_active_tok": 389.5,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5176817940552018,
      "calib/avg_num_step_conf": 6.0859375,
      "calib/ece": 0.37920948616600786,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00045713906581740975,
      "calib/mean_conf": 0.00023715415019762845,
      "calib/mu_c": 0.0005208333333333333,
      "calib/mu_w": 6.369426751592357e-05,
      "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": 0.001762331763605372,
      "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": 1835.0,
      "completions/max_terminated_length": 1835.0,
      "completions/mean_length": 297.125,
      "completions/mean_terminated_length": 298.29022216796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 99.0,
      "epoch": 0.0608,
      "grad_norm": 0.007988800294697285,
      "learning_rate": 3.972222222222223e-06,
      "loss": 0.0666,
      "num_tokens": 11280256.0,
      "reward": 1.17597496509552,
      "reward_std": 0.19168786704540253,
      "rewards/accuracy_reward_step": 0.375,
      "rewards/final_brier_reward_step": 0.6136687397956848,
      "rewards/format_reward_step": 0.98828125,
      "step": 57
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.42793796211481094,
      "aux_distill/mean_u": 0.24425906299334496,
      "aux_distill/n_active_tok": 380.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.49844444444444447,
      "calib/avg_num_step_conf": 5.94140625,
      "calib/ece": 0.2938823529411765,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0002333333333333333,
      "calib/mean_conf": 0.0002352941176470588,
      "calib/mu_c": 0.00039999999999999996,
      "calib/mu_w": 0.00016666666666666666,
      "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": 0.002156506209390983,
      "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": 1794.0,
      "completions/max_terminated_length": 1794.0,
      "completions/mean_length": 297.2578125,
      "completions/mean_terminated_length": 297.2578125,
      "completions/min_length": 77.0,
      "completions/min_terminated_length": 77.0,
      "epoch": 0.06186666666666667,
      "grad_norm": 0.007588682230561972,
      "learning_rate": 3.944444444444445e-06,
      "loss": 0.0745,
      "num_tokens": 11462674.0,
      "reward": 1.142693042755127,
      "reward_std": 0.1538347750902176,
      "rewards/accuracy_reward_step": 0.29296875,
      "rewards/final_brier_reward_step": 0.7033547163009644,
      "rewards/format_reward_step": 0.99609375,
      "step": 58
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4090555151924491,
      "aux_distill/mean_u": 0.22674918349077935,
      "aux_distill/n_active_tok": 356.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5060975609756098,
      "calib/avg_num_step_conf": 5.60546875,
      "calib/ece": 0.320234375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00024390243902439024,
      "calib/mean_conf": 7.8125e-05,
      "calib/mu_c": 0.00024390243902439024,
      "calib/mu_w": 0.0,
      "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": 0.001247556204896196,
      "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": 905.0,
      "completions/max_terminated_length": 905.0,
      "completions/mean_length": 284.3984375,
      "completions/mean_terminated_length": 285.51373291015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 81.0,
      "epoch": 0.06293333333333333,
      "grad_norm": 0.00699743302538991,
      "learning_rate": 3.916666666666667e-06,
      "loss": 0.0498,
      "num_tokens": 11641728.0,
      "reward": 1.1602336168289185,
      "reward_std": 0.14041298627853394,
      "rewards/accuracy_reward_step": 0.3203125,
      "rewards/final_brier_reward_step": 0.6798421740531921,
      "rewards/format_reward_step": 1.0,
      "step": 59
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4456522464752197,
      "aux_distill/mean_u": 0.2441319979642522,
      "aux_distill/n_active_tok": 359.78125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5079131268840779,
      "calib/avg_num_step_conf": 5.62109375,
      "calib/ece": 0.3514624505928854,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00037887092354069613,
      "calib/mean_conf": 0.00031620553359683795,
      "calib/mu_c": 0.0005617977528089888,
      "calib/mu_w": 0.00018292682926829268,
      "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": 0.002154784690978499,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2042.0,
      "completions/max_terminated_length": 2042.0,
      "completions/mean_length": 303.05078125,
      "completions/mean_terminated_length": 303.05078125,
      "completions/min_length": 87.0,
      "completions/min_terminated_length": 87.0,
      "epoch": 0.064,
      "grad_norm": 0.0078499810770154,
      "learning_rate": 3.88888888888889e-06,
      "loss": 0.1102,
      "num_tokens": 11828165.0,
      "reward": 1.1623023748397827,
      "reward_std": 0.19147999584674835,
      "rewards/accuracy_reward_step": 0.34765625,
      "rewards/final_brier_reward_step": 0.6410109996795654,
      "rewards/format_reward_step": 0.98828125,
      "step": 60
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4245549370534718,
      "aux_distill/mean_u": 0.21719526107186932,
      "aux_distill/n_active_tok": 360.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5060566931070528,
      "calib/avg_num_step_conf": 5.62890625,
      "calib/ece": 0.4565625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00023365922646498185,
      "calib/mean_conf": 0.00046875000000000004,
      "calib/mu_c": 0.0003418803418803419,
      "calib/mu_w": 0.0005755395683453237,
      "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": 0.0033910210022204228,
      "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": 992.0,
      "completions/max_terminated_length": 992.0,
      "completions/mean_length": 281.19921875,
      "completions/mean_terminated_length": 282.3019714355469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.06506666666666666,
      "grad_norm": 0.008309507742524147,
      "learning_rate": 3.861111111111112e-06,
      "loss": 0.0503,
      "num_tokens": 12004216.0,
      "reward": 1.228666067123413,
      "reward_std": 0.13236938416957855,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/final_brier_reward_step": 0.5432695150375366,
      "rewards/format_reward_step": 1.0,
      "step": 61
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4038896760903299,
      "aux_distill/mean_u": 0.21998769765356624,
      "aux_distill/n_active_tok": 381.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.505661231884058,
      "calib/avg_num_step_conf": 5.98828125,
      "calib/ece": 0.2809375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -4.830917874396138e-05,
      "calib/mean_conf": 0.0003125,
      "calib/mu_c": 0.0002777777777777778,
      "calib/mu_w": 0.00032608695652173916,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0027775607554111218,
      "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": 888.0,
      "completions/max_terminated_length": 888.0,
      "completions/mean_length": 312.54296875,
      "completions/mean_terminated_length": 313.7686462402344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 98.0,
      "epoch": 0.06613333333333334,
      "grad_norm": 0.006524847354739904,
      "learning_rate": 3.833333333333334e-06,
      "loss": 0.0302,
      "num_tokens": 12191307.0,
      "reward": 1.1367928981781006,
      "reward_std": 0.15469135344028473,
      "rewards/accuracy_reward_step": 0.28125,
      "rewards/final_brier_reward_step": 0.7149921655654907,
      "rewards/format_reward_step": 0.99609375,
      "step": 62
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43727352656424046,
      "aux_distill/mean_u": 0.26170699023718325,
      "aux_distill/n_active_tok": 403.28125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5011460931706331,
      "calib/avg_num_step_conf": 6.31640625,
      "calib/ece": 0.3579527559055118,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 2.292186341266094e-05,
      "calib/mean_conf": 0.00031496062992125983,
      "calib/mu_c": 0.00032967032967032967,
      "calib/mu_w": 0.00030674846625766873,
      "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": 0.001746541182111719,
      "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": 2360.0,
      "completions/max_terminated_length": 2360.0,
      "completions/mean_length": 331.5703125,
      "completions/mean_terminated_length": 332.87060546875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.0672,
      "grad_norm": 0.007350383326411247,
      "learning_rate": 3.8055555555555556e-06,
      "loss": 0.0818,
      "num_tokens": 12384829.0,
      "reward": 1.1700375080108643,
      "reward_std": 0.19835248589515686,
      "rewards/accuracy_reward_step": 0.35546875,
      "rewards/final_brier_reward_step": 0.6369500160217285,
      "rewards/format_reward_step": 0.9921875,
      "step": 63
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4085767618380487,
      "aux_distill/mean_u": 0.21745177327263837,
      "aux_distill/n_active_tok": 369.25,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5036487005505057,
      "calib/avg_num_step_conf": 5.76953125,
      "calib/ece": 0.4226086956521739,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -6.20919216489566e-05,
      "calib/mean_conf": 0.00031620553359683795,
      "calib/mu_c": 0.00028037383177570094,
      "calib/mu_w": 0.00034246575342465754,
      "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": 0.0023310107032073887,
      "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": 2331.0,
      "completions/max_terminated_length": 2331.0,
      "completions/mean_length": 314.87109375,
      "completions/mean_terminated_length": 314.87109375,
      "completions/min_length": 56.0,
      "completions/min_terminated_length": 56.0,
      "epoch": 0.06826666666666667,
      "grad_norm": 0.007864679209887981,
      "learning_rate": 3.777777777777778e-06,
      "loss": 0.1226,
      "num_tokens": 12569212.0,
      "reward": 1.1973800659179688,
      "reward_std": 0.2078630030155182,
      "rewards/accuracy_reward_step": 0.41796875,
      "rewards/final_brier_reward_step": 0.5705413818359375,
      "rewards/format_reward_step": 0.98828125,
      "step": 64
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4152026427909732,
      "aux_distill/mean_u": 0.22176144665746284,
      "aux_distill/n_active_tok": 322.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5045454545454545,
      "calib/avg_num_step_conf": 5.046875,
      "calib/ece": 0.4296484375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 9.090909090909092e-05,
      "calib/mean_conf": 3.90625e-05,
      "calib/mu_c": 9.090909090909092e-05,
      "calib/mu_w": 0.0,
      "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": 0.0006237781024480981,
      "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": 744.0,
      "completions/max_terminated_length": 744.0,
      "completions/mean_length": 268.62109375,
      "completions/mean_terminated_length": 269.6745300292969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 96.0,
      "epoch": 0.06933333333333333,
      "grad_norm": 0.0070212222635746,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.0323,
      "num_tokens": 12743003.0,
      "reward": 1.2148826122283936,
      "reward_std": 0.12614324688911438,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.5703902244567871,
      "rewards/format_reward_step": 1.0,
      "step": 65
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4280928233638406,
      "aux_distill/mean_u": 0.23642457940125466,
      "aux_distill/n_active_tok": 398.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5028768699654776,
      "calib/avg_num_step_conf": 6.203125,
      "calib/ece": 0.30929411764705883,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00017836593785960876,
      "calib/mean_conf": 0.0005098039215686275,
      "calib/mu_c": 0.0006329113924050633,
      "calib/mu_w": 0.00045454545454545455,
      "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": 0.002371171209071525,
      "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": 1600.0,
      "completions/max_terminated_length": 1600.0,
      "completions/mean_length": 330.25390625,
      "completions/mean_terminated_length": 330.25390625,
      "completions/min_length": 85.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.0704,
      "grad_norm": 0.006328082177788019,
      "learning_rate": 3.7222222222222225e-06,
      "loss": 0.0465,
      "num_tokens": 12933900.0,
      "reward": 1.150583028793335,
      "reward_std": 0.132345050573349,
      "rewards/accuracy_reward_step": 0.30859375,
      "rewards/final_brier_reward_step": 0.6878847479820251,
      "rewards/format_reward_step": 0.99609375,
      "step": 66
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.42184107284992933,
      "aux_distill/mean_u": 0.241509055720451,
      "aux_distill/n_active_tok": 356.53125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4847692307692308,
      "calib/avg_num_step_conf": 5.5703125,
      "calib/ece": 0.4899607843137255,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0003046153846153846,
      "calib/mean_conf": 0.0002352941176470588,
      "calib/mu_c": 8e-05,
      "calib/mu_w": 0.0003846153846153846,
      "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": 0.001515776320791191,
      "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": 2288.0,
      "completions/max_terminated_length": 2288.0,
      "completions/mean_length": 319.81640625,
      "completions/mean_terminated_length": 319.81640625,
      "completions/min_length": 87.0,
      "completions/min_terminated_length": 87.0,
      "epoch": 0.07146666666666666,
      "grad_norm": 0.007012557238340378,
      "learning_rate": 3.694444444444445e-06,
      "loss": 0.0474,
      "num_tokens": 13120781.0,
      "reward": 1.2402722835540771,
      "reward_std": 0.14968647062778473,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/final_brier_reward_step": 0.5078883171081543,
      "rewards/format_reward_step": 0.99609375,
      "step": 67
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4061860046349466,
      "aux_distill/mean_u": 0.21555456906916828,
      "aux_distill/n_active_tok": 340.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5073344914109245,
      "calib/avg_num_step_conf": 5.35546875,
      "calib/ece": 0.3863671875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 8.556906646078617e-05,
      "calib/mean_conf": 0.0003515625,
      "calib/mu_c": 0.00040404040404040404,
      "calib/mu_w": 0.00031847133757961787,
      "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": 0.002042860447655138,
      "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": 873.0,
      "completions/max_terminated_length": 873.0,
      "completions/mean_length": 284.3203125,
      "completions/mean_terminated_length": 285.435302734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 68.0,
      "epoch": 0.07253333333333334,
      "grad_norm": 0.0074111889116466045,
      "learning_rate": 3.6666666666666666e-06,
      "loss": 0.0429,
      "num_tokens": 13297655.0,
      "reward": 1.1935133934020996,
      "reward_std": 0.12575292587280273,
      "rewards/accuracy_reward_step": 0.38671875,
      "rewards/final_brier_reward_step": 0.6135894656181335,
      "rewards/format_reward_step": 1.0,
      "step": 68
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.42660697270184755,
      "aux_distill/mean_u": 0.2442016385441469,
      "aux_distill/n_active_tok": 357.125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5029569892473118,
      "calib/avg_num_step_conf": 5.53515625,
      "calib/ece": 0.36707509881422923,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00037768817204301064,
      "calib/mean_conf": 0.0005138339920948617,
      "calib/mu_c": 0.0007526881720430106,
      "calib/mu_w": 0.000375,
      "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": 0.003101192924258187,
      "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": 2326.0,
      "completions/max_terminated_length": 2326.0,
      "completions/mean_length": 350.0625,
      "completions/mean_terminated_length": 350.0625,
      "completions/min_length": 109.0,
      "completions/min_terminated_length": 109.0,
      "epoch": 0.0736,
      "grad_norm": 0.006129054818302393,
      "learning_rate": 3.638888888888889e-06,
      "loss": 0.098,
      "num_tokens": 13491767.0,
      "reward": 1.170190453529358,
      "reward_std": 0.1372353434562683,
      "rewards/accuracy_reward_step": 0.36328125,
      "rewards/final_brier_reward_step": 0.625537097454071,
      "rewards/format_reward_step": 0.98828125,
      "step": 69
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43751452304422855,
      "aux_distill/mean_u": 0.2280092215886282,
      "aux_distill/n_active_tok": 340.0,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.4961018711018711,
      "calib/avg_num_step_conf": 5.3125,
      "calib/ece": 0.4124603174603174,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -7.796257796257797e-05,
      "calib/mean_conf": 0.00023809523809523812,
      "calib/mu_c": 0.0001923076923076923,
      "calib/mu_w": 0.0002702702702702703,
      "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": 0.001524553389864964,
      "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": 2103.0,
      "completions/max_terminated_length": 2103.0,
      "completions/mean_length": 338.734375,
      "completions/mean_terminated_length": 338.734375,
      "completions/min_length": 93.0,
      "completions/min_terminated_length": 93.0,
      "epoch": 0.07466666666666667,
      "grad_norm": 0.006687454413622618,
      "learning_rate": 3.6111111111111115e-06,
      "loss": 0.0793,
      "num_tokens": 13685475.0,
      "reward": 1.1875770092010498,
      "reward_std": 0.15229782462120056,
      "rewards/accuracy_reward_step": 0.40625,
      "rewards/final_brier_reward_step": 0.5782788991928101,
      "rewards/format_reward_step": 0.984375,
      "step": 70
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41422352846711874,
      "aux_distill/mean_u": 0.2371021008723161,
      "aux_distill/n_active_tok": 377.5,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5053293611386958,
      "calib/avg_num_step_conf": 5.8984375,
      "calib/ece": 0.37366141732283464,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0001065872227739159,
      "calib/mean_conf": 0.00035433070866141734,
      "calib/mu_c": 0.0004210526315789474,
      "calib/mu_w": 0.0003144654088050315,
      "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": 0.0018487176191927393,
      "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": 1979.0,
      "completions/max_terminated_length": 1979.0,
      "completions/mean_length": 365.3671875,
      "completions/mean_terminated_length": 365.3671875,
      "completions/min_length": 85.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.07573333333333333,
      "grad_norm": 0.005943302530795336,
      "learning_rate": 3.5833333333333335e-06,
      "loss": 0.0653,
      "num_tokens": 13883417.0,
      "reward": 1.1778888702392578,
      "reward_std": 0.15389281511306763,
      "rewards/accuracy_reward_step": 0.37109375,
      "rewards/final_brier_reward_step": 0.6214027404785156,
      "rewards/format_reward_step": 0.9921875,
      "step": 71
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43518350925296545,
      "aux_distill/mean_u": 0.2185392387426317,
      "aux_distill/n_active_tok": 331.5,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5210302457466919,
      "calib/avg_num_step_conf": 5.1796875,
      "calib/ece": 0.3633201581027668,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0005279503105590061,
      "calib/mean_conf": 0.00031620553359683795,
      "calib/mu_c": 0.0006521739130434782,
      "calib/mu_w": 0.00012422360248447205,
      "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": 0.001962799844732828,
      "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": 2768.0,
      "completions/max_terminated_length": 2768.0,
      "completions/mean_length": 308.15625,
      "completions/mean_terminated_length": 308.15625,
      "completions/min_length": 93.0,
      "completions/min_terminated_length": 93.0,
      "epoch": 0.0768,
      "grad_norm": 0.008181699551641941,
      "learning_rate": 3.555555555555556e-06,
      "loss": 0.1023,
      "num_tokens": 14066713.0,
      "reward": 1.164294958114624,
      "reward_std": 0.17150144279003143,
      "rewards/accuracy_reward_step": 0.359375,
      "rewards/final_brier_reward_step": 0.6254648566246033,
      "rewards/format_reward_step": 0.984375,
      "step": 72
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41626708675175905,
      "aux_distill/mean_u": 0.194263666773112,
      "aux_distill/n_active_tok": 314.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.49814814814814823,
      "calib/avg_num_step_conf": 4.9140625,
      "calib/ece": 0.47023529411764703,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -3.7037037037037084e-05,
      "calib/mean_conf": 0.0003529411764705882,
      "calib/mu_c": 0.0003333333333333333,
      "calib/mu_w": 0.0003703703703703704,
      "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": 0.0018452220166303669,
      "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": 2391.0,
      "completions/max_terminated_length": 2391.0,
      "completions/mean_length": 315.80859375,
      "completions/mean_terminated_length": 315.80859375,
      "completions/min_length": 115.0,
      "completions/min_terminated_length": 115.0,
      "epoch": 0.07786666666666667,
      "grad_norm": 0.006817559711635113,
      "learning_rate": 3.5277777777777784e-06,
      "loss": 0.0434,
      "num_tokens": 14254592.0,
      "reward": 1.2306232452392578,
      "reward_std": 0.15328508615493774,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.5276527404785156,
      "rewards/format_reward_step": 0.99609375,
      "step": 73
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4087336999364197,
      "aux_distill/mean_u": 0.17648553013055804,
      "aux_distill/n_active_tok": 318.25,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5107051282051283,
      "calib/avg_num_step_conf": 5.0234375,
      "calib/ece": 0.40909448818897637,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00021410256410256412,
      "calib/mean_conf": 0.0003543307086614173,
      "calib/mu_c": 0.0004807692307692308,
      "calib/mu_w": 0.0002666666666666667,
      "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": 0.0018487176191927398,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2514.0,
      "completions/max_terminated_length": 2514.0,
      "completions/mean_length": 302.38671875,
      "completions/mean_terminated_length": 303.57257080078125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 89.0,
      "epoch": 0.07893333333333333,
      "grad_norm": 0.008606214076280594,
      "learning_rate": 3.5e-06,
      "loss": 0.0454,
      "num_tokens": 14435931.0,
      "reward": 1.1955060958862305,
      "reward_std": 0.1825999915599823,
      "rewards/accuracy_reward_step": 0.40625,
      "rewards/final_brier_reward_step": 0.5863245725631714,
      "rewards/format_reward_step": 0.9921875,
      "step": 74
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37989877723157406,
      "aux_distill/mean_u": 0.2001054968370109,
      "aux_distill/n_active_tok": 312.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5027925531914894,
      "calib/avg_num_step_conf": 4.88671875,
      "calib/ece": 0.6295669291338583,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 5.585106382978726e-05,
      "calib/mean_conf": 0.0003543307086614173,
      "calib/mu_c": 0.000375,
      "calib/mu_w": 0.00031914893617021275,
      "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": 0.0018487176191927398,
      "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": 2718.0,
      "completions/max_terminated_length": 2718.0,
      "completions/mean_length": 307.0546875,
      "completions/mean_terminated_length": 307.0546875,
      "completions/min_length": 103.0,
      "completions/min_terminated_length": 103.0,
      "epoch": 0.08,
      "grad_norm": 0.008363153785467148,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 0.0629,
      "num_tokens": 14619289.0,
      "reward": 1.3088264465332031,
      "reward_std": 0.15875719487667084,
      "rewards/accuracy_reward_step": 0.62890625,
      "rewards/final_brier_reward_step": 0.3676527142524719,
      "rewards/format_reward_step": 0.9921875,
      "step": 75
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4291513739153743,
      "aux_distill/mean_u": 0.23722198629035704,
      "aux_distill/n_active_tok": 308.78125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.49385964912280694,
      "calib/avg_num_step_conf": 4.82421875,
      "calib/ece": 0.47391304347826085,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 4.072681704260652e-05,
      "calib/mean_conf": 0.0003952569169960474,
      "calib/mu_c": 0.0004166666666666667,
      "calib/mu_w": 0.00037593984962406017,
      "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": 0.0024835222248861676,
      "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": 2518.0,
      "completions/max_terminated_length": 2518.0,
      "completions/mean_length": 312.01171875,
      "completions/mean_terminated_length": 312.01171875,
      "completions/min_length": 108.0,
      "completions/min_terminated_length": 108.0,
      "epoch": 0.08106666666666666,
      "grad_norm": 0.008254244923591614,
      "learning_rate": 3.444444444444445e-06,
      "loss": 0.0841,
      "num_tokens": 14802220.0,
      "reward": 1.2228484153747559,
      "reward_std": 0.1601463109254837,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.5199155807495117,
      "rewards/format_reward_step": 0.98828125,
      "step": 76
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4120447961613536,
      "aux_distill/mean_u": 0.22283618500480631,
      "aux_distill/n_active_tok": 331.25,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.4801587301587301,
      "calib/avg_num_step_conf": 5.1796875,
      "calib/ece": 0.49556,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0005529953917050691,
      "calib/mean_conf": 0.00044000000000000007,
      "calib/mu_c": 0.00016129032258064516,
      "calib/mu_w": 0.0007142857142857143,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0024096472770926453,
      "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": 2523.0,
      "completions/max_terminated_length": 2523.0,
      "completions/mean_length": 319.32421875,
      "completions/mean_terminated_length": 319.32421875,
      "completions/min_length": 64.0,
      "completions/min_terminated_length": 64.0,
      "epoch": 0.08213333333333334,
      "grad_norm": 0.009399495087563992,
      "learning_rate": 3.416666666666667e-06,
      "loss": 0.1611,
      "num_tokens": 14988631.0,
      "reward": 1.2227314710617065,
      "reward_std": 0.20367680490016937,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/final_brier_reward_step": 0.4923378825187683,
      "rewards/format_reward_step": 0.9765625,
      "step": 77
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4121271399781108,
      "aux_distill/mean_u": 0.21187937211055408,
      "aux_distill/n_active_tok": 332.75,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.48820649755229195,
      "calib/avg_num_step_conf": 5.19921875,
      "calib/ece": 0.4206299212598425,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00028100960010172296,
      "calib/mean_conf": 0.0006299212598425197,
      "calib/mu_c": 0.00046728971962616824,
      "calib/mu_w": 0.0007482993197278912,
      "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": 0.0030086571928413857,
      "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": 2161.0,
      "completions/max_terminated_length": 2161.0,
      "completions/mean_length": 346.95703125,
      "completions/mean_terminated_length": 346.95703125,
      "completions/min_length": 118.0,
      "completions/min_terminated_length": 118.0,
      "epoch": 0.0832,
      "grad_norm": 0.007775504142045975,
      "learning_rate": 3.3888888888888893e-06,
      "loss": 0.0838,
      "num_tokens": 15185476.0,
      "reward": 1.2013624906539917,
      "reward_std": 0.14862890541553497,
      "rewards/accuracy_reward_step": 0.41796875,
      "rewards/final_brier_reward_step": 0.5745999813079834,
      "rewards/format_reward_step": 0.9921875,
      "step": 78
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40545395016670227,
      "aux_distill/mean_u": 0.21084205152926944,
      "aux_distill/n_active_tok": 333.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.49871700879765396,
      "calib/avg_num_step_conf": 5.26953125,
      "calib/ece": 0.5151171875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 4.643206256109482e-05,
      "calib/mean_conf": 0.0005078125,
      "calib/mu_c": 0.0005303030303030304,
      "calib/mu_w": 0.00048387096774193554,
      "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": 0.002366749134328298,
      "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": 760.0,
      "completions/max_terminated_length": 760.0,
      "completions/mean_length": 340.0703125,
      "completions/mean_terminated_length": 341.4039306640625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 63.0,
      "epoch": 0.08426666666666667,
      "grad_norm": 0.00761050172150135,
      "learning_rate": 3.3611111111111117e-06,
      "loss": 0.0376,
      "num_tokens": 15378910.0,
      "reward": 1.2580829858779907,
      "reward_std": 0.12420278042554855,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/final_brier_reward_step": 0.4849160313606262,
      "rewards/format_reward_step": 1.0,
      "step": 79
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41106561524793506,
      "aux_distill/mean_u": 0.21736490320900576,
      "aux_distill/n_active_tok": 301.875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5141188828472566,
      "calib/avg_num_step_conf": 4.69140625,
      "calib/ece": 0.46619607843137256,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00037815126050420167,
      "calib/mean_conf": 0.00047058823529411766,
      "calib/mu_c": 0.0006722689075630252,
      "calib/mu_w": 0.00029411764705882356,
      "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": 0.0027607397500859676,
      "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": 1661.0,
      "completions/max_terminated_length": 1661.0,
      "completions/mean_length": 289.74609375,
      "completions/mean_terminated_length": 289.74609375,
      "completions/min_length": 98.0,
      "completions/min_terminated_length": 98.0,
      "epoch": 0.08533333333333333,
      "grad_norm": 0.009661182761192322,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.0838,
      "num_tokens": 15555245.0,
      "reward": 1.228824257850647,
      "reward_std": 0.15840938687324524,
      "rewards/accuracy_reward_step": 0.46484375,
      "rewards/final_brier_reward_step": 0.5318671464920044,
      "rewards/format_reward_step": 0.99609375,
      "step": 80
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41248337598517537,
      "aux_distill/mean_u": 0.23188753956743624,
      "aux_distill/n_active_tok": 289.5,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5129652605459057,
      "calib/avg_num_step_conf": 4.5859375,
      "calib/ece": 0.48771653543307086,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00033746898263027294,
      "calib/mean_conf": 0.00047244094488188977,
      "calib/mu_c": 0.0006451612903225806,
      "calib/mu_w": 0.0003076923076923077,
      "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": 0.002299697931733895,
      "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": 1484.0,
      "completions/max_terminated_length": 1484.0,
      "completions/mean_length": 317.109375,
      "completions/mean_terminated_length": 318.35296630859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 115.0,
      "epoch": 0.0864,
      "grad_norm": 0.009779230691492558,
      "learning_rate": 3.3055555555555558e-06,
      "loss": 0.053,
      "num_tokens": 15742673.0,
      "reward": 1.234684705734253,
      "reward_std": 0.16173626482486725,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.5084320306777954,
      "rewards/format_reward_step": 0.9921875,
      "step": 81
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.405979058239609,
      "aux_distill/mean_u": 0.1937273843168663,
      "aux_distill/n_active_tok": 243.75,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.49200253887654716,
      "calib/avg_num_step_conf": 3.80859375,
      "calib/ece": 0.5431746031746031,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00024373214852427803,
      "calib/mean_conf": 0.0004761904761904762,
      "calib/mu_c": 0.00036496350364963507,
      "calib/mu_w": 0.0006086956521739131,
      "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": 0.0023084189797220617,
      "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": 0.01171875,
      "completions/max_length": 2440.0,
      "completions/max_terminated_length": 2440.0,
      "completions/mean_length": 279.09765625,
      "completions/mean_terminated_length": 282.4071350097656,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 76.0,
      "epoch": 0.08746666666666666,
      "grad_norm": 0.0082706892862916,
      "learning_rate": 3.277777777777778e-06,
      "loss": 0.033,
      "num_tokens": 15919674.0,
      "reward": 1.248239517211914,
      "reward_std": 0.1699363887310028,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/final_brier_reward_step": 0.4456976652145386,
      "rewards/format_reward_step": 0.98046875,
      "step": 82
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3855320755392313,
      "aux_distill/mean_u": 0.20919488492473376,
      "aux_distill/n_active_tok": 240.5,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.48275862068965514,
      "calib/avg_num_step_conf": 3.7578125,
      "calib/ece": 0.5407905138339921,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00027812735967782537,
      "calib/mean_conf": 0.0007114624505928853,
      "calib/mu_c": 0.0005839416058394161,
      "calib/mu_w": 0.0008620689655172415,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.9765625,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0027201028512387755,
      "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": 0.01171875,
      "completions/max_length": 876.0,
      "completions/max_terminated_length": 876.0,
      "completions/mean_length": 332.66796875,
      "completions/mean_terminated_length": 336.6126708984375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 99.0,
      "epoch": 0.08853333333333334,
      "grad_norm": 0.007690061815083027,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.0329,
      "num_tokens": 16112101.0,
      "reward": 1.2483553886413574,
      "reward_std": 0.13580173254013062,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/final_brier_reward_step": 0.4498359262943268,
      "rewards/format_reward_step": 0.9765625,
      "step": 83
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4101240658201277,
      "aux_distill/mean_u": 0.20711848674623146,
      "aux_distill/n_active_tok": 207.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5073301991834716,
      "calib/avg_num_step_conf": 3.24609375,
      "calib/ece": 0.4622745098039216,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00022825683533341575,
      "calib/mean_conf": 0.00047058823529411766,
      "calib/mu_c": 0.0005932203389830508,
      "calib/mu_w": 0.00036496350364963507,
      "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": 0.0022953742159469645,
      "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": 2607.0,
      "completions/max_terminated_length": 2607.0,
      "completions/mean_length": 287.5390625,
      "completions/mean_terminated_length": 287.5390625,
      "completions/min_length": 58.0,
      "completions/min_terminated_length": 58.0,
      "epoch": 0.0896,
      "grad_norm": 0.010937503539025784,
      "learning_rate": 3.2222222222222227e-06,
      "loss": 0.028,
      "num_tokens": 16291631.0,
      "reward": 1.2268332242965698,
      "reward_std": 0.12932856380939484,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/final_brier_reward_step": 0.5356976389884949,
      "rewards/format_reward_step": 0.99609375,
      "step": 84
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3972956840880215,
      "aux_distill/mean_u": 0.2200048133976546,
      "aux_distill/n_active_tok": 222.0,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5182291666666666,
      "calib/avg_num_step_conf": 3.46875,
      "calib/ece": 0.4829435483870968,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00014062500000000002,
      "calib/mean_conf": 0.0009274193548387097,
      "calib/mu_c": 0.001,
      "calib/mu_w": 0.000859375,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.004703590576146321,
      "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": 2342.0,
      "completions/max_terminated_length": 2342.0,
      "completions/mean_length": 325.8125,
      "completions/mean_terminated_length": 327.0902099609375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 72.0,
      "epoch": 0.09066666666666667,
      "grad_norm": 0.009536146186292171,
      "learning_rate": 3.1944444444444443e-06,
      "loss": 0.1186,
      "num_tokens": 16482863.0,
      "reward": 1.1918638944625854,
      "reward_std": 0.2307528555393219,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.48919644951820374,
      "rewards/format_reward_step": 0.95703125,
      "step": 85
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.396264610812068,
      "aux_distill/mean_u": 0.1669753048582908,
      "aux_distill/n_active_tok": 150.78125,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5228428050052137,
      "calib/avg_num_step_conf": 2.359375,
      "calib/ece": 0.44710843373493975,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.953125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.004846193952033369,
      "calib/mean_conf": 0.0026907630522088354,
      "calib/mu_c": 0.005357142857142858,
      "calib/mu_w": 0.000510948905109489,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.03170504050452634,
      "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": 0.01171875,
      "completions/max_length": 1915.0,
      "completions/max_terminated_length": 1915.0,
      "completions/mean_length": 301.15625,
      "completions/mean_terminated_length": 304.727294921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 66.0,
      "epoch": 0.09173333333333333,
      "grad_norm": 0.009685726836323738,
      "learning_rate": 3.1666666666666667e-06,
      "loss": 0.0231,
      "num_tokens": 16665471.0,
      "reward": 1.1756794452667236,
      "reward_std": 0.2164359986782074,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/final_brier_reward_step": 0.5232340097427368,
      "rewards/format_reward_step": 0.953125,
      "step": 86
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3704542489722371,
      "aux_distill/mean_u": 0.1592296599675276,
      "aux_distill/n_active_tok": 132.0,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5052309782608695,
      "calib/avg_num_step_conf": 2.06640625,
      "calib/ece": 0.634047619047619,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 5.434782608695963e-06,
      "calib/mean_conf": 0.0008730158730158729,
      "calib/mu_c": 0.0008750000000000001,
      "calib/mu_w": 0.0008695652173913042,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0029600089204716695,
      "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": 0.00390625,
      "completions/max_length": 2846.0,
      "completions/max_terminated_length": 2846.0,
      "completions/mean_length": 306.0625,
      "completions/mean_terminated_length": 307.26275634765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 62.0,
      "epoch": 0.0928,
      "grad_norm": 0.01069539226591587,
      "learning_rate": 3.138888888888889e-06,
      "loss": 0.1338,
      "num_tokens": 16849319.0,
      "reward": 1.283745288848877,
      "reward_std": 0.1876409649848938,
      "rewards/accuracy_reward_step": 0.625,
      "rewards/final_brier_reward_step": 0.3487406373023987,
      "rewards/format_reward_step": 0.96875,
      "step": 87
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.396932493429631,
      "aux_distill/mean_u": 0.1686609234334076,
      "aux_distill/n_active_tok": 154.28125,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5094345760947361,
      "calib/avg_num_step_conf": 2.4140625,
      "calib/ece": 0.563,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00014639859457349228,
      "calib/mean_conf": 0.001,
      "calib/mu_c": 0.0010638297872340428,
      "calib/mu_w": 0.0009174311926605505,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.98046875,
      "calib/nonempty_step_conf_rate": 0.94140625,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0036055512754639895,
      "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": 0.00390625,
      "completions/max_length": 3062.0,
      "completions/max_terminated_length": 3062.0,
      "completions/mean_length": 377.2734375,
      "completions/mean_terminated_length": 378.7529602050781,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 95.0,
      "epoch": 0.09386666666666667,
      "grad_norm": 0.009100557304918766,
      "learning_rate": 3.1111111111111116e-06,
      "loss": 0.1337,
      "num_tokens": 17055749.0,
      "reward": 1.2193293571472168,
      "reward_std": 0.21440431475639343,
      "rewards/accuracy_reward_step": 0.55078125,
      "rewards/final_brier_reward_step": 0.39959609508514404,
      "rewards/format_reward_step": 0.9375,
      "step": 88
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.397604210767895,
      "aux_distill/mean_u": 0.1975274313688232,
      "aux_distill/n_active_tok": 137.25,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.4970833333333333,
      "calib/avg_num_step_conf": 2.14453125,
      "calib/ece": 0.5191600000000001,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00014743589743589753,
      "calib/mean_conf": 0.00084,
      "calib/mu_c": 0.0007692307692307692,
      "calib/mu_w": 0.0009166666666666668,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.9296875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0033006665993401997,
      "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": 0.00390625,
      "completions/max_length": 2465.0,
      "completions/max_terminated_length": 2465.0,
      "completions/mean_length": 350.140625,
      "completions/mean_terminated_length": 351.51373291015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.09493333333333333,
      "grad_norm": 0.010221146047115326,
      "learning_rate": 3.0833333333333336e-06,
      "loss": 0.1189,
      "num_tokens": 17254273.0,
      "reward": 1.1858930587768555,
      "reward_std": 0.2175370752811432,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.43428629636764526,
      "rewards/format_reward_step": 0.921875,
      "step": 89
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39298654487356544,
      "aux_distill/mean_u": 0.22362391264261372,
      "aux_distill/n_active_tok": 144.53125,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5160235448005231,
      "calib/avg_num_step_conf": 2.2578125,
      "calib/ece": 0.5571084337349397,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.92578125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0005487246566383257,
      "calib/mean_conf": 0.0011244979919678717,
      "calib/mu_c": 0.001366906474820144,
      "calib/mu_w": 0.0008181818181818182,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.9765625,
      "calib/nonempty_step_conf_rate": 0.9296875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0046071101685783095,
      "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": 0.0,
      "completions/max_length": 2835.0,
      "completions/max_terminated_length": 2835.0,
      "completions/mean_length": 334.0234375,
      "completions/mean_terminated_length": 334.0234375,
      "completions/min_length": 68.0,
      "completions/min_terminated_length": 68.0,
      "epoch": 0.096,
      "grad_norm": 0.011843850836157799,
      "learning_rate": 3.055555555555556e-06,
      "loss": 0.1311,
      "num_tokens": 17443103.0,
      "reward": 1.2019033432006836,
      "reward_std": 0.25413912534713745,
      "rewards/accuracy_reward_step": 0.54296875,
      "rewards/final_brier_reward_step": 0.3920878767967224,
      "rewards/format_reward_step": 0.92578125,
      "step": 90
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3809904200024903,
      "aux_distill/mean_u": 0.13741306957077365,
      "aux_distill/n_active_tok": 114.8125,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.46223776223776225,
      "calib/avg_num_step_conf": 1.79296875,
      "calib/ece": 0.5761290322580646,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00244955044955045,
      "calib/mean_conf": 0.0020161290322580645,
      "calib/mu_c": 0.000979020979020979,
      "calib/mu_w": 0.003428571428571429,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.98046875,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.0007661290322580645,
      "calib/std_conf": 0.012603112692177985,
      "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": 0.00390625,
      "completions/max_length": 2177.0,
      "completions/max_terminated_length": 2177.0,
      "completions/mean_length": 320.0546875,
      "completions/mean_terminated_length": 321.309814453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 107.0,
      "epoch": 0.09706666666666666,
      "grad_norm": 0.008399066515266895,
      "learning_rate": 3.0277777777777776e-06,
      "loss": 0.1069,
      "num_tokens": 17632749.0,
      "reward": 1.2387499809265137,
      "reward_std": 0.18196678161621094,
      "rewards/accuracy_reward_step": 0.55859375,
      "rewards/final_brier_reward_step": 0.40328124165534973,
      "rewards/format_reward_step": 0.95703125,
      "step": 91
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41714442893862724,
      "aux_distill/mean_u": 0.15397619196037046,
      "aux_distill/n_active_tok": 116.75,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.47993827160493824,
      "calib/avg_num_step_conf": 1.82421875,
      "calib/ece": 0.5702380952380952,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0005092592592592591,
      "calib/mean_conf": 0.0011904761904761906,
      "calib/mu_c": 0.0009722222222222223,
      "calib/mu_w": 0.0014814814814814814,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.95703125,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0035872664698379433,
      "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": 0.0078125,
      "completions/max_length": 1886.0,
      "completions/max_terminated_length": 1886.0,
      "completions/mean_length": 301.79296875,
      "completions/mean_terminated_length": 304.1692810058594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 113.0,
      "epoch": 0.09813333333333334,
      "grad_norm": 0.010597274638712406,
      "learning_rate": 3e-06,
      "loss": 0.0414,
      "num_tokens": 17816728.0,
      "reward": 1.246633529663086,
      "reward_std": 0.19921471178531647,
      "rewards/accuracy_reward_step": 0.5625,
      "rewards/final_brier_reward_step": 0.4112359583377838,
      "rewards/format_reward_step": 0.95703125,
      "step": 92
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3879448645748198,
      "aux_distill/mean_u": 0.15627032364352322,
      "aux_distill/n_active_tok": 140.25,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.4699326898653797,
      "calib/avg_num_step_conf": 2.19140625,
      "calib/ece": 0.4923107569721116,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.000835661671323343,
      "calib/mean_conf": 0.0017131474103585659,
      "calib/mu_c": 0.0012903225806451613,
      "calib/mu_w": 0.0021259842519685043,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.94921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.005190915529849348,
      "calib/step_conf_rate": 0.94921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2323.0,
      "completions/max_terminated_length": 2323.0,
      "completions/mean_length": 369.08984375,
      "completions/mean_terminated_length": 369.08984375,
      "completions/min_length": 83.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.0992,
      "grad_norm": 0.008991632610559464,
      "learning_rate": 2.9722222222222225e-06,
      "loss": 0.131,
      "num_tokens": 18016991.0,
      "reward": 1.1959228515625,
      "reward_std": 0.21242350339889526,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.4816894233226776,
      "rewards/format_reward_step": 0.94140625,
      "step": 93
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3044614642858505,
      "aux_distill/mean_u": 0.13583506846250099,
      "aux_distill/n_active_tok": 105.75,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.4786786786786786,
      "calib/avg_num_step_conf": 1.65234375,
      "calib/ece": 0.5476016260162602,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.93359375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0004784784784784787,
      "calib/mean_conf": 0.0011788617886178863,
      "calib/mu_c": 0.000962962962962963,
      "calib/mu_w": 0.0014414414414414417,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 0.96875,
      "calib/nonempty_step_conf_rate": 0.94140625,
      "calib/pce": 0.0,
      "calib/std_conf": 0.004111321905441839,
      "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": 0.00390625,
      "completions/max_length": 2710.0,
      "completions/max_terminated_length": 2710.0,
      "completions/mean_length": 324.40234375,
      "completions/mean_terminated_length": 325.6745300292969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.10026666666666667,
      "grad_norm": 0.010803716257214546,
      "learning_rate": 2.944444444444445e-06,
      "loss": 0.1487,
      "num_tokens": 18208718.0,
      "reward": 1.1997181177139282,
      "reward_std": 0.194935142993927,
      "rewards/accuracy_reward_step": 0.52734375,
      "rewards/final_brier_reward_step": 0.4111550450325012,
      "rewards/format_reward_step": 0.93359375,
      "step": 94
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40034619299694896,
      "aux_distill/mean_u": 0.21922548809225598,
      "aux_distill/n_active_tok": 99.75,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5176317501626545,
      "calib/avg_num_step_conf": 1.5625,
      "calib/ece": 0.5754183266932271,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.001318152244632401,
      "calib/mean_conf": 0.002270916334661355,
      "calib/mu_c": 0.002827586206896552,
      "calib/mu_w": 0.0015094339622641509,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.98046875,
      "calib/nonempty_step_conf_rate": 0.9609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.007469880474262826,
      "calib/step_conf_rate": 0.9609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2507.0,
      "completions/max_terminated_length": 2507.0,
      "completions/mean_length": 325.10546875,
      "completions/mean_terminated_length": 325.10546875,
      "completions/min_length": 70.0,
      "completions/min_terminated_length": 70.0,
      "epoch": 0.10133333333333333,
      "grad_norm": 0.009412271901965141,
      "learning_rate": 2.916666666666667e-06,
      "loss": 0.1876,
      "num_tokens": 18398073.0,
      "reward": 1.2457122802734375,
      "reward_std": 0.19631265103816986,
      "rewards/accuracy_reward_step": 0.56640625,
      "rewards/final_brier_reward_step": 0.4015808701515198,
      "rewards/format_reward_step": 0.95703125,
      "step": 95
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3630610196851194,
      "aux_distill/mean_u": 0.14667928688553772,
      "aux_distill/n_active_tok": 90.09375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.4754359925788498,
      "calib/avg_num_step_conf": 1.40625,
      "calib/ece": 0.6922222222222222,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0014753246753246756,
      "calib/mean_conf": 0.0022222222222222222,
      "calib/mu_c": 0.0017714285714285714,
      "calib/mu_w": 0.003246753246753247,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.007806070987896382,
      "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": 0.0,
      "completions/max_length": 2452.0,
      "completions/max_terminated_length": 2452.0,
      "completions/mean_length": 289.3046875,
      "completions/mean_terminated_length": 289.3046875,
      "completions/min_length": 83.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.1024,
      "grad_norm": 0.008919981308281422,
      "learning_rate": 2.888888888888889e-06,
      "loss": 0.1522,
      "num_tokens": 18577951.0,
      "reward": 1.309772253036499,
      "reward_std": 0.1510075330734253,
      "rewards/accuracy_reward_step": 0.68359375,
      "rewards/final_brier_reward_step": 0.29141953587532043,
      "rewards/format_reward_step": 0.9609375,
      "step": 96
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3719734316691756,
      "aux_distill/mean_u": 0.1676030758858746,
      "aux_distill/n_active_tok": 99.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.49187392267914315,
      "calib/avg_num_step_conf": 1.55859375,
      "calib/ece": 0.48415686274509806,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0001976114257572028,
      "calib/mean_conf": 0.0021176470588235297,
      "calib/mu_c": 0.0020161290322580645,
      "calib/mu_w": 0.0022137404580152673,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.005758700353186079,
      "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": 0.0,
      "completions/max_length": 1841.0,
      "completions/max_terminated_length": 1841.0,
      "completions/mean_length": 268.06640625,
      "completions/mean_terminated_length": 268.06640625,
      "completions/min_length": 108.0,
      "completions/min_terminated_length": 108.0,
      "epoch": 0.10346666666666667,
      "grad_norm": 0.012327205389738083,
      "learning_rate": 2.861111111111111e-06,
      "loss": 0.0713,
      "num_tokens": 18751648.0,
      "reward": 1.2157626152038574,
      "reward_std": 0.19388362765312195,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.4940253496170044,
      "rewards/format_reward_step": 0.96875,
      "step": 97
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3803948569111526,
      "aux_distill/mean_u": 0.18490822814105912,
      "aux_distill/n_active_tok": 78.53125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.4868605689501211,
      "calib/avg_num_step_conf": 1.2265625,
      "calib/ece": 0.5311155378486055,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.00796812749003984,
      "calib/gap": -0.0024122974869243523,
      "calib/mean_conf": 0.010079681274900398,
      "calib/mu_c": 0.008955223880597015,
      "calib/mu_w": 0.011367521367521368,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.97265625,
      "calib/pce": 0.003665338645418326,
      "calib/std_conf": 0.08539102480372213,
      "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": 0.0,
      "completions/max_length": 1891.0,
      "completions/max_terminated_length": 1891.0,
      "completions/mean_length": 296.8359375,
      "completions/mean_terminated_length": 296.8359375,
      "completions/min_length": 55.0,
      "completions/min_terminated_length": 55.0,
      "epoch": 0.10453333333333334,
      "grad_norm": 0.010739736258983612,
      "learning_rate": 2.8333333333333335e-06,
      "loss": 0.1146,
      "num_tokens": 18933822.0,
      "reward": 1.2334849834442139,
      "reward_std": 0.20740289986133575,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/final_brier_reward_step": 0.45134490728378296,
      "rewards/format_reward_step": 0.96875,
      "step": 98
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36638830602169037,
      "aux_distill/mean_u": 0.11033139603598505,
      "aux_distill/n_active_tok": 78.0,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5610946745562129,
      "calib/avg_num_step_conf": 1.21875,
      "calib/ece": 0.3180321285140562,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.004231508875739645,
      "calib/mean_conf": 0.0032530120481927714,
      "calib/mu_c": 0.006125,
      "calib/mu_w": 0.0018934911242603554,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.97265625,
      "calib/nonempty_step_conf_rate": 0.94921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.013660405105061599,
      "calib/step_conf_rate": 0.94921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2102.0,
      "completions/max_terminated_length": 2102.0,
      "completions/mean_length": 352.171875,
      "completions/mean_terminated_length": 353.5529479980469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 79.0,
      "epoch": 0.1056,
      "grad_norm": 0.007696256972849369,
      "learning_rate": 2.805555555555556e-06,
      "loss": 0.1199,
      "num_tokens": 19129778.0,
      "reward": 1.111154556274414,
      "reward_std": 0.21740788221359253,
      "rewards/accuracy_reward_step": 0.3125,
      "rewards/final_brier_reward_step": 0.6480902433395386,
      "rewards/format_reward_step": 0.94921875,
      "step": 99
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3693304159678519,
      "aux_distill/mean_u": 0.19013945568804616,
      "aux_distill/n_active_tok": 78.75,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.5161803713527852,
      "calib/avg_num_step_conf": 1.23046875,
      "calib/ece": 0.5249999999999999,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0018408488063660472,
      "calib/mean_conf": 0.004593495934959349,
      "calib/mu_c": 0.005461538461538461,
      "calib/mu_w": 0.003620689655172414,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 0.9609375,
      "calib/nonempty_step_conf_rate": 0.94140625,
      "calib/pce": 0.0005691056910569107,
      "calib/std_conf": 0.01820271461858796,
      "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": 0.00390625,
      "completions/max_length": 2178.0,
      "completions/max_terminated_length": 2178.0,
      "completions/mean_length": 333.40625,
      "completions/mean_terminated_length": 334.7137451171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 63.0,
      "epoch": 0.10666666666666667,
      "grad_norm": 0.008256965316832066,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 0.153,
      "num_tokens": 19322538.0,
      "reward": 1.1979174613952637,
      "reward_std": 0.2293814718723297,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.4388035237789154,
      "rewards/format_reward_step": 0.94140625,
      "step": 100
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3319393121637404,
      "aux_distill/mean_u": 0.09735452389920093,
      "aux_distill/n_active_tok": 72.78125,
      "calib/answer_extract_rate": 0.95703125,
      "calib/auroc": 0.5732998506854894,
      "calib/avg_num_step_conf": 1.13671875,
      "calib/ece": 0.42551020408163265,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.004081632653061225,
      "calib/gap": 0.010085516492466405,
      "calib/mean_conf": 0.009183673469387756,
      "calib/mu_c": 0.014905660377358491,
      "calib/mu_w": 0.004820143884892086,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 0.9609375,
      "calib/nonempty_step_conf_rate": 0.9453125,
      "calib/pce": 0.0010204081632653062,
      "calib/std_conf": 0.06635208299004534,
      "calib/step_conf_rate": 0.9453125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2982.0,
      "completions/max_terminated_length": 2982.0,
      "completions/mean_length": 355.12890625,
      "completions/mean_terminated_length": 357.9252014160156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 91.0,
      "epoch": 0.10773333333333333,
      "grad_norm": 0.00929940678179264,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.1571,
      "num_tokens": 19520443.0,
      "reward": 1.1560661792755127,
      "reward_std": 0.2282753884792328,
      "rewards/accuracy_reward_step": 0.4140625,
      "rewards/final_brier_reward_step": 0.5426011681556702,
      "rewards/format_reward_step": 0.94140625,
      "step": 101
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34266128670424223,
      "aux_distill/mean_u": 0.12935753004547906,
      "aux_distill/n_active_tok": 78.75,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5238063660477453,
      "calib/avg_num_step_conf": 1.23046875,
      "calib/ece": 0.5739759036144578,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.9453125,
      "calib/frac_conf_gt_0.9": 0.004016064257028112,
      "calib/gap": 0.008730769230769231,
      "calib/mean_conf": 0.008353413654618475,
      "calib/mu_c": 0.012,
      "calib/mu_w": 0.0032692307692307695,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.98046875,
      "calib/nonempty_step_conf_rate": 0.953125,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0639347339025536,
      "calib/step_conf_rate": 0.953125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1889.0,
      "completions/max_terminated_length": 1889.0,
      "completions/mean_length": 276.265625,
      "completions/mean_terminated_length": 277.3490295410156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 72.0,
      "epoch": 0.1088,
      "grad_norm": 0.008674833923578262,
      "learning_rate": 2.7222222222222224e-06,
      "loss": 0.0695,
      "num_tokens": 19697863.0,
      "reward": 1.2467310428619385,
      "reward_std": 0.18275701999664307,
      "rewards/accuracy_reward_step": 0.5703125,
      "rewards/final_brier_reward_step": 0.40752461552619934,
      "rewards/format_reward_step": 0.9453125,
      "step": 102
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.43721392983570695,
      "aux_distill/mean_u": 0.1723636638590128,
      "aux_distill/n_active_tok": 73.03125,
      "calib/answer_extract_rate": 0.953125,
      "calib/auroc": 0.493941701862494,
      "calib/avg_num_step_conf": 1.140625,
      "calib/ece": 0.5790163934426229,
      "calib/final_conf_rate": 0.953125,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 0.012295081967213115,
      "calib/gap": 0.001057951949041059,
      "calib/mean_conf": 0.016065573770491802,
      "calib/mu_c": 0.016503496503496504,
      "calib/mu_w": 0.015445544554455445,
      "calib/nonempty_final_conf_rate": 0.953125,
      "calib/nonempty_reasoning_rate": 0.95703125,
      "calib/nonempty_step_conf_rate": 0.953125,
      "calib/pce": 0.004508196721311476,
      "calib/std_conf": 0.11026463976292598,
      "calib/step_conf_rate": 0.953125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2866.0,
      "completions/max_terminated_length": 2866.0,
      "completions/mean_length": 396.515625,
      "completions/mean_terminated_length": 398.07061767578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 74.0,
      "epoch": 0.10986666666666667,
      "grad_norm": 0.009121976792812347,
      "learning_rate": 2.6944444444444444e-06,
      "loss": 0.214,
      "num_tokens": 19903923.0,
      "reward": 1.2337702512741089,
      "reward_std": 0.2346125990152359,
      "rewards/accuracy_reward_step": 0.55859375,
      "rewards/final_brier_reward_step": 0.4011343717575073,
      "rewards/format_reward_step": 0.94921875,
      "step": 103
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3781659468077123,
      "aux_distill/mean_u": 0.16131986090351855,
      "aux_distill/n_active_tok": 85.78125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.478180307967542,
      "calib/avg_num_step_conf": 1.33984375,
      "calib/ece": 0.43587301587301586,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.007936507936507936,
      "calib/gap": 0.0010964155645006683,
      "calib/mean_conf": 0.01253968253968254,
      "calib/mu_c": 0.013153153153153152,
      "calib/mu_w": 0.012056737588652484,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.9765625,
      "calib/pce": 0.003968253968253968,
      "calib/std_conf": 0.08897954560688179,
      "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": 1861.0,
      "completions/max_terminated_length": 1861.0,
      "completions/mean_length": 317.05859375,
      "completions/mean_terminated_length": 317.05859375,
      "completions/min_length": 86.0,
      "completions/min_terminated_length": 86.0,
      "epoch": 0.11093333333333333,
      "grad_norm": 0.00879514031112194,
      "learning_rate": 2.666666666666667e-06,
      "loss": 0.1337,
      "num_tokens": 20091770.0,
      "reward": 1.1931354999542236,
      "reward_std": 0.19186003506183624,
      "rewards/accuracy_reward_step": 0.43359375,
      "rewards/final_brier_reward_step": 0.5464273691177368,
      "rewards/format_reward_step": 0.97265625,
      "step": 104
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3944840473122895,
      "aux_distill/mean_u": 0.1414717727707358,
      "aux_distill/n_active_tok": 80.5,
      "calib/answer_extract_rate": 0.95703125,
      "calib/auroc": 0.49540495867768597,
      "calib/avg_num_step_conf": 1.26171875,
      "calib/ece": 0.5008130081300813,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 0.012195121951219513,
      "calib/gap": -0.02308099173553719,
      "calib/mean_conf": 0.015447154471544714,
      "calib/mu_c": 0.00371900826446281,
      "calib/mu_w": 0.026799999999999997,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 0.9609375,
      "calib/nonempty_step_conf_rate": 0.94140625,
      "calib/pce": 0.012195121951219513,
      "calib/std_conf": 0.10962057275702294,
      "calib/step_conf_rate": 0.94140625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2820.0,
      "completions/max_terminated_length": 2820.0,
      "completions/mean_length": 367.71875,
      "completions/mean_terminated_length": 370.6141662597656,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.112,
      "grad_norm": 0.00771205173805356,
      "learning_rate": 2.6388888888888893e-06,
      "loss": 0.1899,
      "num_tokens": 20291666.0,
      "reward": 1.1714603900909424,
      "reward_std": 0.302039235830307,
      "rewards/accuracy_reward_step": 0.47265625,
      "rewards/final_brier_reward_step": 0.46010822057724,
      "rewards/format_reward_step": 0.9375,
      "step": 105
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4089339836500585,
      "aux_distill/mean_u": 0.1628464583554884,
      "aux_distill/n_active_tok": 73.78125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.48087277736038064,
      "calib/avg_num_step_conf": 1.15234375,
      "calib/ece": 0.5207166007905139,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.003952569169960474,
      "calib/gap": -0.011628374655647382,
      "calib/mean_conf": 0.011773517786561265,
      "calib/mu_c": 0.006212121212121212,
      "calib/mu_w": 0.017840495867768594,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.97265625,
      "calib/pce": 0.005375494071146245,
      "calib/std_conf": 0.06562255566896466,
      "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": 2436.0,
      "completions/max_terminated_length": 2436.0,
      "completions/mean_length": 318.90234375,
      "completions/mean_terminated_length": 320.1529541015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 109.0,
      "epoch": 0.11306666666666666,
      "grad_norm": 0.009178507141768932,
      "learning_rate": 2.6111111111111113e-06,
      "loss": 0.0854,
      "num_tokens": 20477889.0,
      "reward": 1.230987548828125,
      "reward_std": 0.17167076468467712,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/final_brier_reward_step": 0.46197509765625,
      "rewards/format_reward_step": 0.96875,
      "step": 106
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3847769391722977,
      "aux_distill/mean_u": 0.1593563057381601,
      "aux_distill/n_active_tok": 77.78125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.42153976975405544,
      "calib/avg_num_step_conf": 1.21484375,
      "calib/ece": 0.5882629482071713,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.01593625498007968,
      "calib/gap": -0.026485609628466772,
      "calib/mean_conf": 0.022334661354581672,
      "calib/mu_c": 0.011360544217687077,
      "calib/mu_w": 0.03784615384615385,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.01247011952191235,
      "calib/std_conf": 0.12520179549592428,
      "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": 2587.0,
      "completions/max_terminated_length": 2587.0,
      "completions/mean_length": 315.31640625,
      "completions/mean_terminated_length": 316.5529479980469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 91.0,
      "epoch": 0.11413333333333334,
      "grad_norm": 0.0072776032611727715,
      "learning_rate": 2.5833333333333337e-06,
      "loss": 0.0908,
      "num_tokens": 20663226.0,
      "reward": 1.2622661590576172,
      "reward_std": 0.20053136348724365,
      "rewards/accuracy_reward_step": 0.57421875,
      "rewards/final_brier_reward_step": 0.3995321989059448,
      "rewards/format_reward_step": 0.9765625,
      "step": 107
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36748338444158435,
      "aux_distill/mean_u": 0.16356597389892147,
      "aux_distill/n_active_tok": 74.03125,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.4810773074661963,
      "calib/avg_num_step_conf": 1.15625,
      "calib/ece": 0.6667469879518073,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00010802469135802462,
      "calib/mean_conf": 0.007951807228915662,
      "calib/mu_c": 0.007916666666666667,
      "calib/mu_w": 0.008024691358024692,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.9765625,
      "calib/nonempty_step_conf_rate": 0.97265625,
      "calib/pce": 0.0,
      "calib/std_conf": 0.018762457354271893,
      "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": 2877.0,
      "completions/max_terminated_length": 2877.0,
      "completions/mean_length": 357.296875,
      "completions/mean_terminated_length": 358.69805908203125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 93.0,
      "epoch": 0.1152,
      "grad_norm": 0.009626587852835655,
      "learning_rate": 2.5555555555555557e-06,
      "loss": 0.0949,
      "num_tokens": 20857926.0,
      "reward": 1.3018684387207031,
      "reward_std": 0.18892589211463928,
      "rewards/accuracy_reward_step": 0.65625,
      "rewards/final_brier_reward_step": 0.3224866986274719,
      "rewards/format_reward_step": 0.96875,
      "step": 108
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.343574243132025,
      "aux_distill/mean_u": 0.10433978655256294,
      "aux_distill/n_active_tok": 77.28125,
      "calib/answer_extract_rate": 0.9453125,
      "calib/auroc": 0.4994190814652815,
      "calib/avg_num_step_conf": 1.20703125,
      "calib/ece": 0.4774793388429752,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.9140625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.008027610716238382,
      "calib/mean_conf": 0.010123966942148762,
      "calib/mu_c": 0.014237288135593221,
      "calib/mu_w": 0.006209677419354838,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 0.9609375,
      "calib/nonempty_step_conf_rate": 0.9296875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.05285627952372183,
      "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": 0.01171875,
      "completions/max_length": 2876.0,
      "completions/max_terminated_length": 2876.0,
      "completions/mean_length": 404.6328125,
      "completions/mean_terminated_length": 409.43084716796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 93.0,
      "epoch": 0.11626666666666667,
      "grad_norm": 0.008727338165044785,
      "learning_rate": 2.5277777777777778e-06,
      "loss": 0.1671,
      "num_tokens": 21066112.0,
      "reward": 1.1536353826522827,
      "reward_std": 0.2425643503665924,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/final_brier_reward_step": 0.47133320569992065,
      "rewards/format_reward_step": 0.9140625,
      "step": 109
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36572849517688155,
      "aux_distill/mean_u": 0.1205168877068942,
      "aux_distill/n_active_tok": 67.28125,
      "calib/answer_extract_rate": 0.96484375,
      "calib/auroc": 0.4670535011801731,
      "calib/avg_num_step_conf": 1.05078125,
      "calib/ece": 0.5018623481781377,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.953125,
      "calib/frac_conf_gt_0.9": 0.004048582995951417,
      "calib/gap": -0.014848544453186466,
      "calib/mean_conf": 0.013765182186234818,
      "calib/mu_c": 0.006370967741935484,
      "calib/mu_w": 0.02121951219512195,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 0.96875,
      "calib/nonempty_step_conf_rate": 0.95703125,
      "calib/pce": 0.006801619433198381,
      "calib/std_conf": 0.07747530311904471,
      "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": 0.0,
      "completions/max_length": 2540.0,
      "completions/max_terminated_length": 2540.0,
      "completions/mean_length": 326.76171875,
      "completions/mean_terminated_length": 326.76171875,
      "completions/min_length": 74.0,
      "completions/min_terminated_length": 74.0,
      "epoch": 0.11733333333333333,
      "grad_norm": 0.008508248254656792,
      "learning_rate": 2.5e-06,
      "loss": 0.1231,
      "num_tokens": 21254683.0,
      "reward": 1.1972490549087524,
      "reward_std": 0.22007989883422852,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.4726230502128601,
      "rewards/format_reward_step": 0.953125,
      "step": 110
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3720876621082425,
      "aux_distill/mean_u": 0.15341549175611296,
      "aux_distill/n_active_tok": 68.3125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5270833333333333,
      "calib/avg_num_step_conf": 1.06640625,
      "calib/ece": 0.5183730158730159,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.002204545454545454,
      "calib/mean_conf": 0.008928571428571428,
      "calib/mu_c": 0.00787878787878788,
      "calib/mu_w": 0.010083333333333333,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.001746031746031746,
      "calib/std_conf": 0.030329585998626372,
      "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": 1772.0,
      "completions/max_terminated_length": 1772.0,
      "completions/mean_length": 327.69921875,
      "completions/mean_terminated_length": 327.69921875,
      "completions/min_length": 84.0,
      "completions/min_terminated_length": 84.0,
      "epoch": 0.1184,
      "grad_norm": 0.008607271127402782,
      "learning_rate": 2.4722222222222226e-06,
      "loss": 0.0551,
      "num_tokens": 21445982.0,
      "reward": 1.239898681640625,
      "reward_std": 0.1710304617881775,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/final_brier_reward_step": 0.47198477387428284,
      "rewards/format_reward_step": 0.9765625,
      "step": 111
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38653289526700974,
      "aux_distill/mean_u": 0.183389772523197,
      "aux_distill/n_active_tok": 62.84375,
      "calib/answer_extract_rate": 0.94140625,
      "calib/auroc": 0.5073285398230089,
      "calib/avg_num_step_conf": 0.98828125,
      "calib/ece": 0.5232780082987551,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.90625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0004362555309734501,
      "calib/mean_conf": 0.007842323651452283,
      "calib/mu_c": 0.008046874999999998,
      "calib/mu_w": 0.007610619469026548,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 0.953125,
      "calib/nonempty_step_conf_rate": 0.91796875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.013827330009136139,
      "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": 0.02734375,
      "completions/max_length": 2946.0,
      "completions/max_terminated_length": 2946.0,
      "completions/mean_length": 389.703125,
      "completions/mean_terminated_length": 400.65863037109375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 99.0,
      "epoch": 0.11946666666666667,
      "grad_norm": 0.008142063394188881,
      "learning_rate": 2.4444444444444447e-06,
      "loss": 0.1322,
      "num_tokens": 21653666.0,
      "reward": 1.1659774780273438,
      "reward_std": 0.24817338585853577,
      "rewards/accuracy_reward_step": 0.5,
      "rewards/final_brier_reward_step": 0.42570507526397705,
      "rewards/format_reward_step": 0.90625,
      "step": 112
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32456047693267465,
      "aux_distill/mean_u": 0.10668147590575155,
      "aux_distill/n_active_tok": 64.78125,
      "calib/answer_extract_rate": 0.9296875,
      "calib/auroc": 0.5366289458010721,
      "calib/avg_num_step_conf": 1.01171875,
      "calib/ece": 0.6068907563025211,
      "calib/final_conf_rate": 0.9296875,
      "calib/format_rate": 0.921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.001116736152471709,
      "calib/mean_conf": 0.006554621848739496,
      "calib/mu_c": 0.006986301369863014,
      "calib/mu_w": 0.005869565217391305,
      "calib/nonempty_final_conf_rate": 0.9296875,
      "calib/nonempty_reasoning_rate": 0.9375,
      "calib/nonempty_step_conf_rate": 0.9296875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.011147368441759219,
      "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": 2135.0,
      "completions/max_terminated_length": 2135.0,
      "completions/mean_length": 338.66015625,
      "completions/mean_terminated_length": 344.0357360839844,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 99.0,
      "epoch": 0.12053333333333334,
      "grad_norm": 0.007761706598103046,
      "learning_rate": 2.4166666666666667e-06,
      "loss": 0.1085,
      "num_tokens": 21845563.0,
      "reward": 1.2109379768371582,
      "reward_std": 0.21525073051452637,
      "rewards/accuracy_reward_step": 0.5703125,
      "rewards/final_brier_reward_step": 0.3593761622905731,
      "rewards/format_reward_step": 0.921875,
      "step": 113
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3718952308408916,
      "aux_distill/mean_u": 0.16472218902211183,
      "aux_distill/n_active_tok": 69.8125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5613652868554829,
      "calib/avg_num_step_conf": 1.08984375,
      "calib/ece": 0.5928968253968254,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": 0.011984551396316104,
      "calib/mean_conf": 0.014246031746031748,
      "calib/mu_c": 0.018954248366013074,
      "calib/mu_w": 0.00696969696969697,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.06800317295444387,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1325.0,
      "completions/max_terminated_length": 1325.0,
      "completions/mean_length": 291.2421875,
      "completions/mean_terminated_length": 292.38433837890625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 100.0,
      "epoch": 0.1216,
      "grad_norm": 0.010775484144687653,
      "learning_rate": 2.388888888888889e-06,
      "loss": 0.097,
      "num_tokens": 22025145.0,
      "reward": 1.286218523979187,
      "reward_std": 0.17462275922298431,
      "rewards/accuracy_reward_step": 0.59765625,
      "rewards/final_brier_reward_step": 0.4005621075630188,
      "rewards/format_reward_step": 0.9765625,
      "step": 114
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3595115006901324,
      "aux_distill/mean_u": 0.12506166863110565,
      "aux_distill/n_active_tok": 67.78125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.4552986391129032,
      "calib/avg_num_step_conf": 1.0625,
      "calib/ece": 0.48904761904761906,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": -0.009470766129032263,
      "calib/mean_conf": 0.011746031746031746,
      "calib/mu_c": 0.0069354838709677425,
      "calib/mu_w": 0.016406250000000004,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.004365079365079365,
      "calib/std_conf": 0.06352201674409502,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2080.0,
      "completions/max_terminated_length": 2080.0,
      "completions/mean_length": 299.15625,
      "completions/mean_terminated_length": 300.3294372558594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 106.0,
      "epoch": 0.12266666666666666,
      "grad_norm": 0.011229630559682846,
      "learning_rate": 2.361111111111111e-06,
      "loss": 0.026,
      "num_tokens": 22206993.0,
      "reward": 1.223961591720581,
      "reward_std": 0.18679305911064148,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.4987046718597412,
      "rewards/format_reward_step": 0.98046875,
      "step": 115
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3685760064981878,
      "aux_distill/mean_u": 0.11651055035050567,
      "aux_distill/n_active_tok": 71.59375,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.5117604187371629,
      "calib/avg_num_step_conf": 1.1171875,
      "calib/ece": 0.5143089430894309,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0034148280659908586,
      "calib/mean_conf": 0.01008130081300813,
      "calib/mu_c": 0.011705426356589149,
      "calib/mu_w": 0.00829059829059829,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 0.98046875,
      "calib/nonempty_step_conf_rate": 0.9765625,
      "calib/pce": 0.0,
      "calib/std_conf": 0.029590592192035723,
      "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": 3009.0,
      "completions/max_terminated_length": 3009.0,
      "completions/mean_length": 353.63671875,
      "completions/mean_terminated_length": 355.0235595703125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 59.0,
      "epoch": 0.12373333333333333,
      "grad_norm": 0.008860284462571144,
      "learning_rate": 2.3333333333333336e-06,
      "loss": 0.1086,
      "num_tokens": 22402044.0,
      "reward": 1.2144132852554321,
      "reward_std": 0.23440314829349518,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.46398282051086426,
      "rewards/format_reward_step": 0.95703125,
      "step": 116
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3026649123057723,
      "aux_distill/mean_u": 0.07919790339599429,
      "aux_distill/n_active_tok": 63.96875,
      "calib/answer_extract_rate": 0.9453125,
      "calib/auroc": 0.4968918069427953,
      "calib/avg_num_step_conf": 0.99609375,
      "calib/ece": 0.4164876033057851,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.008939722008800728,
      "calib/mean_conf": 0.01756198347107438,
      "calib/mu_c": 0.012427184466019418,
      "calib/mu_w": 0.021366906474820146,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 0.97265625,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.004214876033057851,
      "calib/std_conf": 0.06447696655318512,
      "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": 0.01171875,
      "completions/max_length": 2947.0,
      "completions/max_terminated_length": 2947.0,
      "completions/mean_length": 320.8828125,
      "completions/mean_terminated_length": 324.6877746582031,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 103.0,
      "epoch": 0.1248,
      "grad_norm": 0.011900389567017555,
      "learning_rate": 2.305555555555556e-06,
      "loss": 0.0455,
      "num_tokens": 22590790.0,
      "reward": 1.1454674005508423,
      "reward_std": 0.20846307277679443,
      "rewards/accuracy_reward_step": 0.40234375,
      "rewards/final_brier_reward_step": 0.5448409914970398,
      "rewards/format_reward_step": 0.94140625,
      "step": 117
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2679017949849367,
      "aux_distill/mean_u": 0.0878350996942736,
      "aux_distill/n_active_tok": 67.5625,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.5327660994327661,
      "calib/avg_num_step_conf": 1.05859375,
      "calib/ece": 0.53,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.008130081300813009,
      "calib/gap": 0.019511511511511516,
      "calib/mean_conf": 0.025121951219512193,
      "calib/mu_c": 0.03392592592592593,
      "calib/mu_w": 0.014414414414414415,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 0.96875,
      "calib/nonempty_step_conf_rate": 0.94921875,
      "calib/pce": 0.003170731707317074,
      "calib/std_conf": 0.11228495873075943,
      "calib/step_conf_rate": 0.94921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3029.0,
      "completions/max_terminated_length": 3029.0,
      "completions/mean_length": 346.22265625,
      "completions/mean_terminated_length": 347.5804138183594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 95.0,
      "epoch": 0.12586666666666665,
      "grad_norm": 0.01450740359723568,
      "learning_rate": 2.277777777777778e-06,
      "loss": 0.0959,
      "num_tokens": 22783431.0,
      "reward": 1.21885347366333,
      "reward_std": 0.24840456247329712,
      "rewards/accuracy_reward_step": 0.52734375,
      "rewards/final_brier_reward_step": 0.4416132867336273,
      "rewards/format_reward_step": 0.94140625,
      "step": 118
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32812713412567973,
      "aux_distill/mean_u": 0.12670988056552968,
      "aux_distill/n_active_tok": 62.75,
      "calib/answer_extract_rate": 0.8984375,
      "calib/auroc": 0.5259686102054743,
      "calib/avg_num_step_conf": 0.97265625,
      "calib/ece": 0.5122173913043478,
      "calib/final_conf_rate": 0.8984375,
      "calib/format_rate": 0.89453125,
      "calib/frac_conf_gt_0.9": 0.004347826086956522,
      "calib/gap": 0.011715065584957156,
      "calib/mean_conf": 0.013869565217391307,
      "calib/mu_c": 0.01942148760330578,
      "calib/mu_w": 0.007706422018348625,
      "calib/nonempty_final_conf_rate": 0.8984375,
      "calib/nonempty_reasoning_rate": 0.96484375,
      "calib/nonempty_step_conf_rate": 0.9609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.068659718174567,
      "calib/step_conf_rate": 0.9609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2830.0,
      "completions/max_terminated_length": 2830.0,
      "completions/mean_length": 362.19921875,
      "completions/mean_terminated_length": 365.0511779785156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 84.0,
      "epoch": 0.12693333333333334,
      "grad_norm": 0.009933991357684135,
      "learning_rate": 2.25e-06,
      "loss": 0.1331,
      "num_tokens": 22981218.0,
      "reward": 1.1378350257873535,
      "reward_std": 0.33909130096435547,
      "rewards/accuracy_reward_step": 0.47265625,
      "rewards/final_brier_reward_step": 0.4358261823654175,
      "rewards/format_reward_step": 0.89453125,
      "step": 119
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2903097663074732,
      "aux_distill/mean_u": 0.09879584451015042,
      "aux_distill/n_active_tok": 65.34375,
      "calib/answer_extract_rate": 0.91015625,
      "calib/auroc": 0.5062885802469136,
      "calib/avg_num_step_conf": 1.01171875,
      "calib/ece": 0.5657575757575758,
      "calib/final_conf_rate": 0.90234375,
      "calib/format_rate": 0.88671875,
      "calib/frac_conf_gt_0.9": 0.012987012987012988,
      "calib/gap": 0.010914351851851852,
      "calib/mean_conf": 0.027316017316017318,
      "calib/mu_c": 0.03185185185185185,
      "calib/mu_w": 0.0209375,
      "calib/nonempty_final_conf_rate": 0.90234375,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.00432900432900433,
      "calib/std_conf": 0.12327310915368074,
      "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": 0.0,
      "completions/max_length": 2181.0,
      "completions/max_terminated_length": 2181.0,
      "completions/mean_length": 306.7109375,
      "completions/mean_terminated_length": 306.7109375,
      "completions/min_length": 90.0,
      "completions/min_terminated_length": 90.0,
      "epoch": 0.128,
      "grad_norm": 0.011160912923514843,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.0877,
      "num_tokens": 23166424.0,
      "reward": 1.1697453260421753,
      "reward_std": 0.26618820428848267,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/final_brier_reward_step": 0.3824594020843506,
      "rewards/format_reward_step": 0.88671875,
      "step": 120
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.324223575880751,
      "aux_distill/mean_u": 0.08532104738627432,
      "aux_distill/n_active_tok": 65.875,
      "calib/answer_extract_rate": 0.8828125,
      "calib/auroc": 0.5099842767295597,
      "calib/avg_num_step_conf": 1.015625,
      "calib/ece": 0.5158407079646018,
      "calib/final_conf_rate": 0.8828125,
      "calib/format_rate": 0.8828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.002559748427672955,
      "calib/mean_conf": 0.015132743362831859,
      "calib/mu_c": 0.01633333333333333,
      "calib/mu_w": 0.013773584905660377,
      "calib/nonempty_final_conf_rate": 0.8828125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.05632734673157518,
      "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": 2771.0,
      "completions/max_terminated_length": 2771.0,
      "completions/mean_length": 323.62109375,
      "completions/mean_terminated_length": 323.62109375,
      "completions/min_length": 91.0,
      "completions/min_terminated_length": 91.0,
      "epoch": 0.12906666666666666,
      "grad_norm": 0.012316557578742504,
      "learning_rate": 2.1944444444444445e-06,
      "loss": 0.0656,
      "num_tokens": 23354327.0,
      "reward": 1.1233421564102173,
      "reward_std": 0.32146894931793213,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.42637187242507935,
      "rewards/format_reward_step": 0.8828125,
      "step": 121
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.345096837496385,
      "aux_distill/mean_u": 0.1192782161457965,
      "aux_distill/n_active_tok": 65.125,
      "calib/answer_extract_rate": 0.90234375,
      "calib/auroc": 0.5341607565011821,
      "calib/avg_num_step_conf": 1.0078125,
      "calib/ece": 0.5942419913419914,
      "calib/final_conf_rate": 0.90234375,
      "calib/format_rate": 0.8984375,
      "calib/frac_conf_gt_0.9": 0.004329004329004329,
      "calib/gap": 0.010071631205673764,
      "calib/mean_conf": 0.01614761904761905,
      "calib/mu_c": 0.020071631205673762,
      "calib/mu_w": 0.009999999999999998,
      "calib/nonempty_final_conf_rate": 0.90234375,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.07051663941031937,
      "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": 0.01171875,
      "completions/max_length": 2089.0,
      "completions/max_terminated_length": 2089.0,
      "completions/mean_length": 264.42578125,
      "completions/mean_terminated_length": 267.561279296875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 109.0,
      "epoch": 0.13013333333333332,
      "grad_norm": 0.01495366357266903,
      "learning_rate": 2.166666666666667e-06,
      "loss": 0.0432,
      "num_tokens": 23529364.0,
      "reward": 1.1825222969055176,
      "reward_std": 0.26760587096214294,
      "rewards/accuracy_reward_step": 0.55078125,
      "rewards/final_brier_reward_step": 0.3650445342063904,
      "rewards/format_reward_step": 0.8984375,
      "step": 122
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35667229630053043,
      "aux_distill/mean_u": 0.1399172233377629,
      "aux_distill/n_active_tok": 71.625,
      "calib/answer_extract_rate": 0.8828125,
      "calib/auroc": 0.5353483606557377,
      "calib/avg_num_step_conf": 1.10546875,
      "calib/ece": 0.4516814159292035,
      "calib/final_conf_rate": 0.8828125,
      "calib/format_rate": 0.8828125,
      "calib/frac_conf_gt_0.9": 0.017699115044247787,
      "calib/gap": 0.000890605296342991,
      "calib/mean_conf": 0.03,
      "calib/mu_c": 0.03048076923076922,
      "calib/mu_w": 0.02959016393442623,
      "calib/nonempty_final_conf_rate": 0.8828125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.010752212389380532,
      "calib/std_conf": 0.13212691672741708,
      "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": 2506.0,
      "completions/max_terminated_length": 2506.0,
      "completions/mean_length": 330.60546875,
      "completions/mean_terminated_length": 331.9019775390625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 120.0,
      "epoch": 0.1312,
      "grad_norm": 0.012956635095179081,
      "learning_rate": 2.138888888888889e-06,
      "loss": 0.1546,
      "num_tokens": 23719287.0,
      "reward": 1.090217113494873,
      "reward_std": 0.332817405462265,
      "rewards/accuracy_reward_step": 0.40625,
      "rewards/final_brier_reward_step": 0.48512187600135803,
      "rewards/format_reward_step": 0.8828125,
      "step": 123
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.303095584269613,
      "aux_distill/mean_u": 0.10875086624948242,
      "aux_distill/n_active_tok": 69.6875,
      "calib/answer_extract_rate": 0.9453125,
      "calib/auroc": 0.4850164203612479,
      "calib/avg_num_step_conf": 1.0859375,
      "calib/ece": 0.48252066115702474,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.02066115702479339,
      "calib/gap": -0.024482758620689653,
      "calib/mean_conf": 0.03326446280991736,
      "calib/mu_c": 0.02051724137931034,
      "calib/mu_w": 0.04499999999999999,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.01822314049586777,
      "calib/std_conf": 0.14261027448960573,
      "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": 596.0,
      "completions/max_terminated_length": 596.0,
      "completions/mean_length": 271.0859375,
      "completions/mean_terminated_length": 272.1490478515625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 91.0,
      "epoch": 0.13226666666666667,
      "grad_norm": 0.01391842681914568,
      "learning_rate": 2.1111111111111114e-06,
      "loss": 0.0442,
      "num_tokens": 23895501.0,
      "reward": 1.1690056324005127,
      "reward_std": 0.2866358757019043,
      "rewards/accuracy_reward_step": 0.453125,
      "rewards/final_brier_reward_step": 0.49035507440567017,
      "rewards/format_reward_step": 0.94140625,
      "step": 124
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.29174726200290024,
      "aux_distill/mean_u": 0.11426094989438287,
      "aux_distill/n_active_tok": 73.5,
      "calib/answer_extract_rate": 0.91796875,
      "calib/auroc": 0.5308159722222223,
      "calib/avg_num_step_conf": 1.140625,
      "calib/ece": 0.438135593220339,
      "calib/final_conf_rate": 0.921875,
      "calib/format_rate": 0.91796875,
      "calib/frac_conf_gt_0.9": 0.0211864406779661,
      "calib/gap": 0.018342013888888894,
      "calib/mean_conf": 0.036440677966101696,
      "calib/mu_c": 0.046388888888888896,
      "calib/mu_w": 0.028046875000000002,
      "calib/nonempty_final_conf_rate": 0.921875,
      "calib/nonempty_reasoning_rate": 0.98046875,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.00847457627118644,
      "calib/std_conf": 0.1446710202770926,
      "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": 0.015625,
      "completions/max_length": 1559.0,
      "completions/max_terminated_length": 1559.0,
      "completions/mean_length": 266.5859375,
      "completions/mean_terminated_length": 270.8174743652344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 70.0,
      "epoch": 0.13333333333333333,
      "grad_norm": 0.013951667584478855,
      "learning_rate": 2.0833333333333334e-06,
      "loss": 0.0304,
      "num_tokens": 24068555.0,
      "reward": 1.1382172107696533,
      "reward_std": 0.23240628838539124,
      "rewards/accuracy_reward_step": 0.421875,
      "rewards/final_brier_reward_step": 0.5147156119346619,
      "rewards/format_reward_step": 0.91796875,
      "step": 125
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36404131818562746,
      "aux_distill/mean_u": 0.13878008918581453,
      "aux_distill/n_active_tok": 71.1875,
      "calib/answer_extract_rate": 0.89453125,
      "calib/auroc": 0.506964217006541,
      "calib/avg_num_step_conf": 1.1015625,
      "calib/ece": 0.4953947368421052,
      "calib/final_conf_rate": 0.890625,
      "calib/format_rate": 0.890625,
      "calib/frac_conf_gt_0.9": 0.017543859649122806,
      "calib/gap": -0.01719969218930357,
      "calib/mean_conf": 0.035307017543859634,
      "calib/mu_c": 0.026782608695652174,
      "calib/mu_w": 0.04398230088495574,
      "calib/nonempty_final_conf_rate": 0.890625,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.013157894736842105,
      "calib/std_conf": 0.1323916059241489,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2965.0,
      "completions/max_terminated_length": 2965.0,
      "completions/mean_length": 292.92578125,
      "completions/mean_terminated_length": 294.07452392578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 51.0,
      "epoch": 0.1344,
      "grad_norm": 0.013217564672231674,
      "learning_rate": 2.0555555555555555e-06,
      "loss": 0.1714,
      "num_tokens": 24249008.0,
      "reward": 1.1189053058624268,
      "reward_std": 0.3188907504081726,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.44874805212020874,
      "rewards/format_reward_step": 0.890625,
      "step": 126
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3291368675418198,
      "aux_distill/mean_u": 0.11664523604299605,
      "aux_distill/n_active_tok": 66.5,
      "calib/answer_extract_rate": 0.8984375,
      "calib/auroc": 0.5259813932380304,
      "calib/avg_num_step_conf": 1.03125,
      "calib/ece": 0.47982608695652174,
      "calib/final_conf_rate": 0.8984375,
      "calib/format_rate": 0.89453125,
      "calib/frac_conf_gt_0.9": 0.0391304347826087,
      "calib/gap": 0.026621284320399362,
      "calib/mean_conf": 0.05486956521739131,
      "calib/mu_c": 0.06794871794871794,
      "calib/mu_w": 0.04132743362831858,
      "calib/nonempty_final_conf_rate": 0.8984375,
      "calib/nonempty_reasoning_rate": 0.96484375,
      "calib/nonempty_step_conf_rate": 0.9609375,
      "calib/pce": 0.013000000000000005,
      "calib/std_conf": 0.19233966882616488,
      "calib/step_conf_rate": 0.9609375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2925.0,
      "completions/max_terminated_length": 2925.0,
      "completions/mean_length": 293.3359375,
      "completions/mean_terminated_length": 296.8142395019531,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 81.0,
      "epoch": 0.13546666666666668,
      "grad_norm": 0.012097050435841084,
      "learning_rate": 2.027777777777778e-06,
      "loss": 0.1988,
      "num_tokens": 24427774.0,
      "reward": 1.138083577156067,
      "reward_std": 0.3193005919456482,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/final_brier_reward_step": 0.46757346391677856,
      "rewards/format_reward_step": 0.89453125,
      "step": 127
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3628509286791086,
      "aux_distill/mean_u": 0.13717759212656766,
      "aux_distill/n_active_tok": 68.78125,
      "calib/answer_extract_rate": 0.8828125,
      "calib/auroc": 0.50538571202281,
      "calib/avg_num_step_conf": 1.0625,
      "calib/ece": 0.4724888888888889,
      "calib/final_conf_rate": 0.87890625,
      "calib/format_rate": 0.87890625,
      "calib/frac_conf_gt_0.9": 0.017777777777777778,
      "calib/gap": -0.016759068588626646,
      "calib/mean_conf": 0.03355555555555556,
      "calib/mu_c": 0.02476635514018691,
      "calib/mu_w": 0.04152542372881356,
      "calib/nonempty_final_conf_rate": 0.87890625,
      "calib/nonempty_reasoning_rate": 0.98046875,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.015244444444444444,
      "calib/std_conf": 0.13192852535547092,
      "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": 2780.0,
      "completions/max_terminated_length": 2780.0,
      "completions/mean_length": 279.1953125,
      "completions/mean_terminated_length": 280.29022216796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 82.0,
      "epoch": 0.13653333333333334,
      "grad_norm": 0.01438905205577612,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0887,
      "num_tokens": 24605912.0,
      "reward": 1.0940048694610596,
      "reward_std": 0.27575862407684326,
      "rewards/accuracy_reward_step": 0.421875,
      "rewards/final_brier_reward_step": 0.46535351872444153,
      "rewards/format_reward_step": 0.87890625,
      "step": 128
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3082141927443445,
      "aux_distill/mean_u": 0.11286577517381002,
      "aux_distill/n_active_tok": 70.15625,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.4738412698412699,
      "calib/avg_num_step_conf": 1.09375,
      "calib/ece": 0.49418326693227094,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.06374501992031872,
      "calib/gap": -0.017893333333333372,
      "calib/mean_conf": 0.08557768924302789,
      "calib/mu_c": 0.07666666666666665,
      "calib/mu_w": 0.09456000000000002,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03888446215139442,
      "calib/std_conf": 0.2447145766889721,
      "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": 776.0,
      "completions/max_terminated_length": 776.0,
      "completions/mean_length": 226.61328125,
      "completions/mean_terminated_length": 226.61328125,
      "completions/min_length": 88.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.1376,
      "grad_norm": 0.014954674988985062,
      "learning_rate": 1.9722222222222224e-06,
      "loss": 0.051,
      "num_tokens": 24766309.0,
      "reward": 1.2313487529754639,
      "reward_std": 0.21718278527259827,
      "rewards/accuracy_reward_step": 0.4921875,
      "rewards/final_brier_reward_step": 0.49785393476486206,
      "rewards/format_reward_step": 0.98046875,
      "step": 129
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31981189316138625,
      "aux_distill/mean_u": 0.11389575257731732,
      "aux_distill/n_active_tok": 70.53125,
      "calib/answer_extract_rate": 0.95703125,
      "calib/auroc": 0.48054774340431233,
      "calib/avg_num_step_conf": 1.09765625,
      "calib/ece": 0.47861224489795906,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.04897959183673469,
      "calib/gap": 0.05308490692379804,
      "calib/mean_conf": 0.07461224489795919,
      "calib/mu_c": 0.09931297709923663,
      "calib/mu_w": 0.04622807017543859,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.009265306122448977,
      "calib/std_conf": 0.2182367948481023,
      "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": 1839.0,
      "completions/max_terminated_length": 1839.0,
      "completions/mean_length": 228.3671875,
      "completions/mean_terminated_length": 229.26275634765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 74.0,
      "epoch": 0.13866666666666666,
      "grad_norm": 0.015271571464836597,
      "learning_rate": 1.944444444444445e-06,
      "loss": 0.084,
      "num_tokens": 24930059.0,
      "reward": 1.2382566928863525,
      "reward_std": 0.19215700030326843,
      "rewards/accuracy_reward_step": 0.51171875,
      "rewards/final_brier_reward_step": 0.4960445463657379,
      "rewards/format_reward_step": 0.95703125,
      "step": 130
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3160778861492872,
      "aux_distill/mean_u": 0.10070666133135303,
      "aux_distill/n_active_tok": 74.0625,
      "calib/answer_extract_rate": 0.95703125,
      "calib/auroc": 0.4863674496644296,
      "calib/avg_num_step_conf": 1.15625,
      "calib/ece": 0.38048979591836735,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.0653061224489796,
      "calib/gap": 0.011264681208053712,
      "calib/mean_conf": 0.08783673469387754,
      "calib/mu_c": 0.09468750000000002,
      "calib/mu_w": 0.08342281879194631,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.03824489795918368,
      "calib/std_conf": 0.24457671447381732,
      "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": 0.01171875,
      "completions/max_length": 1911.0,
      "completions/max_terminated_length": 1911.0,
      "completions/mean_length": 234.828125,
      "completions/mean_terminated_length": 237.61265563964844,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 101.0,
      "epoch": 0.13973333333333332,
      "grad_norm": 0.015647603198885918,
      "learning_rate": 1.916666666666667e-06,
      "loss": 0.0495,
      "num_tokens": 25096383.0,
      "reward": 1.1477234363555908,
      "reward_std": 0.19904610514640808,
      "rewards/accuracy_reward_step": 0.375,
      "rewards/final_brier_reward_step": 0.5884156227111816,
      "rewards/format_reward_step": 0.95703125,
      "step": 131
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31305989855900407,
      "aux_distill/mean_u": 0.14130636641858182,
      "aux_distill/n_active_tok": 71.90625,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.4506064162754303,
      "calib/avg_num_step_conf": 1.12109375,
      "calib/ece": 0.55212,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.1,
      "calib/gap": -0.04311163275952008,
      "calib/mean_conf": 0.12292,
      "calib/mu_c": 0.10429577464788732,
      "calib/mu_w": 0.1474074074074074,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.053520000000000005,
      "calib/std_conf": 0.29644944526849765,
      "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": 632.0,
      "completions/max_terminated_length": 632.0,
      "completions/mean_length": 221.77734375,
      "completions/mean_terminated_length": 222.64707946777344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 79.0,
      "epoch": 0.1408,
      "grad_norm": 0.015256309881806374,
      "learning_rate": 1.888888888888889e-06,
      "loss": 0.0557,
      "num_tokens": 25258750.0,
      "reward": 1.2594385147094727,
      "reward_std": 0.2623821496963501,
      "rewards/accuracy_reward_step": 0.5546875,
      "rewards/final_brier_reward_step": 0.43684569001197815,
      "rewards/format_reward_step": 0.97265625,
      "step": 132
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3371236342936754,
      "aux_distill/mean_u": 0.10641052422516785,
      "aux_distill/n_active_tok": 72.28125,
      "calib/answer_extract_rate": 0.96484375,
      "calib/auroc": 0.5029411764705882,
      "calib/avg_num_step_conf": 1.125,
      "calib/ece": 0.3203643724696356,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.07692307692307693,
      "calib/gap": 0.005825821237585918,
      "calib/mean_conf": 0.09663967611336033,
      "calib/mu_c": 0.10064935064935063,
      "calib/mu_w": 0.09482352941176471,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.052631578947368425,
      "calib/std_conf": 0.26286271205587297,
      "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": 498.0,
      "completions/max_terminated_length": 498.0,
      "completions/mean_length": 234.28515625,
      "completions/mean_terminated_length": 234.28515625,
      "completions/min_length": 75.0,
      "completions/min_terminated_length": 75.0,
      "epoch": 0.14186666666666667,
      "grad_norm": 0.015624146908521652,
      "learning_rate": 1.8611111111111113e-06,
      "loss": 0.0568,
      "num_tokens": 25425071.0,
      "reward": 1.1076685190200806,
      "reward_std": 0.26900923252105713,
      "rewards/accuracy_reward_step": 0.30078125,
      "rewards/final_brier_reward_step": 0.6489308476448059,
      "rewards/format_reward_step": 0.96484375,
      "step": 133
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33154400484636426,
      "aux_distill/mean_u": 0.13351008039308454,
      "aux_distill/n_active_tok": 75.28125,
      "calib/answer_extract_rate": 0.93359375,
      "calib/auroc": 0.5581487914314662,
      "calib/avg_num_step_conf": 1.16796875,
      "calib/ece": 0.373305439330544,
      "calib/final_conf_rate": 0.93359375,
      "calib/format_rate": 0.9296875,
      "calib/frac_conf_gt_0.9": 0.07112970711297072,
      "calib/gap": 0.0583550441453177,
      "calib/mean_conf": 0.09506276150627616,
      "calib/mu_c": 0.12948979591836735,
      "calib/mu_w": 0.07113475177304965,
      "calib/nonempty_final_conf_rate": 0.93359375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.029163179916317988,
      "calib/std_conf": 0.25451713293291445,
      "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": 1154.0,
      "completions/max_terminated_length": 1154.0,
      "completions/mean_length": 244.79296875,
      "completions/mean_terminated_length": 245.75296020507812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 73.0,
      "epoch": 0.14293333333333333,
      "grad_norm": 0.015033029951155186,
      "learning_rate": 1.8333333333333333e-06,
      "loss": 0.1103,
      "num_tokens": 25596690.0,
      "reward": 1.138121247291565,
      "reward_std": 0.25055497884750366,
      "rewards/accuracy_reward_step": 0.3828125,
      "rewards/final_brier_reward_step": 0.5809300541877747,
      "rewards/format_reward_step": 0.9296875,
      "step": 134
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32621747767552733,
      "aux_distill/mean_u": 0.09390474473256447,
      "aux_distill/n_active_tok": 70.21875,
      "calib/answer_extract_rate": 0.94140625,
      "calib/auroc": 0.5159611380985427,
      "calib/avg_num_step_conf": 1.08984375,
      "calib/ece": 0.4307053941908714,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.07883817427385892,
      "calib/gap": 0.03173976405274115,
      "calib/mean_conf": 0.09883817427385892,
      "calib/mu_c": 0.11609090909090909,
      "calib/mu_w": 0.08435114503816794,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03655601659751036,
      "calib/std_conf": 0.2665257859182153,
      "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": 1175.0,
      "completions/max_terminated_length": 1175.0,
      "completions/mean_length": 223.390625,
      "completions/mean_terminated_length": 223.390625,
      "completions/min_length": 97.0,
      "completions/min_terminated_length": 97.0,
      "epoch": 0.144,
      "grad_norm": 0.017453555017709732,
      "learning_rate": 1.8055555555555557e-06,
      "loss": 0.0854,
      "num_tokens": 25759758.0,
      "reward": 1.1680976152420044,
      "reward_std": 0.2895016372203827,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.5354140400886536,
      "rewards/format_reward_step": 0.94140625,
      "step": 135
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2697015115991235,
      "aux_distill/mean_u": 0.08167140688285077,
      "aux_distill/n_active_tok": 76.6875,
      "calib/answer_extract_rate": 0.94140625,
      "calib/auroc": 0.50177304964539,
      "calib/avg_num_step_conf": 1.19140625,
      "calib/ece": 0.3858506224066391,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.9375,
      "calib/frac_conf_gt_0.9": 0.058091286307053944,
      "calib/gap": 0.03815035460992909,
      "calib/mean_conf": 0.07887966804979252,
      "calib/mu_c": 0.10120000000000001,
      "calib/mu_w": 0.06304964539007092,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.024896265560165977,
      "calib/std_conf": 0.23191783412618772,
      "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": 2800.0,
      "completions/max_terminated_length": 2800.0,
      "completions/mean_length": 220.23828125,
      "completions/mean_terminated_length": 220.23828125,
      "completions/min_length": 61.0,
      "completions/min_terminated_length": 61.0,
      "epoch": 0.14506666666666668,
      "grad_norm": 0.015763340517878532,
      "learning_rate": 1.777777777777778e-06,
      "loss": 0.056,
      "num_tokens": 25924627.0,
      "reward": 1.1440978050231934,
      "reward_std": 0.2571519613265991,
      "rewards/accuracy_reward_step": 0.390625,
      "rewards/final_brier_reward_step": 0.5694457292556763,
      "rewards/format_reward_step": 0.9375,
      "step": 136
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3393157171085477,
      "aux_distill/mean_u": 0.13486390960074313,
      "aux_distill/n_active_tok": 79.1875,
      "calib/answer_extract_rate": 0.96484375,
      "calib/auroc": 0.45054421768707487,
      "calib/avg_num_step_conf": 1.234375,
      "calib/ece": 0.39708502024291503,
      "calib/final_conf_rate": 0.96484375,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.08097165991902834,
      "calib/gap": 0.01586666666666668,
      "calib/mean_conf": 0.0997570850202429,
      "calib/mu_c": 0.1092,
      "calib/mu_w": 0.09333333333333332,
      "calib/nonempty_final_conf_rate": 0.96484375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0459919028340081,
      "calib/std_conf": 0.2703363134483246,
      "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": 2702.0,
      "completions/max_terminated_length": 2702.0,
      "completions/mean_length": 226.609375,
      "completions/mean_terminated_length": 226.609375,
      "completions/min_length": 83.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.14613333333333334,
      "grad_norm": 0.014484074898064137,
      "learning_rate": 1.75e-06,
      "loss": 0.1431,
      "num_tokens": 26089623.0,
      "reward": 1.1627554893493652,
      "reward_std": 0.2251381278038025,
      "rewards/accuracy_reward_step": 0.390625,
      "rewards/final_brier_reward_step": 0.5794172286987305,
      "rewards/format_reward_step": 0.96484375,
      "step": 137
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.30941454344429076,
      "aux_distill/mean_u": 0.09865573523816851,
      "aux_distill/n_active_tok": 76.5625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5532268170426065,
      "calib/avg_num_step_conf": 1.1953125,
      "calib/ece": 0.47723320158102756,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.09090909090909091,
      "calib/gap": 0.0443496240601504,
      "calib/mean_conf": 0.11881422924901183,
      "calib/mu_c": 0.1398496240601504,
      "calib/mu_w": 0.09549999999999999,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.03517786561264821,
      "calib/std_conf": 0.2908638100607603,
      "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": 1248.0,
      "completions/max_terminated_length": 1248.0,
      "completions/mean_length": 194.859375,
      "completions/mean_terminated_length": 194.859375,
      "completions/min_length": 65.0,
      "completions/min_terminated_length": 65.0,
      "epoch": 0.1472,
      "grad_norm": 0.018135685473680496,
      "learning_rate": 1.7222222222222224e-06,
      "loss": 0.0513,
      "num_tokens": 26243843.0,
      "reward": 1.27192223072052,
      "reward_std": 0.2641417384147644,
      "rewards/accuracy_reward_step": 0.51953125,
      "rewards/final_brier_reward_step": 0.5165007710456848,
      "rewards/format_reward_step": 0.98828125,
      "step": 138
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3026889590546489,
      "aux_distill/mean_u": 0.12891794165425197,
      "aux_distill/n_active_tok": 73.5625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5180472440944881,
      "calib/avg_num_step_conf": 1.1484375,
      "calib/ece": 0.4740873015873016,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.06746031746031746,
      "calib/gap": 0.006313700787401566,
      "calib/mean_conf": 0.09273809523809524,
      "calib/mu_c": 0.09591999999999999,
      "calib/mu_w": 0.08960629921259843,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0353968253968254,
      "calib/std_conf": 0.25348844495036854,
      "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": 378.0,
      "completions/max_terminated_length": 378.0,
      "completions/mean_length": 187.58984375,
      "completions/mean_terminated_length": 188.32550048828125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 75.0,
      "epoch": 0.14826666666666666,
      "grad_norm": 0.017611773684620857,
      "learning_rate": 1.6944444444444446e-06,
      "loss": 0.0482,
      "num_tokens": 26394962.0,
      "reward": 1.239492416381836,
      "reward_std": 0.2608584761619568,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/final_brier_reward_step": 0.5180472731590271,
      "rewards/format_reward_step": 0.984375,
      "step": 139
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3559173212852329,
      "aux_distill/mean_u": 0.12057362805447201,
      "aux_distill/n_active_tok": 81.96875,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5053320561941252,
      "calib/avg_num_step_conf": 1.27734375,
      "calib/ece": 0.4995219123505977,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.05179282868525897,
      "calib/gap": 0.031229885057471263,
      "calib/mean_conf": 0.0749003984063745,
      "calib/mu_c": 0.08933333333333333,
      "calib/mu_w": 0.05810344827586207,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.018286852589641432,
      "calib/std_conf": 0.2229347224591291,
      "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": 1865.0,
      "completions/max_terminated_length": 1865.0,
      "completions/mean_length": 210.265625,
      "completions/mean_terminated_length": 210.265625,
      "completions/min_length": 66.0,
      "completions/min_terminated_length": 66.0,
      "epoch": 0.14933333333333335,
      "grad_norm": 0.017557280138134956,
      "learning_rate": 1.6666666666666667e-06,
      "loss": 0.0716,
      "num_tokens": 26553806.0,
      "reward": 1.2641351222991943,
      "reward_std": 0.21921595931053162,
      "rewards/accuracy_reward_step": 0.52734375,
      "rewards/final_brier_reward_step": 0.49311408400535583,
      "rewards/format_reward_step": 0.98046875,
      "step": 140
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.325836184900254,
      "aux_distill/mean_u": 0.1298084705460597,
      "aux_distill/n_active_tok": 89.15625,
      "calib/answer_extract_rate": 0.94921875,
      "calib/auroc": 0.5169692453597075,
      "calib/avg_num_step_conf": 1.38671875,
      "calib/ece": 0.5068312757201645,
      "calib/final_conf_rate": 0.94921875,
      "calib/format_rate": 0.94921875,
      "calib/frac_conf_gt_0.9": 0.06995884773662552,
      "calib/gap": -0.044848936458474445,
      "calib/mean_conf": 0.09094650205761316,
      "calib/mu_c": 0.06842975206611569,
      "calib/mu_w": 0.11327868852459014,
      "calib/nonempty_final_conf_rate": 0.94921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.04991769547325103,
      "calib/std_conf": 0.25741739099632116,
      "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": 721.0,
      "completions/max_terminated_length": 721.0,
      "completions/mean_length": 215.8984375,
      "completions/mean_terminated_length": 215.8984375,
      "completions/min_length": 81.0,
      "completions/min_terminated_length": 81.0,
      "epoch": 0.1504,
      "grad_norm": 0.016918949782848358,
      "learning_rate": 1.638888888888889e-06,
      "loss": 0.0954,
      "num_tokens": 26716172.0,
      "reward": 1.1825156211853027,
      "reward_std": 0.29578882455825806,
      "rewards/accuracy_reward_step": 0.47265625,
      "rewards/final_brier_reward_step": 0.47049999237060547,
      "rewards/format_reward_step": 0.94921875,
      "step": 141
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38834019796922803,
      "aux_distill/mean_u": 0.17126695928741237,
      "aux_distill/n_active_tok": 83.59375,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.4985401459854015,
      "calib/avg_num_step_conf": 1.3046875,
      "calib/ece": 0.4508730158730158,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.051587301587301584,
      "calib/gap": -0.022569343065693442,
      "calib/mean_conf": 0.07626984126984127,
      "calib/mu_c": 0.064,
      "calib/mu_w": 0.08656934306569344,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.03539682539682539,
      "calib/std_conf": 0.22101940634102674,
      "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": 709.0,
      "completions/max_terminated_length": 709.0,
      "completions/mean_length": 201.75,
      "completions/mean_terminated_length": 201.75,
      "completions/min_length": 72.0,
      "completions/min_terminated_length": 72.0,
      "epoch": 0.15146666666666667,
      "grad_norm": 0.019056685268878937,
      "learning_rate": 1.6111111111111113e-06,
      "loss": 0.0345,
      "num_tokens": 26872980.0,
      "reward": 1.2108287811279297,
      "reward_std": 0.21176810562610626,
      "rewards/accuracy_reward_step": 0.453125,
      "rewards/final_brier_reward_step": 0.5349390506744385,
      "rewards/format_reward_step": 0.98046875,
      "step": 142
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3469718978740275,
      "aux_distill/mean_u": 0.11716029338227302,
      "aux_distill/n_active_tok": 98.71875,
      "calib/answer_extract_rate": 0.9296875,
      "calib/auroc": 0.42581805106076953,
      "calib/avg_num_step_conf": 1.53515625,
      "calib/ece": 0.461218487394958,
      "calib/final_conf_rate": 0.9296875,
      "calib/format_rate": 0.9296875,
      "calib/frac_conf_gt_0.9": 0.07563025210084033,
      "calib/gap": -0.07309385113268607,
      "calib/mean_conf": 0.09903361344537816,
      "calib/mu_c": 0.05757281553398059,
      "calib/mu_w": 0.13066666666666665,
      "calib/nonempty_final_conf_rate": 0.9296875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.06373949579831932,
      "calib/std_conf": 0.26121787914065403,
      "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": 2387.0,
      "completions/max_terminated_length": 2387.0,
      "completions/mean_length": 243.36328125,
      "completions/mean_terminated_length": 243.36328125,
      "completions/min_length": 84.0,
      "completions/min_terminated_length": 84.0,
      "epoch": 0.15253333333333333,
      "grad_norm": 0.015829890966415405,
      "learning_rate": 1.5833333333333333e-06,
      "loss": 0.1274,
      "num_tokens": 27042617.0,
      "reward": 1.1177458763122559,
      "reward_std": 0.28702062368392944,
      "rewards/accuracy_reward_step": 0.40234375,
      "rewards/final_brier_reward_step": 0.5011168122291565,
      "rewards/format_reward_step": 0.9296875,
      "step": 143
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3681814162991941,
      "aux_distill/mean_u": 0.12000292491714179,
      "aux_distill/n_active_tok": 91.28125,
      "calib/answer_extract_rate": 0.95703125,
      "calib/auroc": 0.5029694381422661,
      "calib/avg_num_step_conf": 1.421875,
      "calib/ece": 0.4917959183673469,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.07346938775510205,
      "calib/gap": 0.00044174562925396177,
      "calib/mean_conf": 0.09955102040816327,
      "calib/mu_c": 0.09976377952755905,
      "calib/mu_w": 0.09932203389830509,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.03648979591836735,
      "calib/std_conf": 0.26153901345723407,
      "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": 2921.0,
      "completions/max_terminated_length": 2921.0,
      "completions/mean_length": 212.92578125,
      "completions/mean_terminated_length": 213.76080322265625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 80.0,
      "epoch": 0.1536,
      "grad_norm": 0.01637650653719902,
      "learning_rate": 1.5555555555555558e-06,
      "loss": 0.1096,
      "num_tokens": 27201254.0,
      "reward": 1.2170963287353516,
      "reward_std": 0.25823676586151123,
      "rewards/accuracy_reward_step": 0.49609375,
      "rewards/final_brier_reward_step": 0.48497384786605835,
      "rewards/format_reward_step": 0.95703125,
      "step": 144
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2961839628405869,
      "aux_distill/mean_u": 0.11787739525770995,
      "aux_distill/n_active_tok": 81.8125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.461596484323757,
      "calib/avg_num_step_conf": 1.27734375,
      "calib/ece": 0.5660865612648222,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.043478260869565216,
      "calib/gap": 0.014169264069264068,
      "calib/mean_conf": 0.0700399209486166,
      "calib/mu_c": 0.07558441558441559,
      "calib/mu_w": 0.06141515151515152,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.01371541501976284,
      "calib/std_conf": 0.2013873581212608,
      "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": 616.0,
      "completions/max_terminated_length": 616.0,
      "completions/mean_length": 187.5390625,
      "completions/mean_terminated_length": 188.27452087402344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 78.0,
      "epoch": 0.15466666666666667,
      "grad_norm": 0.01805811934173107,
      "learning_rate": 1.527777777777778e-06,
      "loss": 0.0142,
      "num_tokens": 27351968.0,
      "reward": 1.3120663166046143,
      "reward_std": 0.20374266803264618,
      "rewards/accuracy_reward_step": 0.6015625,
      "rewards/final_brier_reward_step": 0.43272656202316284,
      "rewards/format_reward_step": 0.98828125,
      "step": 145
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36236625676974654,
      "aux_distill/mean_u": 0.09993783496658051,
      "aux_distill/n_active_tok": 94.34375,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5411502576242864,
      "calib/avg_num_step_conf": 1.4765625,
      "calib/ece": 0.30735138339920953,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.05928853754940711,
      "calib/gap": 0.06291440607157776,
      "calib/mean_conf": 0.08917035573122531,
      "calib/mu_c": 0.1306988372093023,
      "calib/mu_w": 0.06778443113772455,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.028300395256916994,
      "calib/std_conf": 0.23328645191461295,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 540.0,
      "completions/max_terminated_length": 540.0,
      "completions/mean_length": 202.4453125,
      "completions/mean_terminated_length": 203.23922729492188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 76.0,
      "epoch": 0.15573333333333333,
      "grad_norm": 0.01728767901659012,
      "learning_rate": 1.5e-06,
      "loss": 0.0343,
      "num_tokens": 27511010.0,
      "reward": 1.169335126876831,
      "reward_std": 0.22755557298660278,
      "rewards/accuracy_reward_step": 0.3359375,
      "rewards/final_brier_reward_step": 0.6785140633583069,
      "rewards/format_reward_step": 0.98828125,
      "step": 146
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35689425515010953,
      "aux_distill/mean_u": 0.13463799166264762,
      "aux_distill/n_active_tok": 102.34375,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.48808092948717946,
      "calib/avg_num_step_conf": 1.59765625,
      "calib/ece": 0.3636904761904761,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.051587301587301584,
      "calib/gap": 0.01811698717948719,
      "calib/mean_conf": 0.08805555555555554,
      "calib/mu_c": 0.09927083333333335,
      "calib/mu_w": 0.08115384615384616,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0353968253968254,
      "calib/std_conf": 0.23207609608466293,
      "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": 2326.0,
      "completions/max_terminated_length": 2326.0,
      "completions/mean_length": 221.4296875,
      "completions/mean_terminated_length": 221.4296875,
      "completions/min_length": 70.0,
      "completions/min_terminated_length": 70.0,
      "epoch": 0.1568,
      "grad_norm": 0.016975658014416695,
      "learning_rate": 1.4722222222222225e-06,
      "loss": 0.0652,
      "num_tokens": 27671376.0,
      "reward": 1.1728016138076782,
      "reward_std": 0.2246793806552887,
      "rewards/accuracy_reward_step": 0.375,
      "rewards/final_brier_reward_step": 0.6190406084060669,
      "rewards/format_reward_step": 0.9765625,
      "step": 147
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38051269575953484,
      "aux_distill/mean_u": 0.16149938439742753,
      "aux_distill/n_active_tok": 92.96875,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5304027632950991,
      "calib/avg_num_step_conf": 1.44921875,
      "calib/ece": 0.5001204819277109,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.03614457831325301,
      "calib/gap": 0.037861704900938487,
      "calib/mean_conf": 0.06734939759036146,
      "calib/mu_c": 0.08437956204379564,
      "calib/mu_w": 0.04651785714285715,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.008634538152610442,
      "calib/std_conf": 0.19164109068829516,
      "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": 927.0,
      "completions/max_terminated_length": 927.0,
      "completions/mean_length": 201.98046875,
      "completions/mean_terminated_length": 201.98046875,
      "completions/min_length": 57.0,
      "completions/min_terminated_length": 57.0,
      "epoch": 0.15786666666666666,
      "grad_norm": 0.018086962401866913,
      "learning_rate": 1.4444444444444445e-06,
      "loss": 0.0778,
      "num_tokens": 27828195.0,
      "reward": 1.2653236389160156,
      "reward_std": 0.25430917739868164,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/final_brier_reward_step": 0.48767852783203125,
      "rewards/format_reward_step": 0.97265625,
      "step": 148
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35395267652347684,
      "aux_distill/mean_u": 0.13261109607192345,
      "aux_distill/n_active_tok": 134.15625,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5235483007760235,
      "calib/avg_num_step_conf": 2.09375,
      "calib/ece": 0.38477911646586344,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.04417670682730924,
      "calib/gap": 0.00790272946213541,
      "calib/mean_conf": 0.07827309236947792,
      "calib/mu_c": 0.08297029702970297,
      "calib/mu_w": 0.07506756756756756,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.028714859437751,
      "calib/std_conf": 0.2090536856715205,
      "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": 2419.0,
      "completions/max_terminated_length": 2419.0,
      "completions/mean_length": 253.02734375,
      "completions/mean_terminated_length": 253.02734375,
      "completions/min_length": 61.0,
      "completions/min_terminated_length": 61.0,
      "epoch": 0.15893333333333334,
      "grad_norm": 0.013964850455522537,
      "learning_rate": 1.4166666666666667e-06,
      "loss": 0.079,
      "num_tokens": 27997426.0,
      "reward": 1.1745210886001587,
      "reward_std": 0.2611221969127655,
      "rewards/accuracy_reward_step": 0.39453125,
      "rewards/final_brier_reward_step": 0.5912296772003174,
      "rewards/format_reward_step": 0.96875,
      "step": 149
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36969612818211317,
      "aux_distill/mean_u": 0.1544741904551301,
      "aux_distill/n_active_tok": 101.53125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5018785415126878,
      "calib/avg_num_step_conf": 1.58984375,
      "calib/ece": 0.4910588235294118,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.03137254901960784,
      "calib/gap": -0.0012176644493717692,
      "calib/mean_conf": 0.05945098039215686,
      "calib/mu_c": 0.05886363636363637,
      "calib/mu_w": 0.06008130081300814,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.01643137254901961,
      "calib/std_conf": 0.17386182407759287,
      "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": 644.0,
      "completions/max_terminated_length": 644.0,
      "completions/mean_length": 197.3671875,
      "completions/mean_terminated_length": 198.1411895751953,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 77.0,
      "epoch": 0.16,
      "grad_norm": 0.015584699809551239,
      "learning_rate": 1.3888888888888892e-06,
      "loss": 0.0096,
      "num_tokens": 28152912.0,
      "reward": 1.2674424648284912,
      "reward_std": 0.15812954306602478,
      "rewards/accuracy_reward_step": 0.515625,
      "rewards/final_brier_reward_step": 0.5075414180755615,
      "rewards/format_reward_step": 0.99609375,
      "step": 150
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36593104852363467,
      "aux_distill/mean_u": 0.15258560407885197,
      "aux_distill/n_active_tok": 96.84375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5031104730653542,
      "calib/avg_num_step_conf": 1.51171875,
      "calib/ece": 0.3605577689243028,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.02390438247011952,
      "calib/gap": -0.03334974022422751,
      "calib/mean_conf": 0.05243027888446216,
      "calib/mu_c": 0.03130434782608695,
      "calib/mu_w": 0.06465408805031446,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.023227091633466136,
      "calib/std_conf": 0.15073283837724566,
      "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": 1860.0,
      "completions/max_terminated_length": 1860.0,
      "completions/mean_length": 225.625,
      "completions/mean_terminated_length": 226.5098114013672,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 75.0,
      "epoch": 0.16106666666666666,
      "grad_norm": 0.015279371291399002,
      "learning_rate": 1.3611111111111112e-06,
      "loss": 0.0523,
      "num_tokens": 28317696.0,
      "reward": 1.1589202880859375,
      "reward_std": 0.21208204329013824,
      "rewards/accuracy_reward_step": 0.359375,
      "rewards/final_brier_reward_step": 0.618621826171875,
      "rewards/format_reward_step": 0.98046875,
      "step": 151
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39372802060097456,
      "aux_distill/mean_u": 0.1188828538735574,
      "aux_distill/n_active_tok": 93.15625,
      "calib/answer_extract_rate": 0.9609375,
      "calib/auroc": 0.5239783856805134,
      "calib/avg_num_step_conf": 1.453125,
      "calib/ece": 0.4027235772357724,
      "calib/final_conf_rate": 0.9609375,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.032520325203252036,
      "calib/gap": -0.005854103343465061,
      "calib/mean_conf": 0.07126016260162601,
      "calib/mu_c": 0.0679047619047619,
      "calib/mu_w": 0.07375886524822696,
      "calib/nonempty_final_conf_rate": 0.9609375,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.02357723577235773,
      "calib/std_conf": 0.18500834247937423,
      "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": 3031.0,
      "completions/max_terminated_length": 3031.0,
      "completions/mean_length": 228.0078125,
      "completions/mean_terminated_length": 228.0078125,
      "completions/min_length": 63.0,
      "completions/min_terminated_length": 63.0,
      "epoch": 0.16213333333333332,
      "grad_norm": 0.015285255387425423,
      "learning_rate": 1.3333333333333334e-06,
      "loss": 0.1525,
      "num_tokens": 28481458.0,
      "reward": 1.174981713294983,
      "reward_std": 0.23847061395645142,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.5687136650085449,
      "rewards/format_reward_step": 0.9609375,
      "step": 152
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4158263560384512,
      "aux_distill/mean_u": 0.185887581378134,
      "aux_distill/n_active_tok": 101.03125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5082277694982613,
      "calib/avg_num_step_conf": 1.578125,
      "calib/ece": 0.4423224409448819,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.01968503937007874,
      "calib/gap": 0.023963474913065073,
      "calib/mean_conf": 0.052874409448818896,
      "calib/mu_c": 0.06532786885245902,
      "calib/mu_w": 0.041364393939393944,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.007440944881889762,
      "calib/std_conf": 0.14663943411123173,
      "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": 743.0,
      "completions/max_terminated_length": 743.0,
      "completions/mean_length": 218.94140625,
      "completions/mean_terminated_length": 219.80001831054688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 82.0,
      "epoch": 0.1632,
      "grad_norm": 0.01481254305690527,
      "learning_rate": 1.3055555555555556e-06,
      "loss": 0.0337,
      "num_tokens": 28644827.0,
      "reward": 1.249547004699707,
      "reward_std": 0.18368719518184662,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.5537816286087036,
      "rewards/format_reward_step": 0.9921875,
      "step": 153
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.350020554382354,
      "aux_distill/mean_u": 0.1530069324122927,
      "aux_distill/n_active_tok": 91.84375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5121492346938775,
      "calib/avg_num_step_conf": 1.43359375,
      "calib/ece": 0.43182539682539683,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.047619047619047616,
      "calib/gap": -0.007178571428571423,
      "calib/mean_conf": 0.07523809523809524,
      "calib/mu_c": 0.07125000000000001,
      "calib/mu_w": 0.07842857142857143,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03130952380952381,
      "calib/std_conf": 0.21098677565462198,
      "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": 725.0,
      "completions/max_terminated_length": 725.0,
      "completions/mean_length": 201.703125,
      "completions/mean_terminated_length": 202.49412536621094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 71.0,
      "epoch": 0.16426666666666667,
      "grad_norm": 0.016052914783358574,
      "learning_rate": 1.2777777777777779e-06,
      "loss": 0.0341,
      "num_tokens": 28800903.0,
      "reward": 1.209600806236267,
      "reward_std": 0.21740567684173584,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/final_brier_reward_step": 0.5598265528678894,
      "rewards/format_reward_step": 0.984375,
      "step": 154
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3580281659960747,
      "aux_distill/mean_u": 0.14145642484386875,
      "aux_distill/n_active_tok": 103.09375,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.4906623026569118,
      "calib/avg_num_step_conf": 1.609375,
      "calib/ece": 0.40308300395256913,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.03557312252964427,
      "calib/gap": -0.008018867924528314,
      "calib/mean_conf": 0.0666403162055336,
      "calib/mu_c": 0.06198113207547169,
      "calib/mu_w": 0.07,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.025375494071146247,
      "calib/std_conf": 0.18552611088804694,
      "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": 619.0,
      "completions/max_terminated_length": 619.0,
      "completions/mean_length": 194.28515625,
      "completions/mean_terminated_length": 195.0470733642578,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 70.0,
      "epoch": 0.16533333333333333,
      "grad_norm": 0.01730683632194996,
      "learning_rate": 1.25e-06,
      "loss": 0.0453,
      "num_tokens": 28957856.0,
      "reward": 1.2017738819122314,
      "reward_std": 0.21324856579303741,
      "rewards/accuracy_reward_step": 0.4140625,
      "rewards/final_brier_reward_step": 0.5871413946151733,
      "rewards/format_reward_step": 0.98828125,
      "step": 155
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36394436936825514,
      "aux_distill/mean_u": 0.15244325122522395,
      "aux_distill/n_active_tok": 121.28125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5520793222949558,
      "calib/avg_num_step_conf": 1.89453125,
      "calib/ece": 0.37584980237154153,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.02766798418972332,
      "calib/gap": 0.0351219355666795,
      "calib/mean_conf": 0.05893280632411067,
      "calib/mu_c": 0.07933962264150944,
      "calib/mu_w": 0.04421768707482994,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.007905138339920948,
      "calib/std_conf": 0.1648236793891626,
      "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": 2092.0,
      "completions/max_terminated_length": 2092.0,
      "completions/mean_length": 241.1171875,
      "completions/mean_terminated_length": 242.06275939941406,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 71.0,
      "epoch": 0.1664,
      "grad_norm": 0.011360744014382362,
      "learning_rate": 1.2222222222222223e-06,
      "loss": 0.0355,
      "num_tokens": 29124342.0,
      "reward": 1.2130236625671387,
      "reward_std": 0.18996354937553406,
      "rewards/accuracy_reward_step": 0.4140625,
      "rewards/final_brier_reward_step": 0.6096410155296326,
      "rewards/format_reward_step": 0.98828125,
      "step": 156
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.401463286485523,
      "aux_distill/mean_u": 0.21597904442972382,
      "aux_distill/n_active_tok": 127.75,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.4743506493506494,
      "calib/avg_num_step_conf": 2.0078125,
      "calib/ece": 0.42352000000000006,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.024,
      "calib/gap": -0.004194805194805196,
      "calib/mean_conf": 0.057440000000000005,
      "calib/mu_c": 0.05509090909090908,
      "calib/mu_w": 0.059285714285714275,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.02048,
      "calib/std_conf": 0.17260778197984006,
      "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": 0.015625,
      "completions/max_length": 1851.0,
      "completions/max_terminated_length": 1851.0,
      "completions/mean_length": 228.9609375,
      "completions/mean_terminated_length": 232.59524536132812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 57.0,
      "epoch": 0.16746666666666668,
      "grad_norm": 0.012861998751759529,
      "learning_rate": 1.1944444444444446e-06,
      "loss": 0.0008,
      "num_tokens": 29286684.0,
      "reward": 1.2028257846832275,
      "reward_std": 0.20734016597270966,
      "rewards/accuracy_reward_step": 0.43359375,
      "rewards/final_brier_reward_step": 0.5619015693664551,
      "rewards/format_reward_step": 0.9765625,
      "step": 157
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3736504097469151,
      "aux_distill/mean_u": 0.16085979852261284,
      "aux_distill/n_active_tok": 116.84375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.60516899878602,
      "calib/avg_num_step_conf": 1.82421875,
      "calib/ece": 0.5218476190476191,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.047619047619047616,
      "calib/gap": 0.028447920260686217,
      "calib/mean_conf": 0.07735873015873017,
      "calib/mu_c": 0.08988936170212765,
      "calib/mu_w": 0.06144144144144143,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.019841269841269837,
      "calib/std_conf": 0.21477512539229582,
      "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": 2533.0,
      "completions/max_terminated_length": 2533.0,
      "completions/mean_length": 238.93359375,
      "completions/mean_terminated_length": 238.93359375,
      "completions/min_length": 70.0,
      "completions/min_terminated_length": 70.0,
      "epoch": 0.16853333333333334,
      "grad_norm": 0.013917519710958004,
      "learning_rate": 1.1666666666666668e-06,
      "loss": 0.0936,
      "num_tokens": 29453091.0,
      "reward": 1.283625841140747,
      "reward_std": 0.2249603569507599,
      "rewards/accuracy_reward_step": 0.55078125,
      "rewards/final_brier_reward_step": 0.4813140332698822,
      "rewards/format_reward_step": 0.984375,
      "step": 158
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.387286314740777,
      "aux_distill/mean_u": 0.15654495993202383,
      "aux_distill/n_active_tok": 123.28125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5469138102334823,
      "calib/avg_num_step_conf": 1.92578125,
      "calib/ece": 0.44444881889763777,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.027559055118110236,
      "calib/gap": 0.021331346249379032,
      "calib/mean_conf": 0.05948818897637795,
      "calib/mu_c": 0.07057377049180327,
      "calib/mu_w": 0.04924242424242424,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.011811023622047246,
      "calib/std_conf": 0.17298256095899606,
      "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": 1930.0,
      "completions/max_terminated_length": 1930.0,
      "completions/mean_length": 224.86328125,
      "completions/mean_terminated_length": 224.86328125,
      "completions/min_length": 82.0,
      "completions/min_terminated_length": 82.0,
      "epoch": 0.1696,
      "grad_norm": 0.013575537130236626,
      "learning_rate": 1.138888888888889e-06,
      "loss": 0.0541,
      "num_tokens": 29615440.0,
      "reward": 1.2475013732910156,
      "reward_std": 0.20087464153766632,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.5496902465820312,
      "rewards/format_reward_step": 0.9921875,
      "step": 159
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40404290379956365,
      "aux_distill/mean_u": 0.16472119918915018,
      "aux_distill/n_active_tok": 120.34375,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5437062937062938,
      "calib/avg_num_step_conf": 1.87890625,
      "calib/ece": 0.4110441767068273,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.05220883534136546,
      "calib/gap": 0.004880591107006227,
      "calib/mean_conf": 0.07795180722891565,
      "calib/mu_c": 0.08075471698113208,
      "calib/mu_w": 0.07587412587412586,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.03164658634538152,
      "calib/std_conf": 0.22251623668645712,
      "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": 2743.0,
      "completions/max_terminated_length": 2743.0,
      "completions/mean_length": 247.48828125,
      "completions/mean_terminated_length": 249.43701171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 82.0,
      "epoch": 0.17066666666666666,
      "grad_norm": 0.012900142930448055,
      "learning_rate": 1.111111111111111e-06,
      "loss": 0.0777,
      "num_tokens": 29783637.0,
      "reward": 1.1860899925231934,
      "reward_std": 0.21723809838294983,
      "rewards/accuracy_reward_step": 0.4140625,
      "rewards/final_brier_reward_step": 0.5713988542556763,
      "rewards/format_reward_step": 0.97265625,
      "step": 160
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3169996486976743,
      "aux_distill/mean_u": 0.15129769459006281,
      "aux_distill/n_active_tok": 127.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4888630319148936,
      "calib/avg_num_step_conf": 1.984375,
      "calib/ece": 0.5971653543307086,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.047244094488188976,
      "calib/gap": -0.012497340425531897,
      "calib/mean_conf": 0.08,
      "calib/mu_c": 0.075375,
      "calib/mu_w": 0.0878723404255319,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.023622047244094488,
      "calib/std_conf": 0.22045407685048604,
      "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": 2533.0,
      "completions/max_terminated_length": 2533.0,
      "completions/mean_length": 216.55859375,
      "completions/mean_terminated_length": 216.55859375,
      "completions/min_length": 59.0,
      "completions/min_terminated_length": 59.0,
      "epoch": 0.17173333333333332,
      "grad_norm": 0.014349176548421383,
      "learning_rate": 1.0833333333333335e-06,
      "loss": 0.0883,
      "num_tokens": 29942996.0,
      "reward": 1.3284180164337158,
      "reward_std": 0.19347649812698364,
      "rewards/accuracy_reward_step": 0.62890625,
      "rewards/final_brier_reward_step": 0.4068359434604645,
      "rewards/format_reward_step": 0.9921875,
      "step": 161
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34928833367303014,
      "aux_distill/mean_u": 0.1428228857288194,
      "aux_distill/n_active_tok": 132.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4971402573768361,
      "calib/avg_num_step_conf": 2.0625,
      "calib/ece": 0.5887843137254902,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.050980392156862744,
      "calib/gap": -0.01196347328740413,
      "calib/mean_conf": 0.07396078431372549,
      "calib/mu_c": 0.06936305732484077,
      "calib/mu_w": 0.0813265306122449,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.023529411764705882,
      "calib/std_conf": 0.21804928034324803,
      "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": 2439.0,
      "completions/max_terminated_length": 2439.0,
      "completions/mean_length": 230.4765625,
      "completions/mean_terminated_length": 230.4765625,
      "completions/min_length": 79.0,
      "completions/min_terminated_length": 79.0,
      "epoch": 0.1728,
      "grad_norm": 0.01401078887283802,
      "learning_rate": 1.0555555555555557e-06,
      "loss": 0.0889,
      "num_tokens": 30106142.0,
      "reward": 1.3188691139221191,
      "reward_std": 0.2151828408241272,
      "rewards/accuracy_reward_step": 0.61328125,
      "rewards/final_brier_reward_step": 0.41508203744888306,
      "rewards/format_reward_step": 0.99609375,
      "step": 162
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40522405598312616,
      "aux_distill/mean_u": 0.1774422503844015,
      "aux_distill/n_active_tok": 133.1875,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.5390625,
      "calib/avg_num_step_conf": 2.078125,
      "calib/ece": 0.4032258064516129,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.028225806451612902,
      "calib/gap": -0.0003739316239316087,
      "calib/mean_conf": 0.056370967741935486,
      "calib/mu_c": 0.05615384615384617,
      "calib/mu_w": 0.05652777777777778,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.020120967741935482,
      "calib/std_conf": 0.17469020818164752,
      "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": 2263.0,
      "completions/max_terminated_length": 2263.0,
      "completions/mean_length": 265.3125,
      "completions/mean_terminated_length": 265.3125,
      "completions/min_length": 81.0,
      "completions/min_terminated_length": 81.0,
      "epoch": 0.17386666666666667,
      "grad_norm": 0.011856562457978725,
      "learning_rate": 1.0277777777777777e-06,
      "loss": 0.1035,
      "num_tokens": 30278894.0,
      "reward": 1.178366780281067,
      "reward_std": 0.25111645460128784,
      "rewards/accuracy_reward_step": 0.40625,
      "rewards/final_brier_reward_step": 0.5754835605621338,
      "rewards/format_reward_step": 0.96875,
      "step": 163
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38988359714858234,
      "aux_distill/mean_u": 0.16001530292552524,
      "aux_distill/n_active_tok": 142.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5338594064386318,
      "calib/avg_num_step_conf": 2.22265625,
      "calib/ece": 0.41811023622047244,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.027559055118110236,
      "calib/gap": 0.009088279678068412,
      "calib/mean_conf": 0.053937007874015744,
      "calib/mu_c": 0.05901785714285714,
      "calib/mu_w": 0.04992957746478873,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.015551181102362205,
      "calib/std_conf": 0.1687595199079184,
      "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": 2365.0,
      "completions/max_terminated_length": 2365.0,
      "completions/mean_length": 248.578125,
      "completions/mean_terminated_length": 248.578125,
      "completions/min_length": 73.0,
      "completions/min_terminated_length": 73.0,
      "epoch": 0.17493333333333333,
      "grad_norm": 0.012431781738996506,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0477,
      "num_tokens": 30448666.0,
      "reward": 1.2250921726226807,
      "reward_std": 0.2109067142009735,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.5751843452453613,
      "rewards/format_reward_step": 0.9921875,
      "step": 164
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.42134454753249884,
      "aux_distill/mean_u": 0.1822649157286872,
      "aux_distill/n_active_tok": 128.5625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5013222394220845,
      "calib/avg_num_step_conf": 2.01171875,
      "calib/ece": 0.3761023622047244,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.047244094488188976,
      "calib/gap": 0.01310887512899897,
      "calib/mean_conf": 0.07460629921259843,
      "calib/mu_c": 0.08245098039215687,
      "calib/mu_w": 0.0693421052631579,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.024566929133858266,
      "calib/std_conf": 0.21081124361086834,
      "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": 558.0,
      "completions/max_terminated_length": 558.0,
      "completions/mean_length": 222.37109375,
      "completions/mean_terminated_length": 223.24314880371094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 77.0,
      "epoch": 0.176,
      "grad_norm": 0.014216553419828415,
      "learning_rate": 9.722222222222224e-07,
      "loss": 0.0429,
      "num_tokens": 30611169.0,
      "reward": 1.1994494199752808,
      "reward_std": 0.18080687522888184,
      "rewards/accuracy_reward_step": 0.3984375,
      "rewards/final_brier_reward_step": 0.6098363399505615,
      "rewards/format_reward_step": 0.9921875,
      "step": 165
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3586527188308537,
      "aux_distill/mean_u": 0.18100545828037087,
      "aux_distill/n_active_tok": 132.875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.517906683480454,
      "calib/avg_num_step_conf": 2.07421875,
      "calib/ece": 0.5074603174603175,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.03968253968253968,
      "calib/gap": -0.027437578814627994,
      "calib/mean_conf": 0.0692063492063492,
      "calib/mu_c": 0.05592307692307692,
      "calib/mu_w": 0.08336065573770492,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.030396825396825394,
      "calib/std_conf": 0.2027474117604479,
      "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": 2017.0,
      "completions/max_terminated_length": 2017.0,
      "completions/mean_length": 249.8984375,
      "completions/mean_terminated_length": 250.87844848632812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 84.0,
      "epoch": 0.17706666666666668,
      "grad_norm": 0.01299862191081047,
      "learning_rate": 9.444444444444445e-07,
      "loss": 0.0718,
      "num_tokens": 30781327.0,
      "reward": 1.24409019947052,
      "reward_std": 0.21570822596549988,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.4881804585456848,
      "rewards/format_reward_step": 0.984375,
      "step": 166
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40623397240415215,
      "aux_distill/mean_u": 0.1911809320311116,
      "aux_distill/n_active_tok": 133.59375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5539028731248411,
      "calib/avg_num_step_conf": 2.0859375,
      "calib/ece": 0.5149603174603175,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.023809523809523808,
      "calib/gap": 0.016235697940503425,
      "calib/mean_conf": 0.04845238095238095,
      "calib/mu_c": 0.055797101449275355,
      "calib/mu_w": 0.03956140350877193,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.007896825396825399,
      "calib/std_conf": 0.15272779536959882,
      "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": 1564.0,
      "completions/max_terminated_length": 1564.0,
      "completions/mean_length": 241.88671875,
      "completions/mean_terminated_length": 242.83531188964844,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 66.0,
      "epoch": 0.17813333333333334,
      "grad_norm": 0.011331038549542427,
      "learning_rate": 9.166666666666666e-07,
      "loss": 0.0552,
      "num_tokens": 30948858.0,
      "reward": 1.271348237991333,
      "reward_std": 0.1793292909860611,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/final_brier_reward_step": 0.480196475982666,
      "rewards/format_reward_step": 0.984375,
      "step": 167
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3805929427035153,
      "aux_distill/mean_u": 0.1850864661732162,
      "aux_distill/n_active_tok": 154.375,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5907431551499348,
      "calib/avg_num_step_conf": 2.41015625,
      "calib/ece": 0.4818145161290322,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.024193548387096774,
      "calib/gap": 0.03916166883963494,
      "calib/mean_conf": 0.05044354838709678,
      "calib/mu_c": 0.06907692307692308,
      "calib/mu_w": 0.029915254237288132,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.004032258064516128,
      "calib/std_conf": 0.15678538867999778,
      "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": 2438.0,
      "completions/max_terminated_length": 2438.0,
      "completions/mean_length": 259.44921875,
      "completions/mean_terminated_length": 259.44921875,
      "completions/min_length": 87.0,
      "completions/min_terminated_length": 87.0,
      "epoch": 0.1792,
      "grad_norm": 0.014961393550038338,
      "learning_rate": 8.88888888888889e-07,
      "loss": 0.1209,
      "num_tokens": 31119949.0,
      "reward": 1.2485013008117676,
      "reward_std": 0.2524811327457428,
      "rewards/accuracy_reward_step": 0.51171875,
      "rewards/final_brier_reward_step": 0.5048152208328247,
      "rewards/format_reward_step": 0.96875,
      "step": 168
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34547490254044533,
      "aux_distill/mean_u": 0.10757020926135626,
      "aux_distill/n_active_tok": 120.625,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5413037909836066,
      "calib/avg_num_step_conf": 1.8828125,
      "calib/ece": 0.46304000000000006,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.024,
      "calib/gap": 0.012348872950819656,
      "calib/mean_conf": 0.05712,
      "calib/mu_c": 0.06344262295081966,
      "calib/mu_w": 0.05109375000000001,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.016080000000000004,
      "calib/std_conf": 0.1751036995611458,
      "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": 1253.0,
      "completions/max_terminated_length": 1253.0,
      "completions/mean_length": 230.125,
      "completions/mean_terminated_length": 230.125,
      "completions/min_length": 66.0,
      "completions/min_terminated_length": 66.0,
      "epoch": 0.18026666666666666,
      "grad_norm": 0.013571527786552906,
      "learning_rate": 8.611111111111112e-07,
      "loss": 0.0784,
      "num_tokens": 31283045.0,
      "reward": 1.2304667234420776,
      "reward_std": 0.20049837231636047,
      "rewards/accuracy_reward_step": 0.48046875,
      "rewards/final_brier_reward_step": 0.5273398160934448,
      "rewards/format_reward_step": 0.97265625,
      "step": 169
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37008119793608785,
      "aux_distill/mean_u": 0.14451895100963041,
      "aux_distill/n_active_tok": 137.09375,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.49952320406865863,
      "calib/avg_num_step_conf": 2.140625,
      "calib/ece": 0.459800796812749,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.04780876494023904,
      "calib/gap": 0.002264462809917364,
      "calib/mean_conf": 0.07709163346613547,
      "calib/mu_c": 0.07826446280991736,
      "calib/mu_w": 0.076,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.02741035856573705,
      "calib/std_conf": 0.21855899759106026,
      "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": 2515.0,
      "completions/max_terminated_length": 2515.0,
      "completions/mean_length": 244.1640625,
      "completions/mean_terminated_length": 244.1640625,
      "completions/min_length": 71.0,
      "completions/min_terminated_length": 71.0,
      "epoch": 0.18133333333333335,
      "grad_norm": 0.011340836994349957,
      "learning_rate": 8.333333333333333e-07,
      "loss": 0.0953,
      "num_tokens": 31449703.0,
      "reward": 1.2274580001831055,
      "reward_std": 0.21996477246284485,
      "rewards/accuracy_reward_step": 0.47265625,
      "rewards/final_brier_reward_step": 0.5291347503662109,
      "rewards/format_reward_step": 0.98046875,
      "step": 170
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3528564595617354,
      "aux_distill/mean_u": 0.1634984313581802,
      "aux_distill/n_active_tok": 142.0625,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5196801544082168,
      "calib/avg_num_step_conf": 2.21875,
      "calib/ece": 0.32158730158730164,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.03571428571428571,
      "calib/gap": 0.039684290342593215,
      "calib/mean_conf": 0.06309523809523811,
      "calib/mu_c": 0.0887640449438202,
      "calib/mu_w": 0.04907975460122699,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.01575396825396825,
      "calib/std_conf": 0.18882243395482923,
      "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": 2720.0,
      "completions/max_terminated_length": 2720.0,
      "completions/mean_length": 255.046875,
      "completions/mean_terminated_length": 255.046875,
      "completions/min_length": 66.0,
      "completions/min_terminated_length": 66.0,
      "epoch": 0.1824,
      "grad_norm": 0.012555527500808239,
      "learning_rate": 8.055555555555557e-07,
      "loss": 0.0855,
      "num_tokens": 31621891.0,
      "reward": 1.1656486988067627,
      "reward_std": 0.20087698101997375,
      "rewards/accuracy_reward_step": 0.34765625,
      "rewards/final_brier_reward_step": 0.6555160284042358,
      "rewards/format_reward_step": 0.98046875,
      "step": 171
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35162057261914015,
      "aux_distill/mean_u": 0.13648460286482703,
      "aux_distill/n_active_tok": 131.03125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4659004799191715,
      "calib/avg_num_step_conf": 2.0546875,
      "calib/ece": 0.5692549019607842,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0392156862745098,
      "calib/gap": -0.0409484718363223,
      "calib/mean_conf": 0.07380392156862745,
      "calib/mu_c": 0.05662162162162161,
      "calib/mu_w": 0.09757009345794392,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03133333333333333,
      "calib/std_conf": 0.20427595902504458,
      "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": 718.0,
      "completions/max_terminated_length": 718.0,
      "completions/mean_length": 224.22265625,
      "completions/mean_terminated_length": 225.1019744873047,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.18346666666666667,
      "grad_norm": 0.013728141784667969,
      "learning_rate": 7.777777777777779e-07,
      "loss": -0.0131,
      "num_tokens": 31782644.0,
      "reward": 1.2943949699401855,
      "reward_std": 0.20674556493759155,
      "rewards/accuracy_reward_step": 0.578125,
      "rewards/final_brier_reward_step": 0.43644610047340393,
      "rewards/format_reward_step": 0.99609375,
      "step": 172
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35267816623672843,
      "aux_distill/mean_u": 0.1463806371924626,
      "aux_distill/n_active_tok": 162.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.47975628930817615,
      "calib/avg_num_step_conf": 2.53515625,
      "calib/ece": 0.3547839215686274,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.043137254901960784,
      "calib/gap": 0.02630636792452832,
      "calib/mean_conf": 0.07078470588235294,
      "calib/mu_c": 0.08718750000000001,
      "calib/mu_w": 0.060881132075471696,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.024549019607843142,
      "calib/std_conf": 0.20494122534161185,
      "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": 2343.0,
      "completions/max_terminated_length": 2343.0,
      "completions/mean_length": 267.28515625,
      "completions/mean_terminated_length": 267.28515625,
      "completions/min_length": 65.0,
      "completions/min_terminated_length": 65.0,
      "epoch": 0.18453333333333333,
      "grad_norm": 0.012042745016515255,
      "learning_rate": 7.5e-07,
      "loss": 0.0252,
      "num_tokens": 31954229.0,
      "reward": 1.1889784336090088,
      "reward_std": 0.20391391217708588,
      "rewards/accuracy_reward_step": 0.375,
      "rewards/final_brier_reward_step": 0.6357695460319519,
      "rewards/format_reward_step": 0.9921875,
      "step": 173
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.378698892891407,
      "aux_distill/mean_u": 0.2000926506861302,
      "aux_distill/n_active_tok": 140.96875,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.4848149113660062,
      "calib/avg_num_step_conf": 2.19921875,
      "calib/ece": 0.4685140562248996,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.04819277108433735,
      "calib/gap": -0.06858315954118874,
      "calib/mean_conf": 0.07469879518072288,
      "calib/mu_c": 0.03696428571428572,
      "calib/mu_w": 0.10554744525547446,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.04670682730923695,
      "calib/std_conf": 0.2103945653553783,
      "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": 1930.0,
      "completions/max_terminated_length": 1930.0,
      "completions/mean_length": 257.1015625,
      "completions/mean_terminated_length": 257.1015625,
      "completions/min_length": 84.0,
      "completions/min_terminated_length": 84.0,
      "epoch": 0.1856,
      "grad_norm": 0.012609269469976425,
      "learning_rate": 7.222222222222222e-07,
      "loss": 0.0738,
      "num_tokens": 32124279.0,
      "reward": 1.1833367347717285,
      "reward_std": 0.24715489149093628,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/final_brier_reward_step": 0.5190171599388123,
      "rewards/format_reward_step": 0.97265625,
      "step": 174
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39681370416656137,
      "aux_distill/mean_u": 0.16005521341916332,
      "aux_distill/n_active_tok": 130.4375,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.4434292122971368,
      "calib/avg_num_step_conf": 2.03515625,
      "calib/ece": 0.40518072289156626,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.05220883534136546,
      "calib/gap": 0.0007111756168360073,
      "calib/mean_conf": 0.08261044176706828,
      "calib/mu_c": 0.08301886792452831,
      "calib/mu_w": 0.0823076923076923,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.03104417670682731,
      "calib/std_conf": 0.2206434840179149,
      "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": 896.0,
      "completions/max_terminated_length": 896.0,
      "completions/mean_length": 238.83203125,
      "completions/mean_terminated_length": 239.76864624023438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 66.0,
      "epoch": 0.18666666666666668,
      "grad_norm": 0.013867022469639778,
      "learning_rate": 6.944444444444446e-07,
      "loss": 0.0267,
      "num_tokens": 32291244.0,
      "reward": 1.1870672702789307,
      "reward_std": 0.2362235188484192,
      "rewards/accuracy_reward_step": 0.4140625,
      "rewards/final_brier_reward_step": 0.5733535289764404,
      "rewards/format_reward_step": 0.97265625,
      "step": 175
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4129004646092653,
      "aux_distill/mean_u": 0.22087444549099958,
      "aux_distill/n_active_tok": 140.8125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.493684012066365,
      "calib/avg_num_step_conf": 2.19921875,
      "calib/ece": 0.4260869565217391,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.043478260869565216,
      "calib/gap": 0.02046631473102059,
      "calib/mean_conf": 0.07549407114624507,
      "calib/mu_c": 0.08649572649572648,
      "calib/mu_w": 0.06602941176470589,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.019565217391304353,
      "calib/std_conf": 0.2103907830333878,
      "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": 1410.0,
      "completions/max_terminated_length": 1410.0,
      "completions/mean_length": 247.1328125,
      "completions/mean_terminated_length": 247.1328125,
      "completions/min_length": 74.0,
      "completions/min_terminated_length": 74.0,
      "epoch": 0.18773333333333334,
      "grad_norm": 0.01126535888761282,
      "learning_rate": 6.666666666666667e-07,
      "loss": 0.0758,
      "num_tokens": 32458574.0,
      "reward": 1.23163902759552,
      "reward_std": 0.1989196240901947,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/final_brier_reward_step": 0.5609343647956848,
      "rewards/format_reward_step": 0.98828125,
      "step": 176
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3774943738244474,
      "aux_distill/mean_u": 0.14791538337202564,
      "aux_distill/n_active_tok": 142.5625,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.4859767891682786,
      "calib/avg_num_step_conf": 2.2421875,
      "calib/ece": 0.4026294820717132,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.055776892430278883,
      "calib/gap": 0.022279819471308845,
      "calib/mean_conf": 0.09075697211155379,
      "calib/mu_c": 0.10327272727272728,
      "calib/mu_w": 0.08099290780141843,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.027569721115537845,
      "calib/std_conf": 0.22860529995719947,
      "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": 2377.0,
      "completions/max_terminated_length": 2377.0,
      "completions/mean_length": 251.8984375,
      "completions/mean_terminated_length": 252.88629150390625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 86.0,
      "epoch": 0.1888,
      "grad_norm": 0.012300615198910236,
      "learning_rate": 6.388888888888889e-07,
      "loss": 0.1106,
      "num_tokens": 32626892.0,
      "reward": 1.2100297212600708,
      "reward_std": 0.2131548970937729,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.5802156329154968,
      "rewards/format_reward_step": 0.98046875,
      "step": 177
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3958655558526516,
      "aux_distill/mean_u": 0.18713042017993753,
      "aux_distill/n_active_tok": 149.875,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.48383068647540983,
      "calib/avg_num_step_conf": 2.33984375,
      "calib/ece": 0.4704800000000001,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.016,
      "calib/gap": -0.014066342213114758,
      "calib/mean_conf": 0.04712,
      "calib/mu_c": 0.039918032786885244,
      "calib/mu_w": 0.053984375,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0148,
      "calib/std_conf": 0.13297708674805597,
      "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": 2347.0,
      "completions/max_terminated_length": 2347.0,
      "completions/mean_length": 252.0234375,
      "completions/mean_terminated_length": 252.0234375,
      "completions/min_length": 58.0,
      "completions/min_terminated_length": 58.0,
      "epoch": 0.18986666666666666,
      "grad_norm": 0.011569799855351448,
      "learning_rate": 6.111111111111112e-07,
      "loss": 0.0964,
      "num_tokens": 32797482.0,
      "reward": 1.2241487503051758,
      "reward_std": 0.25436022877693176,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.5186101198196411,
      "rewards/format_reward_step": 0.9765625,
      "step": 178
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39037828613072634,
      "aux_distill/mean_u": 0.17526136001380732,
      "aux_distill/n_active_tok": 165.3125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5905647163684982,
      "calib/avg_num_step_conf": 2.58203125,
      "calib/ece": 0.5101190476190478,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.011904761904761904,
      "calib/gap": 0.03383268606353855,
      "calib/mean_conf": 0.04432539682539682,
      "calib/mu_c": 0.05949640287769784,
      "calib/mu_w": 0.025663716814159295,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0014285714285714284,
      "calib/std_conf": 0.127216621241411,
      "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": 2311.0,
      "completions/max_terminated_length": 2311.0,
      "completions/mean_length": 266.0546875,
      "completions/mean_terminated_length": 266.0546875,
      "completions/min_length": 76.0,
      "completions/min_terminated_length": 76.0,
      "epoch": 0.19093333333333334,
      "grad_norm": 0.012259440496563911,
      "learning_rate": 5.833333333333334e-07,
      "loss": 0.0898,
      "num_tokens": 32971856.0,
      "reward": 1.2747502326965332,
      "reward_std": 0.17485995590686798,
      "rewards/accuracy_reward_step": 0.54296875,
      "rewards/final_brier_reward_step": 0.48700040578842163,
      "rewards/format_reward_step": 0.9765625,
      "step": 179
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3385851050261408,
      "aux_distill/mean_u": 0.14330181234511938,
      "aux_distill/n_active_tok": 140.0625,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.4555697823303457,
      "calib/avg_num_step_conf": 2.19140625,
      "calib/ece": 0.5428968253968254,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.015873015873015872,
      "calib/gap": -0.023460947503201025,
      "calib/mean_conf": 0.04432539682539683,
      "calib/mu_c": 0.03408450704225352,
      "calib/mu_w": 0.057545454545454545,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.011865079365079365,
      "calib/std_conf": 0.13172520274245528,
      "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": 1013.0,
      "completions/max_terminated_length": 1013.0,
      "completions/mean_length": 259.75,
      "completions/mean_terminated_length": 261.7952880859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 77.0,
      "epoch": 0.192,
      "grad_norm": 0.01267837081104517,
      "learning_rate": 5.555555555555555e-07,
      "loss": 0.015,
      "num_tokens": 33142208.0,
      "reward": 1.2711178064346313,
      "reward_std": 0.2109297513961792,
      "rewards/accuracy_reward_step": 0.5546875,
      "rewards/final_brier_reward_step": 0.4484855532646179,
      "rewards/format_reward_step": 0.984375,
      "step": 180
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3581990972161293,
      "aux_distill/mean_u": 0.1333516353267837,
      "aux_distill/n_active_tok": 126.53125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5822570123939987,
      "calib/avg_num_step_conf": 1.9765625,
      "calib/ece": 0.388804780876494,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.02390438247011952,
      "calib/gap": 0.034264187866927594,
      "calib/mean_conf": 0.05549800796812749,
      "calib/mu_c": 0.07542857142857143,
      "calib/mu_w": 0.041164383561643834,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.012988047808764941,
      "calib/std_conf": 0.16675019784722173,
      "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": 2409.0,
      "completions/max_terminated_length": 2409.0,
      "completions/mean_length": 236.81640625,
      "completions/mean_terminated_length": 237.7451171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 82.0,
      "epoch": 0.19306666666666666,
      "grad_norm": 0.012049591168761253,
      "learning_rate": 5.277777777777779e-07,
      "loss": 0.0878,
      "num_tokens": 33309097.0,
      "reward": 1.201343059539795,
      "reward_std": 0.20447835326194763,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.601905107498169,
      "rewards/format_reward_step": 0.98046875,
      "step": 181
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3256945216562599,
      "aux_distill/mean_u": 0.1309827357755496,
      "aux_distill/n_active_tok": 129.15625,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.4998382714452064,
      "calib/avg_num_step_conf": 2.015625,
      "calib/ece": 0.5071887550200803,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.024096385542168676,
      "calib/gap": -0.011099754172596711,
      "calib/mean_conf": 0.05755020080321286,
      "calib/mu_c": 0.052290076335877865,
      "calib/mu_w": 0.06338983050847458,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.019317269076305224,
      "calib/std_conf": 0.1616556608992178,
      "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": 2266.0,
      "completions/max_terminated_length": 2266.0,
      "completions/mean_length": 245.984375,
      "completions/mean_terminated_length": 246.94903564453125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 72.0,
      "epoch": 0.19413333333333332,
      "grad_norm": 0.011237970553338528,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.0491,
      "num_tokens": 33478229.0,
      "reward": 1.2409536838531494,
      "reward_std": 0.19657526910305023,
      "rewards/accuracy_reward_step": 0.51171875,
      "rewards/final_brier_reward_step": 0.4858136773109436,
      "rewards/format_reward_step": 0.97265625,
      "step": 182
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3604277758859098,
      "aux_distill/mean_u": 0.14891531079840287,
      "aux_distill/n_active_tok": 121.3125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.48015417940876665,
      "calib/avg_num_step_conf": 1.89453125,
      "calib/ece": 0.4139525691699604,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.023715415019762844,
      "calib/gap": 0.0016055045871559495,
      "calib/mean_conf": 0.05569169960474308,
      "calib/mu_c": 0.05660550458715595,
      "calib/mu_w": 0.055,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.01940711462450593,
      "calib/std_conf": 0.16343646494401404,
      "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": 1800.0,
      "completions/max_terminated_length": 1800.0,
      "completions/mean_length": 248.0859375,
      "completions/mean_terminated_length": 248.0859375,
      "completions/min_length": 79.0,
      "completions/min_terminated_length": 79.0,
      "epoch": 0.1952,
      "grad_norm": 0.011038307100534439,
      "learning_rate": 4.7222222222222226e-07,
      "loss": 0.058,
      "num_tokens": 33648419.0,
      "reward": 1.2105414867401123,
      "reward_std": 0.21297506988048553,
      "rewards/accuracy_reward_step": 0.42578125,
      "rewards/final_brier_reward_step": 0.5812394618988037,
      "rewards/format_reward_step": 0.98828125,
      "step": 183
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39351186295971274,
      "aux_distill/mean_u": 0.18142064296222152,
      "aux_distill/n_active_tok": 134.8125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.45229928861788615,
      "calib/avg_num_step_conf": 2.10546875,
      "calib/ece": 0.47968127490039836,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.02390438247011952,
      "calib/gap": -0.03226689532520326,
      "calib/mean_conf": 0.05442231075697211,
      "calib/mu_c": 0.037967479674796745,
      "calib/mu_w": 0.070234375,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.02203187250996016,
      "calib/std_conf": 0.15731045300848923,
      "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": 2236.0,
      "completions/max_terminated_length": 2236.0,
      "completions/mean_length": 261.609375,
      "completions/mean_terminated_length": 261.609375,
      "completions/min_length": 76.0,
      "completions/min_terminated_length": 76.0,
      "epoch": 0.19626666666666667,
      "grad_norm": 0.012052321806550026,
      "learning_rate": 4.444444444444445e-07,
      "loss": 0.0551,
      "num_tokens": 33820671.0,
      "reward": 1.225361704826355,
      "reward_std": 0.23105868697166443,
      "rewards/accuracy_reward_step": 0.48046875,
      "rewards/final_brier_reward_step": 0.50931715965271,
      "rewards/format_reward_step": 0.98046875,
      "step": 184
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3678031573072076,
      "aux_distill/mean_u": 0.17732553736026746,
      "aux_distill/n_active_tok": 148.75,
      "calib/answer_extract_rate": 0.95703125,
      "calib/auroc": 0.4503010033444816,
      "calib/avg_num_step_conf": 2.3203125,
      "calib/ece": 0.43204081632653063,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.95703125,
      "calib/frac_conf_gt_0.9": 0.024489795918367346,
      "calib/gap": 0.015204013377926422,
      "calib/mean_conf": 0.06036734693877551,
      "calib/mu_c": 0.06843478260869565,
      "calib/mu_w": 0.05323076923076923,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.011510204081632655,
      "calib/std_conf": 0.17128645226548408,
      "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": 2368.0,
      "completions/max_terminated_length": 2368.0,
      "completions/mean_length": 262.9921875,
      "completions/mean_terminated_length": 264.0235595703125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 97.0,
      "epoch": 0.19733333333333333,
      "grad_norm": 0.010781550779938698,
      "learning_rate": 4.1666666666666667e-07,
      "loss": 0.0398,
      "num_tokens": 33994917.0,
      "reward": 1.1965997219085693,
      "reward_std": 0.19675306975841522,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.537730872631073,
      "rewards/format_reward_step": 0.95703125,
      "step": 185
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40662080235779285,
      "aux_distill/mean_u": 0.20044403718500448,
      "aux_distill/n_active_tok": 142.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5070610924474306,
      "calib/avg_num_step_conf": 2.265625,
      "calib/ece": 0.42274509803921567,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.023529411764705882,
      "calib/gap": 0.005263157894736831,
      "calib/mean_conf": 0.055686274509803936,
      "calib/mu_c": 0.058596491228070174,
      "calib/mu_w": 0.053333333333333344,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.015686274509803925,
      "calib/std_conf": 0.16465120474316494,
      "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": 867.0,
      "completions/max_terminated_length": 867.0,
      "completions/mean_length": 239.94921875,
      "completions/mean_terminated_length": 240.8902130126953,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 73.0,
      "epoch": 0.1984,
      "grad_norm": 0.012303094379603863,
      "learning_rate": 3.8888888888888895e-07,
      "loss": 0.0567,
      "num_tokens": 34161384.0,
      "reward": 1.23370361328125,
      "reward_std": 0.18402235209941864,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.5728757381439209,
      "rewards/format_reward_step": 0.99609375,
      "step": 186
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3461750391870737,
      "aux_distill/mean_u": 0.1459223011667055,
      "aux_distill/n_active_tok": 174.1875,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.48686660460167575,
      "calib/avg_num_step_conf": 2.71875,
      "calib/ece": 0.3996385542168674,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.060240963855421686,
      "calib/gap": -0.0027922596089905716,
      "calib/mean_conf": 0.08959839357429719,
      "calib/mu_c": 0.08796116504854369,
      "calib/mu_w": 0.09075342465753426,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.03779116465863453,
      "calib/std_conf": 0.23633717523704523,
      "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": 2082.0,
      "completions/max_terminated_length": 2082.0,
      "completions/mean_length": 264.1328125,
      "completions/mean_terminated_length": 264.1328125,
      "completions/min_length": 72.0,
      "completions/min_terminated_length": 72.0,
      "epoch": 0.19946666666666665,
      "grad_norm": 0.012116963975131512,
      "learning_rate": 3.611111111111111e-07,
      "loss": 0.1041,
      "num_tokens": 34330546.0,
      "reward": 1.1781506538391113,
      "reward_std": 0.2714693248271942,
      "rewards/accuracy_reward_step": 0.40234375,
      "rewards/final_brier_reward_step": 0.5789574384689331,
      "rewards/format_reward_step": 0.97265625,
      "step": 187
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36286410223692656,
      "aux_distill/mean_u": 0.18276469576506474,
      "aux_distill/n_active_tok": 138.34375,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.4541254554919313,
      "calib/avg_num_step_conf": 2.16015625,
      "calib/ece": 0.5208032128514055,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.012048192771084338,
      "calib/gap": -0.01723256116605934,
      "calib/mean_conf": 0.044658634538152615,
      "calib/mu_c": 0.03683823529411765,
      "calib/mu_w": 0.05407079646017699,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.00963855421686747,
      "calib/std_conf": 0.1185148387492385,
      "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": 0.01171875,
      "completions/max_length": 2265.0,
      "completions/max_terminated_length": 2265.0,
      "completions/mean_length": 260.0546875,
      "completions/mean_terminated_length": 263.1383361816406,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 67.0,
      "epoch": 0.20053333333333334,
      "grad_norm": 0.011851982213556767,
      "learning_rate": 3.3333333333333335e-07,
      "loss": 0.0989,
      "num_tokens": 34501192.0,
      "reward": 1.2500507831573486,
      "reward_std": 0.21181350946426392,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/final_brier_reward_step": 0.4649452865123749,
      "rewards/format_reward_step": 0.97265625,
      "step": 188
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33696536626666784,
      "aux_distill/mean_u": 0.16132667425486644,
      "aux_distill/n_active_tok": 128.0625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5084586466165414,
      "calib/avg_num_step_conf": 2.00390625,
      "calib/ece": 0.4983070866141732,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.031496062992125984,
      "calib/gap": 0.0394874686716792,
      "calib/mean_conf": 0.06106299212598426,
      "calib/mu_c": 0.07878571428571429,
      "calib/mu_w": 0.03929824561403509,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.004094488188976378,
      "calib/std_conf": 0.17610632293269302,
      "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": 536.0,
      "completions/max_terminated_length": 536.0,
      "completions/mean_length": 222.80078125,
      "completions/mean_terminated_length": 223.67453002929688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.2016,
      "grad_norm": 0.014447653666138649,
      "learning_rate": 3.055555555555556e-07,
      "loss": 0.0249,
      "num_tokens": 34665997.0,
      "reward": 1.291475534439087,
      "reward_std": 0.20227526128292084,
      "rewards/accuracy_reward_step": 0.546875,
      "rewards/final_brier_reward_step": 0.4970136880874634,
      "rewards/format_reward_step": 0.9921875,
      "step": 189
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34723477764055133,
      "aux_distill/mean_u": 0.15021027751099436,
      "aux_distill/n_active_tok": 120.40625,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5632671422145107,
      "calib/avg_num_step_conf": 1.890625,
      "calib/ece": 0.45008000000000004,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.036,
      "calib/gap": -0.006197545144913552,
      "calib/mean_conf": 0.06928,
      "calib/mu_c": 0.06598290598290599,
      "calib/mu_w": 0.07218045112781954,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.02568,
      "calib/std_conf": 0.19074664243440825,
      "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": 878.0,
      "completions/max_terminated_length": 878.0,
      "completions/mean_length": 249.45703125,
      "completions/mean_terminated_length": 250.435302734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.20266666666666666,
      "grad_norm": 0.010926991701126099,
      "learning_rate": 2.7777777777777776e-07,
      "loss": 0.0197,
      "num_tokens": 34835466.0,
      "reward": 1.2151250839233398,
      "reward_std": 0.23179908096790314,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/final_brier_reward_step": 0.5396250486373901,
      "rewards/format_reward_step": 0.9765625,
      "step": 190
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31737090460956097,
      "aux_distill/mean_u": 0.10964079806256934,
      "aux_distill/n_active_tok": 147.28125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4878162782001745,
      "calib/avg_num_step_conf": 2.33984375,
      "calib/ece": 0.4236078431372549,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.047058823529411764,
      "calib/gap": -0.003901283809048986,
      "calib/mean_conf": 0.07686274509803921,
      "calib/mu_c": 0.07469026548672567,
      "calib/mu_w": 0.07859154929577465,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.028666666666666663,
      "calib/std_conf": 0.211171245406618,
      "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": 943.0,
      "completions/max_terminated_length": 943.0,
      "completions/mean_length": 231.07421875,
      "completions/mean_terminated_length": 231.98040771484375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.20373333333333332,
      "grad_norm": 0.011647163890302181,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.0283,
      "num_tokens": 34998789.0,
      "reward": 1.2246136665344238,
      "reward_std": 0.18209809064865112,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.5703210830688477,
      "rewards/format_reward_step": 0.99609375,
      "step": 191
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36972838593646884,
      "aux_distill/mean_u": 0.178276600960977,
      "aux_distill/n_active_tok": 134.90625,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5226031746031746,
      "calib/avg_num_step_conf": 2.125,
      "calib/ece": 0.4762948207171314,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0199203187250996,
      "calib/gap": -0.004618412698412691,
      "calib/mean_conf": 0.04960159362549802,
      "calib/mu_c": 0.04730158730158731,
      "calib/mu_w": 0.05192,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.01195219123505976,
      "calib/std_conf": 0.14437611570583614,
      "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": 673.0,
      "completions/max_terminated_length": 673.0,
      "completions/mean_length": 228.26953125,
      "completions/mean_terminated_length": 229.1647186279297,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 65.0,
      "epoch": 0.2048,
      "grad_norm": 0.011698748916387558,
      "learning_rate": 2.2222222222222224e-07,
      "loss": 0.0397,
      "num_tokens": 35162202.0,
      "reward": 1.2384189367294312,
      "reward_std": 0.1519850641489029,
      "rewards/accuracy_reward_step": 0.4921875,
      "rewards/final_brier_reward_step": 0.5119941830635071,
      "rewards/format_reward_step": 0.98046875,
      "step": 192
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3735968554392457,
      "aux_distill/mean_u": 0.1607234325833779,
      "aux_distill/n_active_tok": 113.8125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5134469696969697,
      "calib/avg_num_step_conf": 1.77734375,
      "calib/ece": 0.42417322834645665,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.051181102362204724,
      "calib/gap": -0.015446969696969695,
      "calib/mean_conf": 0.08039370078740159,
      "calib/mu_c": 0.07163636363636364,
      "calib/mu_w": 0.08708333333333333,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.035748031496062996,
      "calib/std_conf": 0.22381235663819363,
      "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": 768.0,
      "completions/max_terminated_length": 768.0,
      "completions/mean_length": 233.796875,
      "completions/mean_terminated_length": 233.796875,
      "completions/min_length": 77.0,
      "completions/min_terminated_length": 77.0,
      "epoch": 0.20586666666666667,
      "grad_norm": 0.012572907842695713,
      "learning_rate": 1.9444444444444447e-07,
      "loss": 0.0844,
      "num_tokens": 35327766.0,
      "reward": 1.2097558975219727,
      "reward_std": 0.2349000871181488,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.5679492354393005,
      "rewards/format_reward_step": 0.9921875,
      "step": 193
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37450212286785245,
      "aux_distill/mean_u": 0.18273606891361022,
      "aux_distill/n_active_tok": 112.59375,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5285008134150919,
      "calib/avg_num_step_conf": 1.78515625,
      "calib/ece": 0.44482213438735185,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.03162055335968379,
      "calib/gap": 0.02257977724940559,
      "calib/mean_conf": 0.06363636363636364,
      "calib/mu_c": 0.07532786885245903,
      "calib/mu_w": 0.05274809160305344,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.013122529644268773,
      "calib/std_conf": 0.1755899659662866,
      "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": 774.0,
      "completions/max_terminated_length": 774.0,
      "completions/mean_length": 215.4296875,
      "completions/mean_terminated_length": 216.27452087402344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 79.0,
      "epoch": 0.20693333333333333,
      "grad_norm": 0.013541937805712223,
      "learning_rate": 1.6666666666666668e-07,
      "loss": 0.0431,
      "num_tokens": 35488860.0,
      "reward": 1.2452245950698853,
      "reward_std": 0.2098076343536377,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.5490429401397705,
      "rewards/format_reward_step": 0.98828125,
      "step": 194
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40552747156471014,
      "aux_distill/mean_u": 0.20745253260927174,
      "aux_distill/n_active_tok": 114.34375,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5586550836550837,
      "calib/avg_num_step_conf": 1.78515625,
      "calib/ece": 0.5103984063745021,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.03187250996015936,
      "calib/gap": 0.03267245817245817,
      "calib/mean_conf": 0.06290836653386454,
      "calib/mu_c": 0.07735714285714285,
      "calib/mu_w": 0.04468468468468468,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.007768924302788845,
      "calib/std_conf": 0.1762624005123713,
      "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": 2493.0,
      "completions/max_terminated_length": 2493.0,
      "completions/mean_length": 227.03515625,
      "completions/mean_terminated_length": 227.03515625,
      "completions/min_length": 71.0,
      "completions/min_terminated_length": 71.0,
      "epoch": 0.208,
      "grad_norm": 0.015685738995671272,
      "learning_rate": 1.3888888888888888e-07,
      "loss": 0.0641,
      "num_tokens": 35652965.0,
      "reward": 1.279039978981018,
      "reward_std": 0.19776293635368347,
      "rewards/accuracy_reward_step": 0.546875,
      "rewards/final_brier_reward_step": 0.4838613271713257,
      "rewards/format_reward_step": 0.98046875,
      "step": 195
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3677436811849475,
      "aux_distill/mean_u": 0.14246220831369982,
      "aux_distill/n_active_tok": 106.03125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.45757123304715686,
      "calib/avg_num_step_conf": 1.66015625,
      "calib/ece": 0.41980392156862745,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.047058823529411764,
      "calib/gap": -0.01334826427771553,
      "calib/mean_conf": 0.08545098039215686,
      "calib/mu_c": 0.07807017543859651,
      "calib/mu_w": 0.09141843971631204,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.02909803921568627,
      "calib/std_conf": 0.2133479450013375,
      "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": 591.0,
      "completions/max_terminated_length": 591.0,
      "completions/mean_length": 199.859375,
      "completions/mean_terminated_length": 200.6431427001953,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 73.0,
      "epoch": 0.20906666666666668,
      "grad_norm": 0.014952057972550392,
      "learning_rate": 1.1111111111111112e-07,
      "loss": 0.045,
      "num_tokens": 35806673.0,
      "reward": 1.2272090911865234,
      "reward_std": 0.20598194003105164,
      "rewards/accuracy_reward_step": 0.4453125,
      "rewards/final_brier_reward_step": 0.5676996111869812,
      "rewards/format_reward_step": 0.99609375,
      "step": 196
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35766037181019783,
      "aux_distill/mean_u": 0.19740065525817932,
      "aux_distill/n_active_tok": 118.71875,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.5157745363342354,
      "calib/avg_num_step_conf": 1.87890625,
      "calib/ece": 0.4115725806451612,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96875,
      "calib/frac_conf_gt_0.9": 0.028225806451612902,
      "calib/gap": 0.016730908850900936,
      "calib/mean_conf": 0.06447580645161291,
      "calib/mu_c": 0.07385321100917432,
      "calib/mu_w": 0.05712230215827338,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.018266129032258063,
      "calib/std_conf": 0.17584968671538206,
      "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": 220.8984375,
      "completions/mean_terminated_length": 221.7647247314453,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 70.0,
      "epoch": 0.21013333333333334,
      "grad_norm": 0.013707943260669708,
      "learning_rate": 8.333333333333334e-08,
      "loss": 0.0255,
      "num_tokens": 35968279.0,
      "reward": 1.200000286102295,
      "reward_std": 0.2242618203163147,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.5718753933906555,
      "rewards/format_reward_step": 0.96875,
      "step": 197
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35180373350158334,
      "aux_distill/mean_u": 0.12420010954716808,
      "aux_distill/n_active_tok": 126.34375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.51478895588702,
      "calib/avg_num_step_conf": 1.97265625,
      "calib/ece": 0.4947222222222223,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.01984126984126984,
      "calib/gap": 0.0274408124404951,
      "calib/mean_conf": 0.056309523809523816,
      "calib/mu_c": 0.06883211678832118,
      "calib/mu_w": 0.041391304347826084,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0036904761904761893,
      "calib/std_conf": 0.1475846254029491,
      "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": 2630.0,
      "completions/max_terminated_length": 2630.0,
      "completions/mean_length": 231.1640625,
      "completions/mean_terminated_length": 231.1640625,
      "completions/min_length": 67.0,
      "completions/min_terminated_length": 67.0,
      "epoch": 0.2112,
      "grad_norm": 0.012614569626748562,
      "learning_rate": 5.555555555555556e-08,
      "loss": 0.0699,
      "num_tokens": 36132841.0,
      "reward": 1.2765079736709595,
      "reward_std": 0.21813374757766724,
      "rewards/accuracy_reward_step": 0.53515625,
      "rewards/final_brier_reward_step": 0.4983285069465637,
      "rewards/format_reward_step": 0.984375,
      "step": 198
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34593457635492086,
      "aux_distill/mean_u": 0.17202861813479953,
      "aux_distill/n_active_tok": 136.84375,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5353656949505465,
      "calib/avg_num_step_conf": 2.13671875,
      "calib/ece": 0.4144979919678715,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.03614457831325301,
      "calib/gap": 0.03869143675169182,
      "calib/mean_conf": 0.0689558232931727,
      "calib/mu_c": 0.09008849557522124,
      "calib/mu_w": 0.051397058823529414,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.014819277108433735,
      "calib/std_conf": 0.19453299566354348,
      "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": 2070.0,
      "completions/max_terminated_length": 2070.0,
      "completions/mean_length": 247.671875,
      "completions/mean_terminated_length": 249.62203979492188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 60.0,
      "epoch": 0.21226666666666666,
      "grad_norm": 0.012673155404627323,
      "learning_rate": 2.777777777777778e-08,
      "loss": 0.0457,
      "num_tokens": 36300445.0,
      "reward": 1.212408423423767,
      "reward_std": 0.24468745291233063,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.5693480372428894,
      "rewards/format_reward_step": 0.97265625,
      "step": 199
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3864405620843172,
      "aux_distill/mean_u": 0.16800184582972094,
      "aux_distill/n_active_tok": 121.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5787498461917067,
      "calib/avg_num_step_conf": 1.8984375,
      "calib/ece": 0.46976470588235286,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0196078431372549,
      "calib/gap": 0.01872462163159838,
      "calib/mean_conf": 0.054313725490196085,
      "calib/mu_c": 0.06356589147286822,
      "calib/mu_w": 0.044841269841269835,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.009098039215686275,
      "calib/std_conf": 0.15270187412284936,
      "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": 2597.0,
      "completions/max_terminated_length": 2597.0,
      "completions/mean_length": 238.53515625,
      "completions/mean_terminated_length": 238.53515625,
      "completions/min_length": 69.0,
      "completions/min_terminated_length": 69.0,
      "epoch": 0.21333333333333335,
      "grad_norm": 0.011946732178330421,
      "learning_rate": 0.0,
      "loss": 0.0595,
      "num_tokens": 36469558.0,
      "reward": 1.2669954299926758,
      "reward_std": 0.16499918699264526,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.5300847887992859,
      "rewards/format_reward_step": 0.99609375,
      "step": 200
    },
    {
      "epoch": 0.21333333333333335,
      "step": 200,
      "total_flos": 0.0,
      "train_loss": 0.07998869574104901,
      "train_runtime": 11768.0831,
      "train_samples_per_second": 4.351,
      "train_steps_per_second": 0.017
    }
  ],
  "logging_steps": 1,
  "max_steps": 200,
  "num_input_tokens_seen": 36469558,
  "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
}